A network-modeling approach to investigating individual differences in articulatory-to-acoustic relationship strategies

This study represents an exploratory analysis of a novel method of investigating variation among individual speakers with respect to the articulatory strategies used to modify acoustic characteristics of their speech. Articulatory data (nasalization, tongue height, breathiness) and acoustic data (F1 frequency) related to the distinction of three nasal-oral vowel contrasts in French were co-registered. Data were collected first from four Southern French (FR) speakers and, subsequently, from nine naïve Australian English listeners who imitated the FR productions. Articulatory measurements were mapped to F1 measurements using relative importance analysis (RIA), and the RIA coefficients were used to create similarity scores among all of the speakers. The similarity scores were then used to build network models for each nasal-oral vowel pair, using the spinglass algorithm to identify communities of shared articulatory-to-acoustic strategies within each network. The results show that network grouping is rarely based on language-dependent articulatory-to-acoustic strategies, but evidence of inter- and intra-speaker consistency is observed: Individual speakers tend to group together in their articulatory-to-acoustic strategies across vowel pairs, and most speakers have consistent articulatory-to-acoustic mappings across vowel pairs. Evidence is also observed which highlights the multi-dimensional nature of vowel nasality, rather than the uni-dimensional assumption of “nasal” vowels as merely oral vowels produced with a lowered velum.