Catégorie d'événements : SRPP

Many phonological theories analyze speech as a hierarchical structure of segments, moras, and syllables, and this sort of structure is useful for describing cross-linguistic variation in phonological patterns. But does a hierarchical organization govern the production of speech ? In this talk I will describe the Selection-coordination theory of speech production, which holds that articulation is not governed by a fixed hierarchical structure, but rather by flexibly assembled sets of articulatory gestures. The theory holds that selection and coordination mechanisms give rise to two prototypical control regimes : competitive control and coordinative control. Over the course of development, children acquire coordinative control over movements that were previously competitively controlled, this process being mediated by the internalization of sensory feedback. In this framework, segments, moras, and syllables are viewed as differently-sized instantiations of a more general motor planning unit, the organization of control in any given utterance is task/context-specific, and hierarchical structure is the byproduct of a developmental progression. Evidence for the theory is drawn from research in motor control, speech development, and phonological and phonetic patterns.

This study addresses how vowel quality interacts with larynx state. It is known that larynx height varies in relation to vowel quality, but inconsistently so (Ewan & Krones, 1974 ; Ladefoged, DeClerk, Lindau, & Papçun, 1972). Some have also suggested that the intrinsic F0 of vowels arises from lingual-laryngeal interaction in producing vowels (Ohala, 1987 ; Whalen, Gick, Kumada, & Honda, 1998). Lingual-laryngeal interactions form an important part of Esling’s (2005) Laryngeal Articulator Model, which advocates a reshaping of vowel space organization to reflect the laryngeal component of vowel articulation, and Moisik (2013) proposes that lingual-laryngeal interaction emerges in phonological phenomena relating vowel quality to laryngeal qualities and pharyngeal consonants. Magnetic resonance imaging (MRI) lends itself well to studying lingual-laryngeal state because of its ability to capture rich information about the vocal tract in a safe way. We have acquired an MRI data set featuring key vowel qualities in a number of different phonetic contexts captured using 2D (axial, coronal, and sagittal) multi-slice sequences and 2D midsagittal static sequences. The goal is to examine the variation in laryngeal structures as a function of vowel quality in several laryngeal contexts (modal and creaky phonation, glottal and epiglottal stop, and raised larynx voice). The talk will provide an outline of this data set, details about our approach to its analysis, and some preliminary results concerning the quantification of larynx height and epilaryngeal cavity dimensions.

La phonologie diachronique connaît depuis les années 1960 un renouveau avec le développement des études de micro-diachronie, c’est-à-dire du changement en temps réel. Dans le cadre d’une réflexion plus large sur l’articulation entre synchronie et diachronie, je m’intéresse dans cette communication à un processus en particulier, la dissimilation. En m’appuyant sur la “loi de Grassmann” en grec ancien, je montre en quoi ce processus échappe aux cadres d’analyse traditionnels de la phonologie diachronique. Je propose de réanalyser le problème à la lumière d’un processus de variation phonétique récemment découvert, la “dissimilation graduelle” (Jatteau & Hejná 2016) : le trait d’aspiration est réduit lorsqu’il se trouve à proximité d’un autre trait d’aspiration. Je présente mes travaux en cours sur le mongol et une variété d’anglais, et développe les implications du processus pour la diachronie et la typologie.

Paradigmatic contrast describes cases where the phonological grammar seems to conspire to avoid pernicious homophony or similarity between two forms in a morphological paradigm. This paper proposes that a puzzling phonotactic restriction in Haitian can be analyzed as a case of paradigmatic contrast motivated by perceptual similarity avoidance.

In Haitian, pronouns can generally reduce when adjacent to a vowel, but the 3d pers plural pronoun yo [jo] cannot reduce to y [j] after a vowel. This is puzzling because postvocalic [j] is allowed in the Haitian lexicon. The reduction of yo to y is argued to be blocked in coda position to avoid a perceptual confusion with the short form l [l] of the 3d pers singular pronoun li [li] in this context. The hypothesis is supported by external phonological evidence showing that the [l]-[j] contrast may be maintained prevocalically but not postvocalically (e.g. Cibaeno, a Spanish dialect) and by internal perceptual evidence showing that Haitian [l] and [j] are more confusable postvocalically than prevocalically.

