A Study on the Variability in English Onset /l/ Production

• Published in: English Language and Linguistics Vol. 16; no. 1; pp. 225 - 245
• Main Author: 이용은
• Format: Journal Article
• Language: English
• Published: 한국영어학학회 01.04.2010
• Subjects: 영어와문학
• ISSN: 1598-9453; 2636-1183
• DOI: 10.17960/ell.2010.16.1.010

Traditionally, the English /l/ has been described as having two allophonic variants as a function of its position in a syllable, clear /l/ (as in love) and dark /l/ (as in feel). In contrast to this, Sproat and Fujimura (1993) provided compelling evidence that English /l/ actually has a continuum of possible pronunciations and that the phonetic variation in /l/ quality is determined by phonetic principles. Specifically, in their study a continuum of /l/ types in the direction from dark toward light was found as the strength of the linguistic boundary immediately following /l/ decreased. Using a naturally occurring speech, the current study investigates the possibility of whether the same kind of phonetic implementation rules could govern the pronounced forms of the syllable-initial /l/. More specifically, the possibility of interest in this study was whether the quality of onset /l/ may also vary significantly as the strength of the prosodic boundary, at the beginning of which the /l/ positioned itself, increases. To test this hypothesis, the current study made acoustic measurements of syllable-initial /l/ at the beginning of three prosodic boundaries, differing from one another in terms of the prosodic strength. The two major findings of the current study were that (i) in terms of /l/ duration, when the light /l/ was located at the beginning of a stronger prosodic boundary, its duration was only marginally longer than when it was positioned at the beginning of a prosodically weaker boundary, (ii) that in terms of /l/ quality, the ‘lightness’ of /l/ (defined as the value of ‘F2 minus F1’) was found to increase in somewhat gradual fashion as the duration of the /l/ increased. Implications of the current findings for the existing models of phonetic implementations of the English /l/ are discussed. KCI Citation Count: 0