Referential iconicity in music and speech within and across sensory modalities

Abstract: Musical meaning is multifaceted: both highly sensory and yet often abstract, able to cross cultural boundaries and yet embedded in specific traditions. For the most part it is not denotational (Monelle, 1991). Nevertheless, in “programmatic music”, musical themes are intended to refer to worldly objects and events on the basis of iconic (and indexical) grounds. Such non-arbitrariness of the sound-sign (Sonesson, 2013) appears to apply to speech as well, where research has established that the iconicity in question is subtle, but systematic enough to be detectible by both adults and children (Ahlner and Zlatev, 2010; Imai and Kita, 2014). Very often, it operates across sensory modalities, so that for example a sound form like lulu is linked to round shapes, while titti is associated with sharp and hard ones. This thesis investigates how referential iconicity in speech operates in relation to music, taking into account different kinds of iconicity, unimodality and cross-modality and finally cultural background. To address these aspects, an experiment in which 21 Swedish and 21 Chinese native speakers had to match musical fragments or spoken word-forms to referents (represented by schematic pictures) was designed. It included two different conditions. In one there were two sound-stimuli and two referents (more contrastive). In the other, a single sound-stimulus was to be matched to one of four alternative referents (less contrastive). The results showed that there was no significant difference between the overall results for music and linguistic tasks, indicating that the psychological, interpretive processes involved are not limited to a single cognitive domain, or semiotic system. As expected, the more contrastive condition was easier for both groups, showing that cultural background played little role for making the appropriate cross-modal mapping when the choice was so constrained. Finally, the fact that participants performed significantly better in morecontrastive tasks than less-contrastive, whilst performing above significant chance in both conditions serves as a clear indicator that interpreting referential music in music and speech sounds involves a combination of primary and secondary iconicity (Sonesson 1997), with a considerable role for the latter.