CINITIAL has the strongest effect on F0 among the four factors in cases of both pooled and individual data. Such an effect indicates that there is a strong correlation between the initial consonant type and tone: H following aspirated and tense consonants and LH following other consonants. Vowel has also an effect on F0 throughout the three temporal locations but its effect is much weaker than that of CINITIAL in size (Hz). However, the effect of Length and CFINAL on F0 is found to be almost negligible. Interactions between factors are found to be due to sporadic - presumably is linguistically unimportant - effects rather than to systematic difference.
A more through phonetic study suggests that there is a strong correlation between consonant types and tone, confirming Jun's earlier observation. For the fact that consonant types correlated with tone, the phonological discussion of Jun's analysis and an alternative analysis is provided. Jun explains the consonant type-tone interaction in terms of the phonologization of laryngeal effect at an intonational level of Korean phonology. However, in Jun's analysis, it still remains clear why there is a correlation if the laryngeal effect is phonologized. Contrary to Jun, the consonant types-tone interaction is well accounted for in the alternative analysis in terms of the voice hypothesis that lax consonants are underlyingly voiced. This interpretation has potential implications for tonogenesis.
Korean Phonology by several Westerners at the Period of Enlightenment
