Phonological variation over time: Boston /æ/

Abstract

In the Boston dialect, the phoneme /æ/ has been observed to vary allophonically. To determine if this variation is systematic over time [Labov, 1966], five speakers, ages 26–71, read a list of forty words containing /æ/ before nasals (man), voiced obstruents (bad), voiceless obstruents (fact) and spirants (pass). 1st and 2nd formant values for each occurrence of /æ/ were obtained by computer analysis, and five F1 vs. F2 vowel space plots were made. For the two oldest speakers, /æ/'s are distributed in a broad (700-Hz) range along the F2 axis, and a narrow (200-Hz) range along the F1 axis. /æ/'s before nasals, with the highest F2 values, are the most fronted. For the younger speakers, /æ/'s have the same broad F2 range and narrow F1 range; but for the phonetic subclass of /æ/'s before nasals, F1 values have been lowered by about 400 Hz to overlap with F1 values for the midvowel /e/. These preliminary results suggest that, in formal speech style, phonetic discreteness or continuousness of sound change is governed by directionality of change within the vowel space: variations in tongue fronting (F2) are continuous, whereas variations in tongue raising (F0) are discrete.