Abstract
A recent study by Cordier (2013) suggests that psycholinguistic formulaic sequences (multiword units that present a processing advantage to the individual speaker) may be more prevalent in L2 speakers than previously thought. The current study adopts the same identifcation process to explore the use of psycholinguistic formulaic sequences in the speech of Japanese Speakers of English (JSE).Eight adult JSE at intermediate or advanced levels of English each performed two speaking tasks: a structured interview and a narration task. Formulaic sequences were identifed on the basis of hierarchical conditions applied in strict order. The frst condition was fuency and the second condition checked for holisticity (using given diagnostic criteria). For each sample, two measures of formulaicity were calculated: FS% (the percentage of syllables that were part of a formulaic sequence) and ANR (the average number of formulaic syllables per run).The mean formulaicity of the samples (FS%=34.6%, ANR=1.64) suggests that psycholinguistic formulaic sequences, as defned and identifed here, may be a signifcant feature in the speech of intermediate/advanced JSE. The study also confrms the sensitivity of the results to task, with signifcantly more formulaic sequences used in the interview task than in the narration. Overall, the identifcation process was found to be a useful and systematic way of identifying formulaic sequences, but some further refnements of the diagnostic criteria and measures used are also suggested.

Suggested citation