Abstract

The assumption stands in acoustic phonetics that vowel reduction (including undershoot and centralization of short vowels) is greater in casual than controlled speech. However, few studies have investigated the relative magnitude of language‐general (phonetic) effects relative to dialect‐specific (sociolinguistic) ones. This paper investigates reduction in vowels undergoing sound‐change. Pacific Northwestern English (PNWE) /æ/, /ε/ are rising to the spectral location of /ey/, a “merger by approximation” [W. Labov, Princ. Ling. Change, Blackwell (1994)]. Sociolinguistic literature guided selection of two stable/changing pairs (in which one member is stable, the other involved in the merger): /iy ey/, /ɪε/. The normalized corpus (n=1500) was balanced for the following place and voicing and included three conditions (wordlist, reading, and casual). Speakers were partitioned into groups (merged/unmerged). Wassink’s [J. Acoust. Soc. Am. 119, 2334–2350 (2006)] spectral overlap assessment metric was used to detect differences between volumes of vowel ellipsoids (in F1×F2×duration space) and test for style‐related spectral shift. Euclidean distances (measured between 20% and 80% timepoints) differentiated vowel trajectories recorded in different conditions. Merged and unmerged speakers’ reduction strategies differed. Merged: expected reduction patterns in both short vowels (including raised /ε/). /y/ trajectories show greater gliding in casual speech. Unmerged: /ε/ does not shift position or trajectory as task changes. [Research supported by National Science Foundation BCS#‐0643374.]

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call