Simon WEHRLE & Martine GRICE

University of Cologne


Backchannels in Map Tasks in L1 German, L1 Vietnamese &
L2 German


Backchannel behaviour serves primarily to signal a listener’s understanding of and attention to the interlocutor’s speech (Carletta et al., 1997; Clark & Schaefer, 1989), mainly in the form of short utterances such as “okay” or “mm-hm” or verbatim repetitions of all or part of the interlocutor’s previous utterance (cf. Berry, 1994; Umaporn, 2006). The function of signalling understanding is nowhere more important than in cross-cultural contexts, particularly when one or both of the conversational partners are not fully proficient in the relevant language. The frequency and type of backchannel utterances produced in such situations can in themselves have a profound influence on the successful outcome of the communicative interaction (Li, 2006; see Clancy et al., 1996 on cross-linguistic differences). The lexical or intonational choice of backchannel tokens may furthermore be interpreted very differently across cultures, having an effect on the perceived naturalness, likeability or status of the conversational partner – especially since such judgements and perceptions take place almost entirely at a subconscious level (Ha, 2012; Berry, 1994).
In this pilot study, our aim is to investigate in which way and to what extent the backchannel utterances of Vietnamese natives in L2 German differ from those of German native speakers, whilst at the same time drawing comparison to utterances from L1 Vietnamese in an effort to identify any interfering effects on the second language. We recorded twelve Map Task dialogues with a total of twelve female speakers (ages 18-34) in three groups of languages: L1 German, L1 Vietnamese and L2 German. For our analysis we counted as backchannels only such tokens that fit Carletta et al.’s ACKNOWLEDGE move, excluding e.g. any utterance that formed part of a yes-no question (Kowtko, 1997).
Looking at the non-lexical token “mm(-hm)” and the shared token “okay” (used in both German and Vietnamese), it is clear that the intonation contours of L2 German speakers show marked deviation from L1 German speakers, with clear indications of interference from L1 Vietnamese. The differences here seem to be gradual rather than categorical in nature. At a lexical level, L2 German speakers use the tokens shared across German and Vietnamese (“okay” and “mm(-hm)”) more often than L1 German speakers and conversely use specifically German backchannel tokens less often than native speakers.
The most intriguing finding concerns repetitions used as backchannels. In general, repetitions are used more in L1 Vietnamese (5%) than L1 German (2%), numbers very similar to those for Chinese and Canadian in Li (2006), but strikingly in L2 German, repetitions make up 22% of all backchannels used. In about 39% of these backchannel repetitions, not only is part of the previous utterance copied verbatim, but so too is the intonation contour. Whilst in Vietnamese L1 this is to be expected in repetitions simply because of the specification of lexical tones, in the few cases where L1 German speakers produce repetition backchannels, the contour was always clearly different, which indicates an interesting, higher-level interference from the native language onto the L2.
One additional finding concerns the repetition of backchannel words, e.g. “okayokay, jaja”, which occurred in 4-5% of cases in L1 Vietnamese and L2 German but not at all in German L1. These repetitions may signify added interest or attention in Vietnamese but are almost certain to be interpreted in a negative way, as curt or dismissive, by native German listeners. This makes all the clearer the implications of our findings for cross-cultural communication and language learning, in that inappropriate backchannel signals may create misunderstandings and potentially be detrimental to creating rapport and convergence.