Productions of FACE, VAGUE, LEG, DRESS, BAG, and TRAP vowels are separated by region: Northern California, Southern California, and Central Valley. Labels are placed at the mean F1 & F2 for each region.
Productions of FACE, VAGUE, LEG, DRESS, BAG, and TRAP vowels are separated by individual speaker. Labels are placed at the mean F1 & F2 for each participant
Some speakers show very distinct VAGUE and LEG
Others show moderate overlap
A few speakers, such as NorCal's 539596, show essentially complete overlap, suggesting a VAGUE-LEG merger in production in some speakers
We can calculate Pillai Scores for how much VAGUE & LEG overlap (see previous chart) and then plot these scores by region.
Marginal effect for Region predicting Pillai score, p = 0.0536
i.e. very close to statistically signifiant that NorCal has more VAGUE-LEG overlap than does SoCal or Valley
(SoCal and Valley are not significantly different from each other)
A Pearson correlation test shows that greater VAGUE-LEG overlap correlates with lesser LEG-DRESS overlap, p = 0.006, suggesting that within shifted speakers, LEG is becoming more like VAGUE and less like DRESS
A few months ago, I stumbled upon these videos of speakers from Wisconsin producing BAG and VAGUE vowels. They show how individual speakers can vary significantly in how much these vowels are raised and diphthongized, even in the same geographic location where this sound change is well established.