Dinkin (2013) proposed that /æl/ has merged to /awl/ in the Philadelphia dialect. He argued that:
- /l/ has replaced /w/ as the offglide target for /aw/.
- This facilitated the merger of /æl/ to /awl/.
I extend this argument to /uwl/ and /owl/. Using vowel trajectory data from the Philadelphia Neighborhood corpus (Labov & Rosenfelder, 2011, Rosenfelder et al, 2014), I argue that
- /uwl/ and /owl/ have different glide targets from /uw/ and /ow/.
- Their glide target is probably /l/.
This means that the the process affecting /uw, ow, aw/ can be generalized to one phonological process.
This is an interesting example of the phonological class [Vw] doing double duty:
- Defining a set of vowels to undergo a phonological process.
- Defining a set of vowels to undergo a diachronic phonetic change (Labov, Rosenfelder & Fruehwald, 2013).