Compensation for altered feedback is sensitive to speaking style

Abstract

Many speakers oppose alterations to auditory feedback, using a higher pitch when they hear their pitch lowered or a lower first formant when they hear that formant raised. But there is substantial variation in compensation within and across individuals. This case study asks whether a speaker's individual vowel space influences compensation for altered auditory feedback. A generalized linear regression model was used to investigate the influence of vowel density on trial-to-trial formant movement. First, two density maps, showing “hot spot” regions where the subject produced many vowels and “cold spot” regions where the subject produced few vowels, were constructed for (1) casual speech (from a 30-min mock interview) and (2) citation form speech (from 360 HVD words). Maps were produced for one non-phonetician speaker of California English. Afterward, the speaker participated in Five altered feedback experiments on Five separate days. The analysis showed that the subject's between-trial formant changes were not correlated with spontaneous speech vowel density but were positively correlated with citation form vowel density. This pair of results suggests that speech motor decisions are influenced by speech style, accessing citation form vowels more readily when producing citation form speech.