Segmental and tonal interactions in English and Korean: A phonetic and phonological study.

Abstract

This study is an investigation of the phonetic and phonological aspects of the relationship between segment types and tone in Korean (Seoul and Jeonnam dialects) and English. Through experiments, I addressed: (1) the differences between English and Korean in terms of segmental influences on F₀ contour; (2) the tonal patterns of Korean words; (3) Korean listeners' use of tonal information in the perception of phonation contrasts; and (4) the link between tonogenesis and Korean consonant-tone interactions. The phonetic effects of initial consonants on F₀ and the phonological behavior of these consonants led to the postulation of a tonogenesis hypothesis for Korean. The acoustic experiments analyzed the nature and magnitude of the effect of six segmental factors---initial, medial, and final consonants, vowel quality, vowel length, and rime weight---on F₀ contours for English and Korean monosyllables and Korean multi-syllables. In Korean, initial consonants had the greatest effect on F₀ contour in both spectral (Hz) and temporal (ms) magnitude. In English, the effects of initial consonants on F₀ contours were significantly smaller in magnitude than those of Korean. The Korean multi-syllabic data showed that the tonal patterns H(L) and LH(L) are predictable from the domain-initial consonant type. A perceptual experiment showed that Korean listeners systematically use the acoustic effects of initial consonants on following vowels in identifying phonation type. Investigation of the relative contribution of consonantal and vocalic information showed that the vocalic portion played a greater role than the consonantal portion in stop identification, suggesting that Korean listeners use tonal information in the perception of phonation type. The acoustic and perceptual findings are interpreted as evidence that Korean is undergoing tonogenesis: voiceless (aspirated and tense) consonants correlate with H(L) and voiced (lax, sonorant, and glottal 'ieung') consonants with LH(L). Under this account, lax consonants are underlyingly voiced and become devoiced domain-initially. The tonal differences are attributed to the shift of the feature [stiff] (=[voice]) from an initial consonant to the vowel, followed by tonal enhancement. This account captures similarities between tonal phenomena in Korean and other languages that have undergone tonogenesis. Finally, a phonological analysis of the differences between Korean and English is provided within the Optimality-Theoretic framework.