| This thesis is an investigation of the role of formulaic language in second language (L2) speech fluency development, within a cognitive and information processing framework. Fluency has been studied and defined in terms of temporal variables of speech such as rate of speech, pause frequency and distribution, and the length of fluent runs between pauses. It has been suggested by several researchers that the key to fluency in spontaneous speech is mastery of a repertoire of formulaic language sequences, multiword strings processed mentally as single words (Schmidt, 1991; Towell, Hawkins, and Bazergui, 1996; Chambers, 1998). If formulaic sequences are automatized or stored and retrieved as wholes from long term memory so as to allow longer lexical units to be produced within the limits of controlled processing (McLaughlin, Rossman, and McLeod, 1983; Kahnemann and Treismann, 1984; DeKeyser, 2001) and short term memory (Anderson, 1983; Baddeley, 1988), then they may facilitate spontaneous speech under the constraints of real time. The present study was designed to examine whether this could be so.; The study draws on a synthesis of research from three areas: fluency and its development in second language (L2) speech; formulaic language, multi-word lexical units which are stored and retrieved in long-term memory so as to be retrieved as wholes; social and cultural factors related to fluency development and formulaic language use, including first language and culture, voice, and identity. The research was interpreted in light of psycholinguistic knowledge about mental processes underlying L2 speech production, particularly the growing evidence that formulaic language sequences are fundamental to fluent language production as they allow production to occur despite the restrictions of controlled processing and the constraints of short term memory capacity.; The hypotheses which frame the research centre around the idea that increased use of formulaic language units by learners over time facilitates the development of speech fluency as measured by temporal variables such as speech rate, pause phenomena, and the length of fluent runs occurring between pauses. Specifically, it was hypothesized that, with continued learning and experience, L2 speech would exhibit a faster rate of production, a greater proportion of production time spent speaking as opposed to pausing, longer runs between pauses, and that formulaic sequences would appear more frequently in the longer runs between pauses. (Abstract shortened by UMI.)... |
