Abstract:

Corpus linguistics, the analysis of large, computerized collections (corpora) of real-life language, is a valuable tool for the language learning researcher. Analysis with corpus linguistics can reveal patterns that show how a language is really used in real-life situations. For teachers, using such documented patterns and derived real-world examples can be an effective teaching tool. Moreover, learners can also use corpora tools to discover language patterns for themselves, and answer their own questions about language use.

Even despite its immense promise, corpus linguistics has been little used to formulate practical applications for English as a Second speaking and pronunciation teachers; the reasons gap for this are multi-fold. In general, corpus linguistics research is rather theoretical in nature. Furthermore, in part, because of the high cost of transcription, corpus linguistics tends focuses on written over transcribed, spoken language. Furthermore, in ESL research overall, pronunciation is a less-studied, albeit significant area. All of these factors jointly have created a dearth of practical corpus linguistics research that offers ESL speaking and pronunciation teaching implications; therefore, this research uses corpus linguistics to offer insight into what a large sample of English speakers actually said in the context of a university campus, and draws specific implications for pronunciation teaching in particular.

Specifically, this particular corpus linguistics study looks at lexical bundles—herein defined as 4-word phrases that occur more often than chance. Bundles were extracted using AntConc (a free software package) and MICASE (Michigan Corpus of Spoken Academic English), a transcribed corpus of 1.84 million words that were spoken on a university campus.

The lexical bundles most often used were analyzed to determine:
1) What do these bundles actually do in a communicative context – in fact, what do they mean?
2) Secondly, what are the implications of these findings for speaking and pronunciation teaching?

Through the data, it will be shown that in university speech, a majority of speech is participant-oriented, that is, speech that frames opinions or directly engages the listener. Without adequate learner use of these bundles and clear pronunciation of these bundles, communication may be hampered. Specific to pronunciation, in particular, connected speech will be shown as an essential pronunciation feature in the most highly used bundles, and specific implications for teaching will be outlined.

Keywords:

English as a Second Language, Corpus Linguistics, Pronunciation, Lexical Bundles, Connected Speech.