If correct, this analysis provides further evidence for the role of phonetic detail in shaping morphological paradigms.

Ce projet de recherche vise à étudier la production et la perception de la parole chez les apprenants francophones de l’allemand. Un corpus de parole de 7 heures correspondant à trois tâches (imitation, lecture, description) a été enregistré. Il comprend des germanophones natifs et des apprenants francophones. Nous avons analysée les productions des segments intéressants d’après le cadre du SLM. Une étude de perception en EEG utilisant [h-ʔ], [ʃ-ç] et les voyelles courtes et longues a été réalisée sur des germanophones natifs et des apprenants francophones. Enfin, l’impact de l’enseignement sur l’amélioration des production et perception a été examiné à travers une étude longitudinale. L’étude de production montre que, suivant les tâches, les apprenants produisent le [h] en début de mot sans problème majeur. De même, ils peuvent produire des voyelles de durée contrastive. Cependant, pour les trois tâches, les apprenants ont plus de difficultés pour la production de la qualité vocalique, de [ç] et [ŋ]. Fait notable, la perception ne reflète pas toujours la production. Les apprenants tendent à ne pas percevoir le [h] en début de mot alors que la production de ce segment en répétition est bonne. À l’inverse, les apprenants perçoivent le contraste [ʃ-ç] mais sa production reste difficile. Seulement dans les voyelles courtes et longues, la perception reflète la production.

L’étude d’enseignement montre que la conscience linguistique affecte différemment perception et production : une conscience linguistique accrue permet d’affiner la perception de phonèmes à contenu acoustique complexe et la production des phonèmes faciles à produire du point de vue articulatoire.

Les langues expriment la structure de l’information à l’aide de moyens grammaticaux variés. L’attention des chercheurs s’est surtout portée sur les langues à proéminence accentuelle (focus en tant que proéminence), comme les langues germaniques et la plupart des langues romanes. Les langues intonatives, à accent lexico-tonal, les langues à tons, et les langues de phrasage utilisent des réflexes syntaxiques, morphologiques et prosodiques pour signaler un constituent focalisé, donné ou de topique. Néanmoins, la mélodie intonative est toujours compositionelle, établie par la séquence des tons : accent lexical, tons de frontières et tons purement intonatifs. Les tons ne sont pas nécessairement associés à des syllables proéminentes, mais peuvent tout simplement être alignés avec leur domaine prosodique en fonction de la syntaxe et la sémantique de la phrase dans laquelle ils se trouvent. Certaines langues alignent le focus avec les frontières phonologiques (focus en tant qu’alignement) et n’utilisent aucune proéminence focale. Je montrerai dans mon exposé que l’on peut établir une corrélation entre les différents moyens d’expression du focus et le type intonatif de chaque langue. J’illustrerai ces thèmes théoriques par des exemples émanant de langues différentes.

Les technologies numériques facilitent l’échange et le partage des résultats de la recherche, ainsi que des outils, méthodes et données. On constate néanmoins une certaine déperdition d’outils et de données : un cas extrême est celui de travaux de phonétique expérimentale dont il ne reste au final que des publications, elles-mêmes placées derrière une barrière de péage qui limite leur diffusion. Il paraît pourtant clair que la phonétique/phonologie aurait tout à gagner à une association plus étroite entre théorie et empirie, qui permettrait d’inscrire les travaux de recherche dans une logique cumulative. L’exposé présente certaines idées du mouvement pour le libre accès et la « science ouverte » (« Open Science »), et des exemples de réalisations dans le domaine du partage d’outils logiciels, de l’archivage pérenne des bases de données, et de l’édition scientifique.

Quelques liens et références :

 Le cercle vicieux de l’édition scientifique (vidéo). « A regarder par tous si ce n’est déjà fait »
 « Au pays des merveilles du libre accès : la préparation d’un premier ouvrage chez Language Science Press ». Billet du Carnet de la bibliothèque des Langues O’, Le Carreau de la BULAC. https://bulac.hypotheses.org/5761
 Revue Lab Phon, en libre accès depuis cette année : https://www.journal-labphon.org/
 Au sujet de la collecte de données : Niebuhr et Michaud (2015) Speech data acquisition : the underestimated challenge. (Pour lecteurs motivés : 40 pages)

