The Short-a Nasal System: how does it get so tense?

• Main Author: De Decker, Paul
• Format: Conference Object
• Language: English
• Published: 2014
• Subjects: Ferguson; Boberg; Newfoundland
• Online Access: https://research.library.mun.ca/6422/; https://research.library.mun.ca/6422/1/CLA_poster_2014.pdf; https://research.library.mun.ca/6422/3/CLA_poster_2014.pdf

This paper investigates the production of the low front vowel /æ/ which undergoes “tensing” and “raising” (Ferguson, 1972; Labov, 1989) when followed by a nasal consonant in some dialects of English (Labov, Ash and Boberg 2006). Speakers in their early twenties, and life long residents of the province of Newfoundland read a word list which included five tokens of both hand and hat. All speakers exhibited significantly lower F1 and higher F2 values for the vowel in the word hand compared to hat. As a way to explain the acoustic differences found in nasal systems, we test the model outlined by Krakow et al. (1988), that such lowering and raising is a result of co-articulation between the vowel and the following nasal consonant. We predict that: 1. If nasal co-articulation is responsible for tensing, then nasality, measured by A1-P0 (Chen 1997, Chen et al. 2007) should be higher in the nasal environment compared oral ones (i.e. elsewhere). 2. (a) Nasality and (b) its concomitant effects on F1 and F2 should be weakest early in the vowel and strongest immediately preceding the nasal consonant. This follows from Cohn's (1993) formulation of co-articulatory nasalization as a phonetically gradient rule as opposed to a categorical phonological one. All acoustic analyses were conducted using Praat (Boersma and Weenink 2013) at two temporal locations (20% and 80%) into each vowel token. Two-tailed t-tests revealed significantly lower A1-P0 values (more nasality) in the nasal environment for Speaker 1 (t (18) = 5.5011, p < 0.0001), but not for Speaker 2 (t (18) = 1.9266, p = 0.07) or Speaker 3, (t (18) = 0.2687, p = 0.79). This presents a challenge to Hypothesis (1) above suggesting that tensing might not directly related to nasalization. To test Hypothesis (2), three two-tailed t-tests were run for each speaker to examine the effect of nasality over the course of /æ/. No statistically significant differences were found in A1-P0 across the duration of the vowel for Speaker 1 (t (8) =1.008, p = 0.34). However, significant ...