Universal constraints on the sound structure of language: Beyond articulation

Abstract

Across languages, certain onset structures (e.g., blif) are systematically preferred to others (e.g., lbif). Moreover, similar preferences are evident in the behavior of individual speakers even when none of these structures exists in their language. For instance, English speakers exhibit the preference (indicated by >) bnif>bdif>lbif despite no experience with either syllable type (Berent et al., 2007) and similar findings were obtained in Spanish (Berent, Lennertz & Rosselli, 2011), Korean (Berent et al., 2008) and Chinese (Zhao & Berent, 2013). æThis converge suggests the possibility of universal grammatical constraints on language structures. But on an alternative account, these preferences might be only due to auditory and articulatory factors. Our past research has challenged the auditory explanation by replicating these preferences with printed materials (Berent & Lennertz, 2010). However, these results leave open the possibility that peopleÍs preferences might be informed by subvocal articulation of these stimuli. In this view, bnif is preferred to lbif because the former is easier to articulate. Our present experiments address this possibility. æTo this end, we gauge peopleÍs linguistic preferences while suppressing their articulation (by asking them to bite on tongue depressors). The results from three experiments demonstrate either full or nearly full sensitivity to linguistic structure, and only minimal effects of suppression. These findings suggest that people possess broad phonological preferences that are inexplicable by articulatory factors.