[V̰] is the new [ʔ]: On glottalization in Wisconsin English.
Sarah Holmstrom & Charlotte Vanhecke
UW-Madison, Linguistics PhD program
Online – contact rashields@wisc.edu for the link.
While /t/-glottalization has been a well-attested phenomenon since the 1940s, few studies have described its specific patterns of variation and change in American English. Trager (1942) recorded [ʔ] as an allophone of /t/ in American English and proposed that where there is free variation, the glottal stop is associated with more formality. In more recent work, Eddington and Taylor (2009) attested the highest glottalization rates in young female speakers, suggesting a change in progress. The present study is the first to track real-time changes in /t/-glottalization, its stylistic implications, and its implications for theories of frequency and lenition using historical data from Wisconsin English speakers. Using historical data exposes real-time patterns of change, both in glottalization rates and in the social and stylistic associations.
The results corroborate earlier claims that glottalization is steadily spreading in American English, as the 2018 data consistently show higher glottalization rates than the 1950s WELS data across both speech styles. We provide a finer-grained analysis of glottalized /t/ which distinguishes between glottal stop [ʔ] and a further stage of lenition where a longer period of laryngealization (creak) on the preceding vowel or nasal [V̰] compensates for the absence of a discernible. Awareness of this additional stage in the lenition process can shed light on patterns of variation and change that might otherwise be obscured.
The use of real-time data in the present study allows for an investigation into frequency effects on /t/-glottalization, which is a form of lenition, over time. Previous research on frequency effects on regular sound change has established that in many cases, high frequency words change at a faster rate than low frequency words. Pierrehumbert (2001) and Bybee (2002) assert that lenition shows frequency effects, while proponents of the Constant Rate Hypothesis (Fruehwald et al. 2013, Bermúdez-Otero et al. 2015) argue that the rate of change in lenition remains the same across all contexts. The results of this study do not show significant effects of frequency on glottalization, providing support for the Constant Rate Hypothesis from real-time data.
Interpretation of speech style variation is less straightforward. Preliminary results show marked stylistic differences in glottalization rates between the reading task and conversational speech for most speakers, but the precise patterning is highly idiosyncratic. In the WELS data, slightly higher lenition rates in the reading task compared to the conversation confirm Eddington and Taylor’s intuition that glottalization carries no stigma. The numbers also suggest that glottalization signalled formality in the 1950s, echoing Trager’s (1942) findings. The general trend among younger speakers appears to be a shift within the broader glottalization category, where glottal stops regain ground over laryngealized vowels/nasals in conversational speech styles, while overall glottalization rates remain constant. The closer phonetic analysis and real-time diachronic perspective of this study improve our understanding of glottalization in American English.