An Empirical Study On Features And Reasons Of Disfluencies In Consecutive Interpreting Output Of Trainees Interpreters

Everyday spontaneous speech, as a result of speech production difficulties, is peppered with disfluent phenomenon such as filled pauses, lengthenings, repeats and repairs. It is even more so with consecutive interpreting, a much more complex and energy taxing task. Despite conscientious efforts of interpreters to stay clear of them, disfluencies are ever present in the output of C-E CI. Yet, studies on disfluencies in CI are still relatively lacking, especially those probing into underlying triggers.Building on work of previous researchers, this thesis sets out to examine disfluencies in the output of trainee C-E CI interpreters in terms of frequency, features and root causes. This will shed some light on the cognitive processes of speech production and CI, and provide valuable information of most troubling areas for student interpreters. To this end, fourteen postgraduates majoring in interpreting from GUDFS are invited to participate in an experiment, whose interpreting outputs as well as explanations during retrospective interviews are kept on record for analysis.The gathered data are analyzed under the guidance of Levelt’s speech production model and Gile’s CI effort model. Disfluencies are categorized into filled pauses, interruptions(lengthening, repeat, restructuring and false start) and error type repairs(phonological repair, lexical repair and syntactical repair). The triggers as accounted for by subjects fall into three general categories, 1)speech production causes, including conceptual difficulty, grammatical encoding difficulty, phonological encoding difficulty and self-correction; 2) CI specific causes, which are note deciphering difficulty, retrieving source text difficulty, divided attention and strategy employing; 3)other factors like influence of source text, habit and no particular reason.The findings of this research demonstrate that the most frequent disfluencies in the C-E CI output of trainee interpreters are filled pauses, repeats and lexical repairs, while the least common type is false starts. The experimental data shows that for triggers of disfluencies, grammatical encoding difficulties play a markedly significant role. Other major triggers are source text retrieval difficulties and self-correction. The results also draw the author’s attention to the strategic value of disfluencies in C-E CI interpreting for interpreter are found to deliberately resort to disfluencies in order to deal with tricky situations. Regarding the correlation between disfluency types and underlying triggers, a certain level of correlation has indeed been found. For instance, a certain percentage of vocalized pauses are related to strategy employing, while repeats are often a result of frustrated attempts at retrieving source text information. Restructuring has commonly been attributed to note deciphering difficulties by the subjects.Implications of the conclusions include some insight into speech production system, monitoring in particular. The present study also provides support for previous proposition that certain disfluencies are linked to speech production stages. Other implications pertaining to C-E CI training suggests interpreting trainers need to enhance trainee interpreter’s sensitivity of disfluencies in their output, pay more attention to linguistic competencies like vocabulary building especially active vocabulary, and syntax mastery, as well as developing a consistent note system. Instructors can also expose trainee interpreters to strategic employment of disfluencies to deal with crisis.