High and thin tones in speech and gesture: Convergence in Turkish-Swedish bilinguals' metaphors for musical pitch

Abstract

Bilingualism studies seek to elucidate the nature of the bilingual system focusing on the directionality and effects of crosslinguistic influence. The available evidence points to a non-separability of the systems shown in discussions of convergence (the increasing similarity between a bilingual’s languages), manifested as simplification, expansion or redistribution of preferences (Alferink & Gullberg, 2014). Whereas such research mostly focuses on literal language use, the present study examines bilingual use of metaphors for musical pitch. We ask whether bilinguals with (spatially) incompatible metaphors in their languages vary how they conceptualise and communicate about pitch, or display evidence of convergence. Languages of the world use various metaphors for musical pitch. Languages like Swedish may describe musical pitch as ‘high’ or ‘low’; Turkish instead describes it as ‘thin’ or ‘thick’. We investigated how Turkish-Swedish functional bilinguals living in Sweden described pitch metaphors in speech and gesture in a production task. Participants listened to pairs of sung notes differing only in pitch, and described each pair to a Swedish- or Turkish-speaking confederate performing a stimulus-matching task (language order counter-balanced). Speech was coded for metaphor used and co-occurring gestures for movement targeting height (up/down), location (high/low), and handshape for thickness (flat/curved). We hypothesised that a) participants would mainly describe pitch using language-specific metaphors; b) co-speech gestures would represent the physical dimensions invoked by speech (i.e. vertical gestures for height vs. curved gestures for size). However, since recent findings indicate that the ‘height’ metaphor is sometimes also used in Turkish, whereas the ‘thickness’ metaphor appears to be unavailable in Swedish (authors, in prep.), we also hypothesised that convergence would occur with the ‘height’ metaphor (likelier increase in bilingual Turkish), but not with the ‘thickness’ metaphor (in bilingual Swedish). The speech results show that participants generally used language-specific metaphors in each language (‘height’ and brightness’ in Swedish; ‘thickness’ and ‘height’ in Turkish). However, as predicted, the ‘height’ metaphor was significantly more frequent in Turkish trials following Swedish trials than the ‘thickness’ metaphor in Swedish trials. Swedish descriptions of pitch were mainly accompanied by gestures indicating physical height, and Turkish descriptions by size gestures. But vertical gestures also occurred with ‘thickness’ in Turkish. In such cases, speakers thus appeared to communicate two distinct mappings of space to pitch in speech and gesture. The observed patterns are indicative of unidirectional convergence or redistribution of metaphor preferences in Turkish speech and gesture under the influence of Swedish. Our findings thus support the notion of non-separability of bilinguals’ languages in non-literal language use