A Corpus-based Study of Prefabricated Chunks Used by Professional Interpreters in Cantonese-English Simultaneous Interpreting

Previous studies (e.g. Goldman-Eisler 1958; Becker 1975; Levelt 1989; Wray 2002; Pawley & Syder 1983; Nattinger & DeCarrico 1992; Lewis 1993, 1997) in psycholinguistics and first language acquisition have unanimously demonstrated that fluent speech production depends largely on the successful use of formulaic expressions, i.e. prefabricated chunks as defined in this study, which are often stored and retrieved as a whole, thus reducing the cognitive processing efforts and making up for the deficiencies of short-term memory capacity. This illuminating idea strikes a chord in the interpreting community as interpreting scholars are eager to unveil the working mechanisms behind simultaneous interpreting during which process multiple tasks, i.e. listening and comprehension, short-term memory, and production, compete for limited attentional resources (Gile 1995/2009). In light of this view, empirical interpreting studies (e.g. Vuorikoski 2004; Eyckmans 2007; Van Rietvelde et al. 2010; Wang & Huang 2011, 2013; Wang 2012, 2016; Li & Wang 2013; Li 2016, 2017; Plevoets & Defrancq 2018) have set out to investigate the relationship between the interpreters’ performance and their use of prefabricated chunks. The results demonstrate a positive correlation between the two, which indicates that the interpreters often resort to prefabricated chunks stored holistically in their mind to allocate attentional resources, alleviate cognitive overload and improve their production fluency. Illuminating as these studies are, the majority of them are based on experimental data elicited from student interpreters. Investigations on professional interpreters in real-life settings are rather limited (e.g. Shao 2018; Aston 2018), and there is a lack of comparison as regards the use of prefabricated language between professional interpreters and native speakers. To fill this gap, the present study sets out to investigate how professional interpreters make use of prefabricated chunks to guarantee the successful completion of simultaneous interpreting in comparison with native speakers’ monolingual production. Based on two self-built corpora, i.e. a corpus of Cantonese-English simultaneous interpreting and a corpus of native English, the study attempts to describe and compare the use of three to six lexical chunks (i.e. n-grams) used by professional simultaneous interpreters and native English speakers in light of the categorization proposed by Altenberg (1998). Three hypotheses are proposed given the complex nature of simultaneous interpreting: 1) compared with native English speakers, simultaneous interpreters tend to use more but repetitive prefabricated chunks to minimize processing efforts and improve production automation and fluency; 2) compared with native English speakers, simultaneous interpreters tend to use prefabricated chunks that are both structurally and functionally simple; 3) the types of prefabricated chunks used by simultaneous interpreters are closely related to the setting as well as to the source speeches. The study reveals that the first two hypotheses are refuted while the third hypothesis is confirmed. Specifically speaking, 1) there is no significant difference in terms of the total tokens of prefabricated chunks used by simultaneous interpreters and native English speakers. However, contrary to expectation, the interpreters use more varied chunks to improve encoding efficiency and facilitate production. The author attributes it to the influence of source speeches; 2) compared with native English speakers, simultaneous interpreters use more semi-prefabricated chunks which are not syntactically and semantically fixed but are more or less conventionalized building blocks for discourse. In other words, they use more structurally and functionally varied chunks than native speakers; 3) the use of prefabricated chunks, whether in SI or in native speech production, is closely related to settings. This first-of-its-kind study hopes to shed light on the linguistic patterns of prefabricated chunks used by professional simultaneous interpreters in real-life situations, which may have some implications for interpreter training.