We examined how the prosodic naturalness of Japanese utterances made by Chinese learners improves through the use of the visual and auditory information from OJAD (Online Japanese Accent Dictionary). The study found the following: (1) the presence of visual and auditory presentation—as opposed to the mere reading of text—led to higher evaluation scores; (2) visual and/or audiovisual presentation of the entire sentence string had greater effect; (3) audiovisual presentation was more effective for the higher level learners, and auditory presentation was more effective for the lower ones; and (4) the two kinds of information were verified to be effective regardless of the order of presentation.

In Shanghai Chinese, tone sandhi occurs in polysyllabic words and phrases. The domain of tone sandhi has been investigated in terms of not only theoretical approaches but also dialectological description since 1980s, and recently a laboratory phonological study provides auditory-data about the well-formedness of the sandhi domain. This paper summarizes previous studies of the sandhi domain of Shanghai Chinese and analyzes production data of the domain using conversational text. Although basic characteristics of attested sandhi domains could be explained by previous studies, we found some formation patterns of the domain that have not been reported previously.

In a delayed fundamental frequency (F0) fall or a late fall phenomenon, the F0 fall occurs on the post-accented mora in Japanese speech. This study conducted a large-scale investigation of the occurrence conditions of the delayed F0 fall for 230 words of 48 Tokyo-dialect Japanese speakers (21 males and 27 females). The results showed that the delayed F0 fall occurred more frequently (1) in female speech than in male speech, (2) in initial-accented words than in middle-accented words, (3) in longer words, (4) in words in which the accented mora was followed by a mora with a back vowel.

The voicing contrast for Japanese word-initial stops is primarily realized as differences in the Voice Onset Time (VOT). However, previous studies have pointed out that VOT alone cannot differentiate the two stop categories because their VOT values overlap. A few studies have attempted to find another characteristic by examining post-stop fo and voice quality but have failed to find an answer. This study investigated VOT and post-stop fo using data drawn from eighty-two native speakers from four regions. The result shows that post-stop fo differences can distinguish the two categories, but not alone. The acoustic space of VOT and post-stop fo demonstrates a clear distinction between the two categories, while the effect of VOT and post-stop fo on voicing contrast differs from region to region. The result of this study provides evidence that post-stop fo functions as an acoustic characteristic responsible for voicing contrast along with VOT.

Recent studies show that sound symbolic principles are operative in Pokémon characters’ names; e.g., those characters with names that contain more voiced obstruents tend to be larger and heavier (Kawahara et al. 2018b). One question that arose from this line of research is whether other attributes of Pokémon—specifically their types—show any tangible effects of sound symbolism. This question is related to the more general issue of what kinds of semantic attributes/dimensions can be signaled by sound symbolism. In answer to this question, Hosokawa et al. (2018) showed that the dark type characters are more likely to contain voiced stops and less likely to contain labial consonants in their names than the fairy type characters. The current judgment experiment shows that these associations are productive. Moreover, the effect sizes of sound symbolism were not correlated with each participant’s familiarity with Pokémon, suggesting that the sound symbolic knowledge is more abstract than what can be gleaned from the Pokémon lexicon.

This study examines tone in the Nyato dialect of the nDrapa language (Qiangic). First, tonal patterns and phonetic observations are described, including pitch contour pictures created with Praat. The following tones are identified based on the pitch contour patterns: (i) high level, (ii) falling, (iii) rising, and (iv) rising-falling. Regarding polysyllabic words, the initial two syllables show distinctive pitch patterns. Finally, I examine pitch alternation to analyze the tone bearing unit and conclude that a phonological word, which may consist of multiple words, bears tone in Nyato nDrapa.

Tamang is a Tibeto-Burman language spoken in Nepal and India. We report findings from a newly recorded set of data that adopts words from Mazaudon (1973, 2014) as well as Hyonjan (1993, 1997). Various dialects of Tamang have been described as a language with a four-way tonal contrast (Hari 1970, Mazaudon 1973, 1978, Mazaudon and Michaud 2008, Owen-Smith 2014, Regmi and Regmi 2018). Our results, however, find that the Tamang variety we investigated is better described via voice onset time (VOT) with a three-way laryngeal contrast.

