Working Papers in Linguistics: Volume 45 (February 1995)
Permanent URI for this collection
Papers from the Linguistics Laboratory. Edited by Stefanie Jannedy
Front Matter
pp. i-x
Description |
Full Text PDF
Interaction with autonomy: Defining multiple output models in psycholinguistic theory
Boland, Julie E.; Cutler, Anne pp. 1-10
Description |
Full Text PDF
Vocal tract evolution and vowel production
Cohen, K. Bretonnel pp. 11-35
Description |
Full Text PDF
Final Lowering in Kipare
Herman, Rebecca pp. 36-55
Description |
Full Text PDF
Gestural Phasing as an Explanation for Vowel Devoicing in Turkish
Jannedy, Stefanie pp. 56-84
Description |
Full Text PDF
Intervocalic consonant sequences in Korean
Johnson, Keith; Oh, Mira pp. 85-97
Description |
Full Text PDF
Acoustic and Intonational Correlates of Informational Status of Referring Expressions in Standard Korean
Kang, Hyeon-Seok pp. 98-130
Description |
Full Text PDF
Longitudinal Study of the Acquisition of Taiwanese Initial Stops
Pan, Ho-Hsien pp. 131-159
Description |
Full Text PDF
Perceptual Evidence of Tonal Coarticulation
Peng, Shu-hui pp. 160-165
Description |
Full Text PDF
Temporal Structure of an Estonian Lament: A Case Study
Ross, Jaan; Lehiste, Ilse pp. 166-169
Description |
Full Text PDF
Browse
Recent Submissions
Item Front Matter (Volume 45, February 1995)(Ohio State University. Department of Linguistics, 1995-02)Item Interaction with autonomy: Defining multiple output models in psycholinguistic theory(Ohio State University. Department of Linguistics, 1995-02) Boland, Julie E.; Cutler, AnneThere are currently a number of psycholinguistic models in which processing at a particular level of representation is characterized by the generation of multiple outputs, with resolution involving the use of information from higher levels of processing. Surprisingly, models with this architecture have been characterized as autonomous within the domain of word recognition and as interactive within the domain of sentence processing. We suggest that the apparent internal confusion is not, as might be assumed, due to fundamental differences between lexical and syntactic processing. Rather, we believe that the labels in each domain were chosen in order to obtain maximal contrast between a new model and the model or models that were currently dominating the field.Item Vocal tract evolution and vowel production(Ohio State University. Department of Linguistics, 1995-02) Cohen, K. BretonnelEvolution has left the anatomically modern human with a supralaryngeal airway which is qualitatively different from that of all other animals. The source-filter theory of speech (the dominant theory of speech production in modern phonetics) relates articulations to their acoustic outputs by means of the action of a "filter"--the supralaryngeal vocal tract--on a "source"--air coming from a vibrating larynx. Since the human "filter" is qualitatively different from that of other animals, we should expect that the acoustic outputs that a human could generate would be different from those that other animals can produce. However, no satisfactory account of the nature of these differences has yet been given. A computer model was used to calculate variations in vocal tract transfer functions in human and non-human supralaryngeal airways as the location of a vocalic constriction is varied. The results suggest that the nature of the acoustic differences between the human and non-human anatomies has to do with nonlinearities which characterize articulatory/acoustic relations for the anatomically modern human vocal tract more so than for the non-human supralaryngeal airway. These differences arise from the range of articulations which are producible in a vocal tract with a bend in one wall, as opposed to one where both walls are straight. For the high vowels /i/ and /u/, these nonlinearities lead to areas of formant stability in the human airway which are significantly larger than those in the non-human airway. For the low vowel, the results suggest that the non-human airway should not be able to produce a front/back contrast between low vowels. In contrast, modelling of the modern human vocal tract correctly predicts the possibility of a front/back contrast for low vowels, though without any increase in areas of formant stability. The extrinsic tongue musculature of the human was then compared with that of Pan troglodytes. Using a perturbation theory model to evaluate the acoustic effects of extrinsic tongue muscle activity, it was found that the ability to generate the vowel triangle is related to the functional potentials of the extrinsic tongue musculature. These are not acoustically significant in the non-human. In the human, they give rise to the ability to generate the extreme points of the vowel triangle.Item Final Lowering in Kipare(Ohio State University. Department of Linguistics, 1995-02) Herman, RebeccaData from Kipare, a Bantu language which has both lexical tone and final lowering, were examined to determine the optimal model of final lowering in a tone language. The models evaluated are the register model, in which final lowering is accounted for by manipulating tone categories, and the boundary tone model, in which final lowering is accounted for by a separate tone on the same tier as lexical tones but associated to a higher prosodic position. Aspects of final lowering in Kipare examined in this study include which lexical tone categories show final lowering, whether the lowered tones coincide with tones from another lexical tone category, the locality of the effect, the relation of contrasting utterance types to utterance types showing final lowering, and what the domain of final lowering is.Item Gestural Phasing as an Explanation for Vowel Devoicing in Turkish(Ohio State University. Department of Linguistics, 1995-02) Jannedy, StefanieRecent work in phonetics has suggested that vowel devoicing or schwa deletion, observed in various languages, is a gradient process. This study provides evidence for the previously undocumented process of high vowel devoicing in Turkish. The prosodic and segmental factors rate, stress, preceding environment, following environment, vowel type, and syllable type were investigated. The factors are described, evaluated and ranked according to the results of a multiple regression (Variable Rule) analysis. Where applicable, results are contrasted with findings for i.e., Japanese and Korean. Furthermore, VOT (voice onset time) measurements of the three voiceless