Sound symbolism is modulated by linguistic experience

The bouba-kiki effect, or the symbolic association between certain speech sounds and rounded or angular shapes, is widely thought to be universal. However, two studies have failed to replicate this effect with neurotypical participants in the classical paradigm, one conducted in Papua New Guinea (Ross & Rogers, 1975), and the other conducted in Nepal (Styles & Gawne, 2017). As both experiments employed auditory stimuli inconsistent with the sound structure of the respective native language, Styles and Gawne (2017) proposed that pseudoword legality is a prerequisite for sound-symbolic associations to form. In this study, we conducted the first experimental test of this hypothesis, by assessing participants’ performance on the bouba-kiki task as a function of pseudoword phonotactic legality. Our results indicate that phonotactic violations may disrupt the bouba-kiki effect, albeit only when they cause the speech stimuli to be perceived as significantly strange (not “word-like”). We thus conclude that sound symbolism fails whenever phonotactic violations prevent the assemblance of the phonological representations of the target pseudowords.