It is well known that there is a distinction between constrictive and non-constrictive vowels in the Yi languages of the Tibeto-Burman language family. Although this vocalic distinction has been noted for a long time, its detailed process in producing, in particular, constrictive vowels has not been sufficiently researched until recent years. In this article, we will demonstrate and investigate phonetic features of constrictive/non-constrictive vowels in two Yi languages, namely Axi Yi, and Nuosu Yi, based on the outcomes of acoustic analyses of them. In conclusion, at least among these Yi languages, there are two types in vocalic constriction such as ‘pharyngealisation’ and ‘laryngealisation’.

Drenjongke is a Tibeto-Burman language spoken in Sikkim, India, whose phonetic properties are understudied. This language is reported to have a four-way laryngeal contrast: aspirated, voiceless, voiced, and “devoiced” (van Driem 2016). An acoustic analysis of twelve Drenjongke speakers shows that in addition to differences in VOT, there are systematic differences in F0 and F1 in the following vowel. Our analysis further suggests that high F1 after devoiced consonants is controlled, rather than being an automatic consequence of long VOT. We conclude that Drenjongke speakers use at least three acoustic dimensions (VOT, F0 and F1) to distinguish the four-way laryngeal contrast.

ɧ is defined as ‘simultaneous ʃ and x’ in IPA, and its existence in Swedish is widely attested. This article provides a detailed description of a [ɧ]-like sound attested in Amdo Tibetan (Tibetic, Tibeto-Burman) and defines the symbol ɧ in Amdo Tibetan as ‘simultaneous ɕ and x’. It then uses the synchronic and diachronic perspectives to present the necessity of using the symbol /ɧ/ in the phonology. The article further explores a sound correspondence between /ɧ^h/ and a Literary Tibetan sh simple initial and its typological aspect through a comparison with several Khams Tibetan dialects.

Kadu (Luish group of Tibeto-Burman spoken in Burma) can be divided into three dialects: Moteik Kadu, Molang Kadu, and Mokhwang Kadu. In Kadu languages, generally speaking, voiced obstruents are allophones of corresponding voiceless unaspirated obstruents in intervocalic positions. Thus, they are not phonemic. However, the author has discovered that in the Eastern dialect of Mokhwang Kadu, voiced obstruents have become phonemic as a result of the loss of preceding prefixes in Kadu and related languages.

Beginners of Spanish whose L1 is Japanese often wonder if Spanish /y/ is similar to Japanese /y/ or to Japanese /zy/. Three perception experiments revealed that this Spanish phoneme admits a wide variety of palatal consonants as its allophones, that these consonants can be perceived by Japanese listeners sometimes as /y/ and sometimes as /zy/, or sometimes even as /gy/. In Spanish pronunciation and listening teaching, the instructor can advise Japanese learners of Spanish that they can use either Japanese /y/ or /zy/ to pronounce Spanish /y/, and that this Spanish consonant might sound like Japanese /gy/ in some occasions.

This paper employs ultrasound and nasometric instruments in a variable-rate reading task to examine regressive nasalization in Québec French. We measure tongue height according to nasality of context, as well as percent nasality (via an energy-based ratio formula) and vowel duration. Results show high degrees of nasalization for high vowels and certain mid vowels for the group as a whole, and that certain speakers increase the nasal salience of mid and/or high vowels, while others modulate tongue height to undo displacement of the percepts of vowel height as the result of nasal coupling.

Misperception of voiced and voiceless obstruent length in Italian by native Italian and native Japanese speakers was compared. Both Italian and Japanese use consonant length contrastively. This may lead to the expectation that there is a positive first language transfer for the native Japanese speakers. While the native Japanese listeners’ misperception of Italian length categories was limited, they still made significantly more errors than the native Italian listeners. The native Japanese listeners’ length misperception was more noticeable when the word-medial obstruent was short than when it was long and when it was voiced than when it was voiceless.