An experiment on the duration of English sounds was conducted using small to large units as a corpus: CVC-structured monosyllabic words, disyllabic words in which the stress pattern differentiates between nouns and verbs, phrases and compounds spelled in the same way, and a series of sentences in which the number of function words increases gradually. In the pronunciation of the English subjects, the well-known phenomenon of temporal compensation was corroborated between the vowel and the following voiced/voiceless consonant but not between the long/short vowel and the following consonant in monosyllabic words, or between any segment in the disyllabic words cited above. English pronounced by the Japanese subjects, on the other hand, showed phonetic features of Japanese such as pitch accent and mora-timing.
View full abstract