Search

The paper examines acoustic timing patterns of word-medial voiceless and voiced singleton and geminate plosives of Yanagawa Japanese, one of the Chikugo varieties of Japanese spoken in the center of Kyushu (Japan). Unlike standard Tokyo Japanese, Yanagawa Japanese is characterized by frequent gemination of any types of consonants including voiced obstruents (e.g., /kuzzoko/ [kuddzoko] “sole, the fish,” /miroɡɡe/ [miɾoɡɡe] “a shell fish”). Five Yanagawa Japanese speakers were recorded. Another five speakers of standard Tokyo Japanese were also recorded as a control group. The stimuli consisted of nonsense words with C1V1C(C)2V2 structure (e.g., /kak(k)a/, /kaɡ(ɡ)a/). The findings suggest that the whole word duration containing geminate consonants was longer in proportion to mora count difference those containing singleton consonants, confirming moraic timing at word level in Yanagawa Japanese. When C2 was a geminate (voiceless or voiced), it lengthened the preceding vowel (V1) and shortened the following one (V2) in both dialects. However, the magnitude of influence of geminates to adjacent vowels was dialect specific: V1 duration in Yanagawa was consistently longer than Tokyo when the following consonant was a geminate. The results of the current study point towards the timing differences in singletons and geminates across Japanese dialects.

This study explores the phenomenon of language attrition. Specifically, we investigate the phonetic properties of consonant gemination across three groups of speakers of Palestinian Arabic: monolinguals (i.e., native speakers born in Palestine who have lived there their entire life, n = 5), late bilinguals (i.e., speakers born in Palestine who emigrated to the US during their teens, n = 6), and heritage speakers (i.e., speakers of Palestinian descent, born in the US and who speak both English and Arabic in their daily lives, n = 7). All speakers were in their mid-1920s. The participants were tested using a delayed word repetition task. The stimuli comprised 158 bi-syllabic Arabic minimal and near-minimal pairs (e.g., /ħam:aːm/ “bathroom” versus /ħama:m/ “pigeon”) including long and short stops, fricatives, and sonorants. We controlled for stress and syllabic position. Distractors were also included. The acoustic analysis is underway, and consists of manually aligning the target consonants, extracting the mean consonant duration and comparing it across groups. Additional measures include voicing, aspiration, and formant transitions. The findings will enable us to address the question whether universal phonetic factors (from the perspective of Markedness Theory) have an effect on degree of attrition by specifically comparing consonants from different voicing categories and manners of articulation.

Moroccan Arabic uses geminate/singleton contrasts in medial position but it is controversial whether it maintains them utterance-initially. To address this issue, we made simultaneous ultrasound and acoustic recordings of five native speakers producing target words containing /t/-/tt/ and /d/-/dd/ contrasts utterance-initially and -medially, 10 times each. The ultrasound data were analysed via a novel method of using pixel-derived principal components and linear discrimination to generate a time-varying articulatory “closure” signal directly from the ultrasound images, without the need for manual tracing. This articulatory signal was subsequently used to measure closure duration by determining gestural landmarks from its velocity function. The results provide clear articulatory evidence that speakers produce utterance-initial singletons versus geminates with significantly different closure durations, although the contrast provides no acoustic evidence of closure duration per se. Rather, our results reveal that speakers realize geminate/singleton contrasts via acoustic dimensions not related to consonant duration: The vowel is significantly longer following geminates than singletons, and the stop bursts have greater amplitude for geminates than singletons. Together these findings provide evidence that the gemination contrast is maintained in initial position in Moroccan Arabic, and challenge traditional assumptions that consonant length contrasts are primarily carried by acoustic closure duration differences.

Liquid geminates are uncommon and disfavored in Japanese native phonology. However, instances of emphatic expressions and loanwords from languages with liquid geminates such as Italian or Arabic suggest that they are not impossible. The current study examines the durational, acoustic, and articulatory properties of geminated liquids in Japanese to obtain further insights into the nature of /r/ in Japanese and the process of gemination in general. We report on a production experiment whereby eight native speakers of Japanese pronounced mimetics of the form CVCVCVCV with and without emphatic gemination (e.g., kirakira > kir:akira “shiny”). In addition to the audio, tongue movements were recorded using the EMA technology. Preliminary durational analysis suggests that singleton-geminate ratio is about 1:3, which is slightly larger than the ratios previously reported for geminated obstruents. We explore what articulatory strategies speakers employ to lengthen the liquid consonant whose prototypical singleton production is [ɾ].

