No abstract

No abstract

No abstract

Knowledge and acquisition of formulaic sequences 57 an important influence on language learning in general (Dornyei and Skehan, 2003;Sawyer and Ranta, 2001), it is logical to suspect that they also influence the acquisition of formulaic sequences. Thus we will measure several of these factors (i.e. their age, gender, language aptitude, and motivation) in order to de-! termine their effect on formulaic sequence acquisition.. ' MethodologySelecting the target formulaic sequencesThe target formulaic sequences for this longitudinal study were chosen with three main criteria in mind. First, we needed to make sure that target formu-' laic sequences occurred with some degree of frequency in language use. Second, the target sequences would be incorporated into an EAP teaching environment, and so they needed to be connected with academic discourse. Third, in order to secure the cooperation of the language instructors at the Centre for English Language Education (CELE) at the University of Nottingham, the sequences also needed to be seen as useful to students and worthwhile to teach. Based on these criteria, the following procedure was carried out to identify and select appropriate formulaic sequences for the study. Our initial step was to consult reference materials which listed and discussed formulaic sequences of various kinds. We extracted 97 candidate formulaic sequences of an academic nature from the Biber et alls (1999) analysis of lexical bundles, and 59 candidate formulaic sequences from Nattinger and DeCarrico's (1992) functional analysis of lexical phrases. We then took words from Hylands (2000) list which are used to express doubt and certainty (e.g. dearly and approximately) and which are used as discourse markers (e.g. therefore and finally) and submitted them to a corpus analysis to see if they formed the core of a formulaic sequence (clearly the best). If so, they were added to our candidate list. Once the list of candidate formulaic sequences was compiled, we determined how frequently they occurred in each of three corpora. Frequency figures from the British National Corpus (BNC) gave an indication of how often the sequences occurred in general English, figures from the CANCODE corpus indicated how frequent they were in spoken discourse, and figures from the MICASE corpus showed their frequency in academic spoken discourse. Based on these frequency figures, we were able to identify the formulaic sequence candidates with the highest frequencies in written, spoken, and academic contexts.

No abstract