Motor symptoms of Parkinson’s Disease (PD), even if captured at the level of (non-speech) oral motor performances, can differ from those individuated during the speech motor performance (Bunton, 2008 ; Skodda et al., 2012). As a consequence, the understanding of dysarthria in PD needs to rely on data elicited by speech events rather than non-speech oral motor tasks. For this type of investigations, a good number of studies in several languages (Weismer, 1984 ; Forrest et al., 1989 ; Yunusova et al., 2008 for American English ; Skodda et al., 2011, 2012 for German ; Rusz et al., 2013 for Czech ; Audibert & Fougeron, 2012 for French ; Gili Fivela et al. 2014, 2015 ; Iraci, 2017 for Italian) have already focussed on linguistic units, such as the vowels. However, some ambiguities arise from the findings across different languages. These ambiguities may depend on the language-specific constraints that are attributed to vowels production and that may lead one pathological speaker to differently compensate coherently with its own language requirements. Modulation of vocalic gestures as a function of syllable and stress in Italian are found in the healthy speakers’ literature (e.g., Bertinetto, 1981) and indicate that stress modifications (stressed vs. unstressed vowel) have the effect of a global reduction/enhancement of the gestural articulation (as in hypo/hyper-articulation, Avesani et al., 2009) ; syllabic structure effects seem to manifest at the level of vowels’ duration and timing (Gili Fivela et al., 2007) with acoustic shortenings attributed to more complex syllabic structures, and lengthening of timing intervals as a function of intervocalic consonants insertion. As a consequence, this study provides an observation of speech alterations by 5 Italian PD dysarthric speakers (compared to 5 healthy age-matched controls) when vocalic gestures are expected to be modified in space and time because of stress and syllabic structure changes.

In particular, it is hypothesised that 1a) HC show more centralised unstressed vowels compared to stressed ones, 1b) PD show centralisation of unstressed vowels compared to stressed ones as well, though a general centralisation could be found in both conditions compared to HC ; 2a) HC increase the timing of vowels when more intervening consonants occur in the syllabic structure ; 2b) PD increase the timing of vowels when intervening consonants are added in the syllabic structure though this modification could be obtained through alternative gestural phasing strategies ; 3a) HC reduce the duration of vowels in closed vs. open syllable ; 3b) PD reduce the duration of vowels in closed vs. open syllable though pathological durations may be even more reduced than HC’s in both conditions. Kinematic (tongue dorsum) and acoustic data of ‘CVC(C)V disyllables (/’pi.pa, ‘pip.pa, ‘pa.pi, ‘pap.pi/) have been collected through Electromagnetic Articulography (EMA – AG501). For each vowel and condition (/i/ vs /a/ ; stressed vs unstressed ; CVCV vs CVC.CV), F1, F2, and tongue dorsum (TD) position have been measured as well as timing between TD targets, timing between vocal folds activation, acoustic durations and duration of gestural attainment phases.

Results confirm all hypotheses. However, HC exhibit timing modulation only through vocal folds activation while PD also through TD targets. The most important fact seems the reduced range of PD observations compared to HC in crucial measures. When this happens for the position of TD, the effect is a centralisation of vowels articulation ; concerning duration measures, this coincides with shorter lags than HC when lags are expected to be lengthened, and longer lags when expected to be shortened. Findings show that pathological observations needs to vary across linguistic context in order to be interpreted : in fact, durations have been found to be both reduced or increased depending on the syllabic structure. Moreover, the reduced range of PD results observed in both space and time may be connected to an alteration of the auditory feedback though this consideration is only speculative and needs further studies.

References

Audibert, N. & Fougeron, C. (2012). Distorsions de l’espace vocalique : quelles mesures ? Application à la dysarthrie. Proceedings of the 29èmes Journées d’Etudes sur la Parole/JEP-TALN-RECITAL 2012, pp. 217-24.

Avesani, C., Vayra, M., Zmarich, C. (2009). Coordinazione vocale-consonante e prominenza accentuale in italiano. La sfida della Articulatory Phonology. In Linguistica e modelli tecnologici di ricerca, Roma : Bulzoni, 2009, pp. 353-86.

Bertinetto, P.M. (1981). Strutture prosodiche dell’Italiano. Florence : Accademia della Crusca

Bunton, K. (2008). Speech versus Nonspeech : Different Tasks, Different Neural Organization. In Semin Speech Lang. 2008 Nov ; 29(4) : 267–275.

