This study examines how several acoustic cues affect Japanese speakers' perception of geminates in three English-like nonwords, /tεk/, /tεkt/, and /tεkin/, produced by two native English speakers. A perception test showed a significant difference among the test words, /tek/ being perceived most often as a geminate, and /tekt/ least often. Of the variables examined-duration of /k/, ratio of /k/ to preceding vowel duration (C/V), ratio of /k/ to word duration (C/W), and word duration, our results correlated most highly with word duration. A second perception test varying word duration also showed a significant effect of word duration in all cases but one, a token of/tek/ which had durational properties (C/V and C/W) similar to the Japanese geminate consonant. We thus conclude that word duration is the strongest cue for the perception of gemination in English-like sound strings by Japanese speakers.
In this paper, I review previous studies on the acquisition of geminate consonants by Japanese language learners. The cross-linguistic observations reveal that the characteristics seen in the process of acquisition of Japanese geminate consonants cannot be explained neither by the Contrastive Analysis Hypothesis nor Markedness Differential Hypothesis. Finally, I claim that understanding the rhythmic organization of learners' first language and their pronunciation of Japanese as a second language, is essential to capture the reality of the acquisition of Japanese geminate consonants. The paper also discusses the methodological issues concerning durational measurements to investigate rhythmic organization of languages for future study.
The aim of this paper is to account for the phonological characteristics and prosodic function of a word-final moraic obstruent, namely the obstruent suffix /-Q/, in Japanese mimetics. Though phonological structure containing a word-final moraic obstruent is banned in the general vocabulary of Japanese, mimetics allow a moraic obstruent to appear in word-final position. More than 90% of disyllabic mimetic stems can take /-Q/ as a word-final element. This extraordinary frequency shows that word-final /-Q/ behaves as an unmarked default, which repairs ill-formed structures to satisfy certain prosodic requirements in mimetic phonology. /-Q/ plays a key role in constructing head-final prosodic structure, in which an accented trochaic foot appears in word-final position. The optimality-theoretic account in this paper shows that the head-final pattern is widely observed in and strongly required for well-formed prosody in Japanese mimetics.
The present study investigates why Korean learners of Japanese produce the Japanese voiceless stops in intervocalic position as geminates. The results of the present study demonstrate that the Korean learners matched the Japanese voiceless consonants in intervocalic position as Korean tense consonants, for which they resyllabified the preceding syllable as closed, and shortened the vowel in that syllable. This rendered the closure duration of the consonant in the following syllable enlarge up to that of the geminate. The results of the present study clearly show that the Korean learners' production of the Japanese voiceless consonants as geminates is a direct consequence of the presence of the tense consonants and an epenthetic 's' in Korean phonology, indicating that the L2 sound acquisition is greatly influenced by the first language (L1) phonological system.
This study investigated acoustic cues distinguishing between Japanese single and geminate voiceless consonants /p, t, s/ across different speech rates. Results for the two ratios, a consonant to a preceding vowel (C/V1) and a consonant to a preceding mora (C/M1), show that distributions for a singleton and a geminate are clearly distinct in the latter case only. Also, duration ratio of a consonant to its following vowel (C/V2) is found to play an important role in the distinction, results showing no overlapping between a singleton and a geminate.
This paper examines how Korean learners of Japanese (beginners) perceive a geminate stop; the effect of pitch type and a consonant in the initial syllable following the t-closure. 16 nonsense words (reto, retto meto, metto, keto, ketto, teto, tetto) which have a contrast in the voiceless and the voiced consonants of the initial syllable, as well as in the pitch type (High-Low, Low High) containing /t/ singleton or geminate contrast are presented in a carrier sentence spoken at a normal speaking rate. It was found that (1) error rate of perceiving geminate words as singleton words (t-dropping) is higher than that of perceiving singleton as geminate (t-insertion). (2) error is higher as following; voiced+LH>voiced+ HL≧voiceless+HL>voiceless+LH. (3) error rate of t-insertion is higher in LH singleton words and that of t-dropping is higher in HL geminate words.