Second language English listeners’ relative processing of coherence-based and frequency-based formulas: A corpus-based study

Abstract

Abstract This study investigated L2 English listeners’ processing of formulas, in terms of the impact of two different factors inherent in these formulas. One was the formulas’ level of coherence and the other was the formulas’ level of frequency. High-coherence formulas are considered to have specialized meanings, while high-frequency formulas are considered to be less specialized in meaning, commonly being composed of relatively simple words that often co-occur in speech. In previous research, in an academic context, Ellis, Simpson-Vlach and Maynard (2008. Formulaic language in native and second-language speakers: Psycholinguistics, corpus linguistics, and TESOL. TESOL Quarterly 41. 375–396. doi:10.1002/j.1545-7249.2008.tb00137.x) had found that a high level of coherence was the main factor facilitating L1 users’ receptive processing of formulas, while a high level of frequency was the main factor facilitating advanced L2 users’ receptive processing of formulas. Ellis, Simpson-Vlach and Maynard (2008. Formulaic language in native and second-language speakers: Psycholinguistics, corpus linguistics, and TESOL. TESOL Quarterly 41. 375–396. doi:10.1002/j.1545-7249.2008.tb00137.x), from a usage-based perspective, attributed these differences mainly to the greater length of time the L1 users had spent in learning formulas. Consequently, the current study investigated whether these processing differences between the two user groups in an academic context (seen as a possible developmental trend) would be apparent between proficient and less-proficient L2 listeners in a relatively less-challenging, general English environment. The study was considered important for possibly signaling the types of aural receptive formulas to foreground by L2 general English instructors and materials designers. The research examined two groups of L2 learners, one advanced and the other intermediate level, while they listened to four texts. A paused transcription technique elicited the listeners’ identification of targeted segments from the texts, many of which were classified through corpus analysis as containing more/less-coherent formulas or more/less-frequent formulas. Examination of how these formula types were processed by both proficiency groups, however, did not find major differences between the groups in their processing of the different formula types, and thus little evidence of a possible formula developmental trend.