Forrest, K., Weismer, G. & Turner, G.S. (1989). Kinematic, acoustic, and perceptual analyses of connected speech produced by Parkinsonian and normal geriatric adults. In Journal of the Acoustical Society of America, vol. 85, no. 6, pp. 2608–22.

Gili Fivela, B., Iraci, M.M., Sallustio, V., Grimaldi, M., Zmarich, C., Patrocinio, D. (2014). Italian Vowel and Consonant (co)articulation in Parkinson’s Disease : extreme or reduced articulatory variability ?. In Proceedings of the 10th International Seminar on Speech Production (ISSP’14), 5-8 May 2014, Cologne (Germany), pp. 146-9.

Gili Fivela, B., Iraci, M.M., Grimaldi, M., Zmarich, C. (2015). Consonanti scempie e geminate nel Morbo di Parkinson : la produzione di bilabiali. In C. Avesani, M. Vayra, F. Tamburini (eds.) Il farsi e disfarsi del linguaggio. Acquisizione, mutamento e destrutturazione della struttura sonora del linguaggio / Language acquisition and language loss. Acquisition, change and disorders of the language sound structure, Milano : AISV.

Gili Fivela, B., Zmarich, C., Perrier, P., Savariaux, C., Tisato, G. (2007). Acoustic and kinematic correlates of phonological length contrast in Italian consonants. In Proceedings of International Conference of Phonetic Sciences (ICPhS’07), Saarbrücken, Germany, 6-10 August 2007, pp. 469-72

Iraci, M.M. 2017. Vowels and Consonants and Co-articulation in Parkinson’s Disease – unpublished PhD Thesis, Università del Salento / Universität zu Köln.

Rusz, J., Cmejla, R., Tykalova, T., Ruzickova, H., Klempir, J., Majerova, V., Ruzicka, E. (2013). Imprecise vowel articulation as a potential early marker of Parkinson’s disease : Effect of speaking task. In Journal of the Acoustical Society of America, 134, 2171–2181.

Skodda, S., Gronheit, W., Schlegel, U. (2012). Impairment of Vowel Articulation as a Possible Marker of Disease Progression in Parkinson’s Disease. In PLoS ONE, 7 (2), pp. 1–8.

Skodda, S., Visser, W. & Schlegel, U. (2011). Vowel articulation in Parkinson’s disease. In Journal of Voice, 25, 4, 467–72.

Weismer, G. (1984). Articulatory characteristics of Parkinsonian dysarthria : Segmental and phrase-level timing, spirantization, and glottal-supraglottal coordination. In M. McNeil, J. Rosenbek & A. Aronson (eds.), The dysarthrias : Physiology, acoustics, perception, & management, San Diego, CA : College-Hill, pp. 101–30.

Yunusova, Y., Weismer, G., Westbury, J.R., Lindstrom, M.J. (2008). Articulatory movements during vowels in speakers with dysarthria and healthy control. In Journal of Speech, Language, and Hearing Research 51 (3), pp. 596-611.

Modern Risiangku Tamang (Tibeto-Burman, Nepal) has a system of four tones, two high tones and two low tones, which results from the two-way split of an earlier two-tone system linked to the disappearance of an older voicing contrast on initial consonants (Ci). The voicing contrast has fully disappeared in sonorant and fricative Ci, thus phonologizing the tone system, and partially disappeared in plosive Ci. In the latter case, three cues are used in production : F0, voice quality, and some residual voicing of Ci (Mazaudon & Michaud 2008). To summarize, (1) low tones are produced with breathier voice than high tones ; (2) 20% of the low tone syllables (vs. 0% of the high tone syllables) are produced with prevoiced Ci.

Variation in cue weighting has been reported for a number of languages where tone is still incipient like Korean (Silva 2006) or Afrikaans (Beddor 2015 ; Coetzee, p.c.). In Tamang, tone is phonologized ; however, old features/cues, i.e., Ci prevoicing and breathy voice, are still present in production. This study explores how old and new cues are used in the perception of tones in Risiangku Tamang.

