|
KMID : 0917520000070030021
|
|
Journal of Speech Sciences
2000 Volume.7 No. 3 p.21 ~ p.33
|
|
|
Voicing in intervocalic lax obstruents /p,t,k,c/ of Korean
|
|
Yun Il-Sung
|
|
|
Abstract
|
|
|
There are two hypotheses with reference to voicing in Korean intervocalic lax stops /p, t, k/ and affricate /c/: (1) the phonologically voiceless lax stops /p, t, k/ and affricate /c/ are realised as voiced allophones in the intervocalic position; (2) the shorter the lax consonant, the higher the percentage of voicing. But the literature reveals that there are views rejecting or doubting them. to clarify these, an experiment was carried out, using a Sun Sparcstation, twelve native speakers of Korean and speech materials embedded in sentence frame. The results showed that the extent of voicing in lax stops and affricate was too inconsistent to support the full voicing hypothesis, and shorter duration (faster speech) did not necessarily cause a higher percentage of voicing.
Kewywords: voicing, itervocalic lax obstruents, speech rate, duration
This study was aimed at examining two controversial hypotheses: (1) the phonologically voiceless Korean lax stops /p, t, k/ and affricate /c/ are realised as voiced allophones in the intervocalic position (2) the faster the speech, the more likely a lax stop is to be voiced (i.e. the closure duration of a lax stop is negatively correlated with the percentage of voicing relative to closure duration (PCT)). First, the results of the experiment revealed that full voicing did not always appear in Korean intervocalic lax stops /p, t, k/ and affricate /c/, while substantial devoicing and marked standard deviations of PCT were observed. There was great variability between/within speakers and between phonemes (i.e. the place and manner of articulation). Accordingly the first hypothesis is rejected the conventional full voicing hypothesis should be modified. Second, shorter consonant duration (faster rate of speech) did not necessarily induce higher PCT. That is, the correlations between consonant duration and VCT or PCT, in general, were not significantly high for the pooled data (across all subjects). In addition, the correlations individual speakers yielded between consonant duration and PCT had a great variation ranging from significantly negative correlations to significantly positive correlations. There was great variability between and within speakers, and between phonemes (i.e. the place and manner of articulation). Similar observations concerning tempo effect on voicing of lax obstruents have been reported in English (Kim, 1989; Smith, 1997). In consequence, the second hypothesis the shorter the lax consonant, the higher the percentage of voicing is also rejected.
|
|
|
KEYWORD
|
|
|
|
|
|
FullTexts / Linksout information
|
|
|
|
|
|
Listed journal information
|
|
|
|
|