Mandarin speakers tend to adapt an intervocalic nasal as either an onset of the following syllable (e.g., Brunoa bu.lŭ.nuo) or as a nasal geminate (e.g., Daniel a dān.ni.ěr) (Huang & Lin 2013, 2014). Two forced-choice identification experiments were conducted to test the effects of nasalization (whether the pre-nasal vowel bears stress or not) and duration (whether the pre-nasal vowel is lax or tense): Would stronger nasalization and shorter duration encourage higher nasal gemination rate? The results showed that Mandarin speakers’ choice of repairs was indeed biased by the different phonetic manipulations, suggesting an effect of perceptual similarity. Moreover, the overall preference for the V.NV form over the VN.NV form suggests an influence from the native syllable type frequency (open syllables being more frequent than closed syllables). The across-the-board higher VN.NV responses for lax than for tense vowels regardless of the phonetic manipulations are attributed to the possibility that Mandarin speakers might have perceived the tense vowels as diphthongs (i.e., English /e/ to [ej], /o/ to [ow]) and inserting a nasal coda is illegal in this contexts (*CVGN). That is, the findings suggest that the variations in loanword adaptation were guided by perception, frequency, as well as phonotactics.

Gemination is largely assumed to be cued through duration. While production and perception studies consistently confirm the importance of duration as a cue, certain phones are not as easily distinguished using duration alone. The current study analyzes geminate and singleton minimal pair production and perception in Ikema, an endangered Japonic language, to determine to what extent other cues may contribute to this distinction. Productions of stop, fricative, affricate, nasal, and rhotic geminates and singletons were used to determine acoustic differences and possible cues to gemination. We analyzed duration for all phones, and spectral moment and vocalic measures for fricatives and affricates. Results indicate that while duration is an important cue, spectral moment measures and vocalic measures are better predictors of affricate geminates and contribute significantly to the acoustic distinction among fricatives. Word-initial stops may be difficult to identify as closure duration is a more significant predictor of gemination than voice onset time. Listener perception was analyzed using a lexical decision task. Results indicate that while Ikema speakers distinguish geminates from singletons in novel situations, non-word perception may not be consistent cross-linguistically due to varying exposure to the concept, and community-internal acceptability of pronunciation variation.

This study examines how plural noun reduplication in Tohono O’odham (TO) language is phonetically realized. To make a noun plural, the first CV sequence (i.e., base; e.g., goks “a dog” → gogoks “dogs”) or the first consonant of the base (e.g., pado “a pig” → papdo “pigs”) are reduplicated after the base (Hill & Zepeda, 1998). Base and reduplicant have been regarded phonologically equivalent (i.e., consonant gemination; Fitzgerald, 2003). In a sense that TO has the strong-weak stress pattern across syllables (e.g., towa “turkey”), stress may influence reduplicants' duration. The CV reduplicant at the second syllable with the weak stress may not be phonologically same as the base and the C reduplicant at the coda of the base with the strong stress. If CV and C reduplicants are equivalent even with stress, an additional prosodic unit (i.e., mora) may exist. The analysis of TO speech supports that the base (CV1) is longer than the reduplicant (CV2 or C2) even with other contextual variables. Reduplication process may be stress-timed, but it is unknown whether speakers equally treat base and reduplicants. This paper will also discuss a need to examine how TO speakers control the timing with rhythm between base and reduplicant.This study examines how plural noun reduplication in Tohono O’odham (TO) language is phonetically realized. To make a noun plural, the first CV sequence (i.e., base; e.g., goks “a dog” → gogoks “dogs”) or the first consonant of the base (e.g., pado “a pig” → papdo “pigs”) are reduplicated after the base (Hill & Zepeda, 1998). Base and reduplicant have been regarded phonologically equivalent (i.e., consonant gemination; Fitzgerald, 2003). In a sense that TO has the strong-weak stress pattern across syllables (e.g., towa “turkey”), stress may influence reduplicants' duration. The CV reduplicant at the second syllable with the weak stress may not be phonologically same as the base and the C reduplicant at the coda of the base with the strong stress. If CV and C reduplicants are equivalent even with stress, an additional prosodic unit (i.e., mora) may exist. The analysis of TO speech supports that the base (CV1) is longer than the reduplicant (CV2 or C2) even with other contextual variables. Reduplication pro...

Lamkang is an under-documented endangered Kuki-Chin language of the Northeast India spoken in 40 villages spread across Manipur’s Chandel district with under 10,000 Lamkang-Naga speakers (Ethnologue 2014). In Lamkang, doubled consonants have three different sources: (1) gemination when stems are followed by enclitics of the shape V; (2) total assimilation of enclitics to stem codas; (3) ambisyllables in prefixes of the shape CV (where V is an epenthetic vowel). We examine the acoustic properties of the doubled consonants in these three morphological environments to show that the length of the closures reflect the morphological derivational history. The paper investigates a method for measuring doubled consonant length: We measure from peak intensity of vowels on either side of the consonant closure, rather than the length of the closure. Orthographic practice reflects these differences as doubled consonants in the prefixes are shorter than the phonologically derived consonants with enclitics. Copied conso...