We conducted a four-alternative forced-choice identification test on Risiangku listeners living in Nepal with a quasi-quadruplet. All stimuli were synthesized with equalized segmental duration and intensity contour. Five parameters were manipulated : degree of breathiness (modal, breathy, or super breathy) ; Ci prevoicing (present or absent) ; F0 of V1 onset (115, 130, 145, or 160 Hz) ; F0 slope of V1 (rising or falling) ; F0 slope of V2 (rising or falling). The resulting 96 stimuli were presented in a carrier sentence to 28 listeners individually.

Concerning the difference between high and low tones, results show that F0 onset, Ci prevoicing, and voice quality are all in play. The identification rate of high tones (T1-2) increases with the F0 onset. However, the slopes of all the identification curves are less steep than what we would expect of a sharp categorical perception. Listeners also rely heavily on both Ci prevoicing and voice quality : prevoiced Ci stimuli yield low tone identification in 50% of cases even with the highest F0 onset ; and the high tone identification rate is lower for breathy than for modal stimuli, and for super breathy than for breathy stimuli. Statistical (GLM) and classification (J48 tree, SMO) analyses confirmed a preponderance of the F0 onset cue in high vs. low tone identification, followed by reliance on the prevoicing of the Ci when present. The use of voice quality comes in to resolve conflicting cues : i.e., a high F0 with Ci prevoicing or a low F0 without Ci prevoicing. Moreover, when cues are conflicting, response times increase.

Our study shows a predominance of F0 as an identification cue, which is consistent with the historical evolution. However, the old features/cues of [voice] and [breathiness] continue to be used in perception as in production. The substantial role of [voice] in perception in spite of its reduced role in production might suggest that its progressive disappearance is production-driven.

References :

Beddor, P. S. (2015). The relation between language users’ perception and production repertoires. Proc. of the 18th ICPhS, Glasgow, UK

Mazaudon, M., & Michaud, A. (2008). Tonal contrasts and initial consonants : a case study of Tamang, a ’missing link’ in tonogenesis. Phonetica, 65, 231-256.

Silva, D. J. (2006). Acoustic evidence for the emergence of tonal contrast in contemporary Korean. Phonology, 23, 287-308

While recent studies in L3 phonology largely focus on the consecutive learning of foreign languages (Llama et al. 2010 ; Wrembel 2014), little is known about the acquisition of L3 phonology in learners who acquired more than one language during early childhood. This particularly holds for foreign language learning in multilingual speakers who speak a so-called heritage language (Valdés 2000) in addition to their dominant language. In my talk, I will present data produced by adolescents who speak Chinese, Russian or Turkish as an HL along with German and take English, French and/or Russian courses at school. Control-data from monolingually raised German learners are taken into account. In addition to the recordings, extra-linguistic data were collected from all participants (background questionnaire, semi-focused interview, phonological awareness test following Osburne 2003). I will focus on the outcomes of (completed and ongoing) research on two related phenomena, namely global speech rhythm and Voice Onset Time (VOT ; Lisker & Abramson 1964). The results obtained from the analyses performed on the speech data show cross-linguistic influence (CLI) in terms of both positive and negative transfer from the HLs, which may constitute a multilingual advantage depending on the characteristics of the interacting languages. The extra-linguistic data indicate that positive transfer from the HLs is favored by the learner’s language attitudes as well as by their individual degree of multilingual and phonological awareness. By and large, I interpret the findings as evidence to suggest that multilingual and phonological awareness should be promoted in school education.

References

• Lisker, L. & Abramson, A. S. 1964. A cross-language study of voicing in initial stops. Word 20, 384–422.

• Llama, R., Cardoso, W. & Collins, L. 2010. The influence of language distance and language status on the acquisition of L3 phonology. International Journal of Multilingualism 7, 39–57.

• Osburne, A. G. 2003. Pronunciation strategies of advanced ESOL learners. International Review of Applied Linguistics in Language Teaching 41, 131–141.

• Valdés, G. 2000. Teaching heritage languages : An introduction for Slavic-language-teaching profes-sionals. In O. Kagan & B. Rifkin (Eds.), Learning and teaching of Slavic languages and cultures : Toward the 21st century. Bloomington : Slavica, 375–403.

• Wrembel, M. 2014. VOT patterns in the acquisition of third language phonology. Concordia Working Papers in Applied Linguistics (COPAL) 5, 750–770.