The purpose of this paper is to examine the ability of Japanese native speakers to identify the Tokyo accent, which is standard Japanese. A total of 126 university students from four areas, where their accents have different systems, participated in a perception test eight times over the course of one month. The data collected were analyzed in terms of the number of tests, intonation types, accent patterns, number of morae, type of morae, and type of special morae. Error analysis was also conducted. The rate of correct answers got better with each test regardless of accent system and reached 87% in the last test for listeners of Tokyo-shiki (Tokyo type) accent, 83% for listeners of no accent, 80% for listeners of Keihan-shiki (Kyoto-Osaka type) accent, and 69% for listeners of Nikei-shiki (two patterns) accent. The findings will help develop tools to improve listening skills for the Tokyo accent.
This study examined the effects of phonological factors on the perception of English syllables and lexical stress by native Japanese speakers. Factors investigated were syllable complexity, position of the consonant cluster, number of syllables, stress position, vowel structure, silent words, and syllabic consonants. The Syllable Count Task (SCT) and the Stress Identification Task (SIT) were conducted. In both tasks, performance declined as the number of syllables increased. In addition, for one-syllable words, as the complexity of the syllable structure increased, SCT performance decreased. Furthermore, as the stress position came closer to the coda position of the word, SIT performance declined.
A perception experiment was conducted to determine how accurately Chinese learners of Japanese can identify four emotions (Happy, Angry, Sad and Neutral) in Japanese emotional speech. The result showed that the accuracy of recognition was worst in “Happy”, followed by “Anger”, “Sad”, and “Neutral”, and was influenced by their learning experience. Acoustic analysis revealed that there were no significant differences between (1) “Happy” and “Angry” in F0mean, F0level, F0max, F0span, Duration and IntensityMean, and (2) “Neutral” and “Sad” in F0level, F0max, F0span, Duration and IntensityMean, thus potentially explaining why these two emotions were especially confused.
This is a report on an examination and evaluation of Japanese learners’ English using multiple metrics: phoneme substitution, epenthesis, elision, primary stress, rhythm and overall goodness. In the examination, the difficulty in producing weak/lax vowels was highlighted. The correlation analysis between overall goodness and each of the other metrics indicated that improving both segmental and prosodic features is essential for Japanese learners to achieve good pronunciation. A systematic pronunciation evaluation using multiple metrics is beneficial not only for a deeper and broader understanding of Japanese English but also to the development of pronunciation teaching and learning.
The objective of this study was to examine whether vowel duration and lexical decision time differ between CnV, CnVC (voiced), and CnVC (voiceless) words. After presenting monosyllabic English and nonsense words with onset consonant clusters to Japanese learners of English and native English speakers, participants were asked to judge whether the test words were English words or nonsense words and pronounced the words. Their response time was measured, and their pronunciation was recorded to compare vowel durations. Their CVC test words had shorter vowels than CV words, and vowels preceding unvoiced consonants were shorter than vowels preceding voiced consonants. These effects were more conspicuous for native English speakers, though the same effects were detected for the Japanese speakers. Comparison of lexical decision time showed native English speakers responded quicker than Japanese speakers, and that nonsense words were more difficult to judge than English words for both English and Japanese speakers. No significantly high correlations were detected between learners’ English proficiency levels and their vowel durations or lexical decision time.
The study examines the phonetic characteristics of voicing contrasts of initial stops in languages. The examination is based on physiological and acoustic data published in the past and newly analyzed for six languages. Acoustic and phonetic features such as VOT, onset F0, F0 contour, burst spectrum, and the tongue contact areas are functional in differentiating homorganic stops in many languages. VOT which is related to laryngeal timing to oral release seems to be a primary cue, but is not functional for stops in Burmese. These cues appear to play a major role in distinguishing the contrasts, together with other features.
A growing body of cross-linguistic research has shown that speakers advance the tongue root during the production of voiced stops to meet aerodynamic requirements for voicing. The current study investigates the tongue root positioning during the production of voiced and voiceless stops in Russian, which has contrastive palatalization. Ultrasound tongue imaging data collected from five native Russian speakers revealed that the tongue root position is affected more by palatalization than by voicing, suggesting that (i) articulatory strategies of accommodating for the aerodynamic constraint are language-specific; (ii) the (dis) preference for particular strategies may be constrained by the language-specific use of articulatory space, such as the presence of palatalization.
The present study investigates the acoustic and articulatory characteristics of voiced plosive geminates produced by a male speaker of Kumamoto dialect in Kyushu District. Two types of Magnetic Resonance Images (MRI), real-time MRI and synchronized MRI, are used for articulatory analyses. The acoustic result showed that relatively long voicing period appeared during the closure of voiced geminates. Plosive /dd/ showed voicing continuing from a preceding vowel (post-voicing) and pre-voicing of the consonant, while /gg/ showed only post-voicing. Analyses of MRI revealed that the distance between lower tongue and pharyngeal wall extended towards the end of the geminate plosives suggesting the occurrence of pharyngeal expansion. This agrees to the result of the speakers of Ikema southern Ryukyuan whose voiced geminates were voiced throughout the consonants. Thus, pharyngeal expansion is regarded to be an major articulatory manifestation to keep and excite voicing during voiced geminates.
This paper reports the phonetic details of geminates in the Kumamoto dialect of Japanese, especially from the perspectives of generation and gender. Data from 105 native speakers were analyzed, and it was demonstrated that generation and gender in voiced geminates have a clear relationship with (a) the proportion of voicing patterns of the closures and (b) the duration of continuous voicing from a preceding vowel. If a generational tendency reflects the diachronic change of voiced geminates, it indicates that although Kumamoto dialect has full-voiced pronunciation with phonologically voiced geminates in traditional pronunciation, it is changing in the younger generation.
Recent studies on sound symbolism have demonstrated that the number of voiced obstruents in Japanese Pokémon names is positively correlated with their evolution levels. This correlation is likely to be rooted in the sound symbolic relationship between voiced obstruents and largeness/heaviness/strength. This study shows that when Japanese children are presented with two non-existing names and a pair of pre-evolution and post-evolution Pokémon characters, they are more likely to associate names having voiced obstruents with post-evolution Pokémon characters. The experiment also shows that Japanese children associate post-evolution characters more with [a] than with [i], which shows that they are also sensitive to vocalic sound symbolism.
Khalkha Mongolian has two types of obstruents, which are transliterated as 〈b, d, g, dz, ǰ〉 and 〈p, t, k, ts, č〉, respectively; however, it is not clear whether the feature distinguishing them is voicing or aspiration. This production study examines the distribution of the VOT values between the two series of word-initial obstruents in Khalkha Mongolian. The data show that VOTs of 〈b, d, g, dz, ǰ〉 generally show positive values and that there is no phonetic ground to view /g/ and /ɢ/ as voiced consonants, although they are phonologically regarded as voiced.
This paper reports that phonetic realization of voiced geminates in the Murayama dialect of Yamagata Japanese. Whether closure voicing ratio is long or not in voiced geminates in Tohoku dialects is controversial. This paper analyzes the data which were taken from elicitation from four native speakers, and demonstrates that both long and short closure voicing ratio exist in the dialect. Moreover, the result demonstrates that closure voicing ratio of underlying voiced geminate is longer than that of derived one for one speaker. Our results imply the necessarity that more comprehensive research on voicing in Tohoku dialects is needed.