The liquid phoneme in Korean is known to have an optional retroflex allophone in syllable-final positions (Lee 1999). However, there lacks literature evaluating when and how this retroflex allophone is realized. To define the distribution of the retroflex allophone of the liquid phoneme in Korean, this study uses a read speech corpus of Seoul Korean to track the rhotacization of the liquid phoneme in coda positions. Rhotacization of the liquid is detected and measured by the lowering of the third formant. Preliminary results show that the liquid phoneme in coda position is realized as retroflex when preceded by a back vowel, and when followed by another liquid phoneme (gemination). The results suggest that the liquid phoneme is not simply optionally rhotacized in coda positions; instead, the retroflex allophone is realized in contexts defined by the preceding vowel and the following consonant. Future research remains to further examine the articulatory and perceptual aspects of the rhotacization.

Quantity contrasts (i.e., singleton vs geminate consonants, short vs long vowels) are common in the languages of the world. Yet, most singleton‐geminate contrasts occur in word‐medial position. Languages that allow gemination in other than word‐medial position (word‐initially or word‐finally) are quite few. Even less frequent is the occurrence in the languages of the world of word‐initial voiceless geminate stops. Pattani Malay (Abramson, 1986), Turgovian Swiss German (Kraehenmann, 2003), Tashlhiyt Berber (Ridouane, 2007), are the only documented languages that allow singleton‐geminate contrasts word‐initially for all native consonants, including both voiced and voiceless plosives. In this work we deal with word‐initial voiceless geminate stops in Tashlhiyt Berber and present results from 3 experiments: 1) a production study including acoustic and electropalatographic measurements 2) a native‐listener experiment with labelling and discrimination, and 3) a cross‐language perceptual experiment, with listeners of two languages differing in germination: Italian (in which gemination only occurs word‐medially) and French (in which gemination only occurs between words in continuous speech).

This research provides evidence toward the existence of geminates in Cypriot Greek (hereinafter, CyG). In the existing literature it is reported that the sole or primary cue to gemination, across languages, is closure duration. This phonetic study verifies that this cross‐linguistic phonetic cue is also present in CyG. Toward this end, statistical analyses evidence highly significant durational differences in the voice onset time (VOT) of geminates versus singletons in CyG [F(164)=p

This study investigates the nature of the acoustic variation in sequences of identical affricates produced by Polish learners of English. In both English and Polish sequences of identical affricates occur across word boundaries, but only in Polish do such sequences also occur root internally and across morpheme boundaries. In Polish sequences of identical affricates are manifested variably both by rearticulation of both affricates and by articulation of a single affricate but with lengthened duration of either the stop or the fricative. To investigate their English, the subjects performed two tasks: repetition of 12 English sentences and orally responding to 17 multiple choice questions. The task produced significant cross‐speaker differences in the phonetics of the geminates, differences correlated with differences in their proficiency levels in English. The more Polish‐like singly articulated long affricates were produced by 22% of the intermediate speakers but by 48% of the advanced speakers, the opposite of what one might expect. The intermediate speakers appear to have paid more attention to the phonetics of the English cues, thus producing more fully rearticulated affricates; the more advanced speakers appear to have paid less attention to the phonetics of the cues, thus reverting more to the norms of Polish pronunciation.

Acoustic analysis of Italian [r] and [l] was carried out. Speech materials consisting of prestressed VCV syllables produced by 3 male and 3 female native speakers of Italian formed the basis of the present study. Each syllable was repeated 3 times by each speaker. The vowel in the syllable was [i,a,u] and the consonant was [r] or [l] in geminated and nongeminated form, leading to 216 syllables. Confirming the results of a previous study on Italian stop consonants [R. Rossetti, 2874 (1994)] results of the present analysis showed that the first vowel duration was systematically shortened in the geminated consonants. The data also showed that in the present case the duration of the consonant as well as, in most cases, of the VC and CV transitions were lengthened in the geminated form. Spectral analysis showed that there is little or no effect on the spectral properties of the consonants due to the presence of gemination. However, [r] in geminated form is in a few cases devoiced. Informal perceptual tests indicate that [r] can be automatically generated from [l] by introducing a silent portion within the stationary portion of [l].

An acoustic study of geminated and nongerminated Italian stops was carried out. The analysis was performed on 648 VCV syllables, where V is one of the vowels [i, a, u] and C is one of the consonants [p, t, k, b, d, g] (in geminated and nongeminated form). The syllables were uttered by 6 speakers, 3 repetitions. The analysis of the segments durations corresponding to the different articulatory phases of the VCV structure showed a strong correlation between the presence of gemination and the first vowel and occlusive silence durations. Statistical analyses using an a posteriori classifier yielded to a classification score of 96.1%. An analysis of the properties of the burst (energy, power, and spectral shape) showed that there is no correlation between these and the presence of gemination. An automatic software system, which allows the modification of a single/geminated VCV utterance into a geminated/single one, was developed. Informal perceptual tests, in which the stimuli were obtained by using the previous program, validated the significance of the first vowel and the occlusion durations as acoustic correlates of gemination.