stops [p t k] were obtained, as well as duration measurements of vowels in open and closed syllables where vowels are significantly longer in Turkish. Generally, most devoicing occurred when the vowel was shorter (i.e., as a result of faster rates of speech, lack of stress, in closed syllables, ect.). These findings accord well with predictions made by a model assuming gradual gestural overlap of adjacent consonantal and vocalic gestures. It will be attempted to explain the findings with differences in phasing between articulatory gestures.Item Intervocalic consonant sequences in Korean(Ohio State University. Department of Linguistics, 1995-02) Johnson, Keith; Oh, MiraThis paper reports the results of an instrumental phonetic study of intervocalic consonant sequences in Korean. The study explored a putative positional neutralization produced at the phonetics/phonology interface. It was designed to determine whether Korean intervocalic laryngeal consonants are phonetically distinct from geminates, plain consonants, or laryngeal consonants in consonant clusters. The results showed that the contrast between intervocalic tensed singletons and geminates was neutralized, and that both of these patterned with heterorganic consonant sequences rather than plain singletons. Moreover, we found that this neutralization persisted across (limited) variation in speaking rate, although intervocalic tense consonants were more compressible in faster speech than were post-consonantal tense consonants.Item Acoustic and Intonational Correlates of Informational Status of Referring Expressions in Standard Korean(Ohio State University. Department of Linguistics, 1995-02) Kang, Hyeon-SeokThis paper examines the phonetic correlates of informational status of referring expressions in Standard Korean. Two experiments are conducted. The results show that speakers of Standard Korean indicate the informational status of referring expressions not only with acoustic cues such as amplitude and duration but also with intonational phrasing. The results also suggest that speakers of Standard Korean rather clearly distinguish "current" and "displaced" entities. The term 'intonational attenuation' is proposed to describe the use of intonational phrasing by some languages to indicate the given status of referring expressions.Item Longitudinal Study of the Acquisition of Taiwanese Initial Stops(Ohio State University. Department of Linguistics, 1995-02) Pan, Ho-HsienThis study focuses on the Taiwanese stop system and is intended to provide information regarding the three-way voicing contrast of Taiwanese stops, as they are acquired by children. There are three voiceless unaspirated stops in Taiwanese, /p, t, k/, three voiceless aspirated stops /ph, th, kh/ and two voiced stops /b, g/. A longitudinal study of two Taiwanese-acquiring girls between the ages of 28 and 29 months to 33 and 40 months respectively, indicated that Taiwanese voiceless unaspirated stops were acquired first, voiceless aspirated stops secondly, and finally voiced stops. Even though the two girls acquired the three-way voicing contrast by the age of three years, their productions were not adult-like, especially for the voiced stops. There were many voiced stops produced with short lag VOTs as was found in the voiceless unaspirated stops.Item Perceptual Evidence of Tonal Coarticulation(Ohio State University. Department of Linguistics, 1995-02) Peng, Shu-huiThe current study examines the extent to which tonal coarticulation is perceptible. Ten tokens each of two Taiwanese syllables (high tone and mid-falling tone) were placed in three contexts in which the following tone was high, mid, or mid-falling. The task of the listeners was to identify the whole phrase in a threeway forced choice after listening to the first syllable. Acoustic analysis showed that the high tone was contextually more stable than the mid-falling tone. Anticipatory assimilation of F0 was found between the mid-falling tone and its following tone. The results of the identification test indicated that contextual variability was perceptually detectable by listeners and contributed to the recognition of the following tone. Consistent with the patterns of assimilation, the following tone was more predictable when the phrase started with a mid-falling tone than when it started with a high tone. As is the case with segmental coaiticulation, tonal coarticulation changes phonetic features of neighboring tones which contribute to the recognition of the tone. However, the variability of the tonal features are constrained by the phonological system of the language to maintain phonological contrast.Item Temporal Structure of an Estonian Lament: A Case Study(Ohio State University. Department of Linguistics, 1995-02) Ross, Jaan; Lehiste, IlseThe acoustic structure of a South-Eastern Estonian lament was analyzed. The lament was recorded in 1972 as performed alternatingly by a soloist and a choir (of several female voices). In this lament, the basic structural unit consists of two four-foot verse lines of alternating trochaeic and dactylic feet (2 + 3+ 2 + 3 syllables). The first verse of the couplet has invariant verbal content throughout the whole lament. The two-line unit is first performed by the soloist and, at the end of the second line, taken over and repeated by the choir. There are 11 structural units in the whole lament. Due to an almost complete one-to-one correspondence between syllables in the text and notes in the melody, the overall number of notes in the lament is n = (2 + 3 + 2 + 3) x 2 x 2 x 11 = 440 reduced by occasional pauses where the performer has to take a breath. Durations of all notes are perceived as equal by a listener, the exception being the very end of each structural unit where the last two notes are sung at a faster speed. The possible influence of a number of factors on syllable/note durations was tested by an ANOVA. These factors included word stress, metrical accent, modus of performance (solo or ensemble), variability of verbal contents (first or second line of the structural unit), position in the verse line, and position of the structural unit in the whole lament. The results were significant for performance modus, syllable/note position, and (somewhat less) for word stress. A closer study of differences between the performances of the soloist and the choir reveals a marked ritenuto at the end of the second verse line sung by the soloist, probably aimed at signaling to the choir that it has to step in.