A method for analyzing the coarticulated CV and VC components of vowel-formant trajectories in CVC syllables

• This paper describes how vowel formants are affected by consonantal contexts.
• It presents a new method for quantifying these coarticulatory effects.
• The method is based on an additive model with linear-scaling properties.
• Contextual effects can now be observed as they change over the duration of a vowel.

To better understand the dynamic structure of vowels in CVC′ contexts one must account for the temporally-overlapping effects of the initial and final consonants C and C′. Here we present a linear-decomposition (LD) method for analyzing these effects as perturbations of the vowel-formant trajectories from their targets. The perturbations are modeled by the superposition of their CV and VC′ components, which are scaled by the differences between the vowel targets and their respective consonant loci. We use a dataset of second-formant frequencies (F2) from bVd, dVd, and gVd syllables containing seven vowels to illustrate how to estimate each element of the model by taking advantage of its additive structure and scaling properties. The model represents a family of formant trajectories unified by its scaling relationships, and the LD method that follows from it reveals how contextual effects combine and change over the duration of a vowel. Formulas for implementing the method are presented in appendices along with the two speakers' F2 datasets employed in this study.