This study analyzes duration of stop syllables in Italian and Japanese. Its results show that the two languages are very similar in terms of syllable length. They also indicate that, in accordance with previous findings, both Italian and Japanese syllable duration is contrastive in terms of gemination, and Italian duration is also contrastive as a function of stress. Finally, they show that both languages present variability in the duration of their syllables, even in cases where the syllable is neither geminate nor stressed. These results point to the following conclusions: (1) other elements outside the syllable itself (i.e., position within word, surrounding syllables, information content, etc.) may be involved in determining syllable duration; and (2) the stress‐timed/syllable‐timed language distinction may not be useful for understanding how languages assign duration for particular parts of the syllable. In particular, the results of this study indicate that stress and syllable position play an important role in determining syllable duration even in syllable‐timed languages.

This study investigated the acoustic characteristics of coarticulatory phenomena in connected Japanese digits. Its aim was to discover acoustic phonetic regularities that could be used in a knowledge‐based continuous speech recognition system. The data consisted of 840 six‐digit strings spoken by seven males and three female native speakers of Japanese. A variety of intra‐ and inter‐digit coarticulatory phenonema were examined using synchronized displays of digitized waveforms, wideband spectrograms, total energy, low‐frequency energy, and zero crossings rate. Among the questions assessed were: whether allophonic processes often cited in the phonological literature are readily observable in the acoustic signal, whether they are consistently observed across tokens and speakers, and whether seemingly free variants occur truly randomly or are conditioned by such factors as phrasal structure, accentual pattern, and speaking rate. The allophonic processes examined included: devoicing/shortening of high vowels, gemination of voiceless stops, deletion/lenition of word‐medial sonorants, glottalization, lenition of [h], vowel nasalization, frication/nasalization of [g], affrication of [z], etc. Current affiliation: Voice Processing Corporation, 1 Broadway, Cambridge, MA 02142.

The duration of phonetic intervals are sometimes longer in the speech of young children than adults, although it is not clear at what age timing can be said to have become “adult‐like.” This study examined speech timing as a function of speaker age. Adults and three groups of children (mean ages: 5;0, 8;0, 10;8) produced utterances like “Bob Bob Bob” with a tube in the corner of the mouth. The duration of “vowels” (/a/, /I/, /i/) and singleton and abutting stops (/b/,/b♯b/,/b♯p/) was defined by means of variation in supraglottal pressure. There were few between‐group differences in the absolute duration of vowels, and none for bilabial closure intervals. All groups showed the same relative effects of phonetic factors. For vowels, the effect of position‐in‐utterance, number of syllables per word, vowel height, and stop voicing were comparable across groups. Position‐in‐utterance, vocalic environment, gemination (i.e., /b/ versus /b♯b/), and position with regard to stress also showed similar effects on consonant duration for each group.

Velum movements have been studied by tracking a metal pellet placed on the velum with computer‐controlled radiography. The tongue gesture was observed simultaneously by tracking two additional pellets on the posterior part of the tongue. A set of utterances consisting of Japanese words in isolation contained syllable initial and syllable final nasal consonants in comparison with non‐nasal segments, i.e., vowel elongation and consonant gemination. It is shown that (1) the syllable final (moraic) nasal in Japanese shows a velum lowering to a considerably larger extent than the syllable initial nasal, (2) there are cases where a transitional raising of the velum above the normal non‐nasalized level is observed, (3) an apparently inherent lingual gesture is found for the labial nasal consonant, (4) there seem to be characteristic gestures of the tongue for the (syllable final) geminate consonant both in the pertinent time segment and in the (impertinent) initial consonant of the syllable. A quantitative analysis of dynamics of the velum movement is discussed with reference to syllable structures.

The acoustic characteristics of speech sounds are often affected by the speaking rate. The purpose of our study was to investigate the extent to which the application of certain low‐level phonological rules are correlated with the speech rate. The corpus consisted of a short paragraph made up of 77 syllables, 47 words, and 4 sentences. The paragraph was constructed such that the application of a low‐level phonological rule, such as gemination, palatalization, and flapping, was permissible, at many of the word boundaries. The paragraph was read several times by 4 speakers, 2 male and 2 female, at their normal speaking rate, as well as twice and half of their normal rates. The sentences were then digitized using the SPIRE facility and the corresponding spectrograms were made. At each word boundary, the experimenters separately determined whether a particular rule had been applied through critical listening and careful examination of the spectrograms and other interactive displays. The results were tabulated...