The articulatory awareness is the part of phonological awareness which is responsible for the distinction of different articulation points of speech sounds. This ability, in addition to improving the perception and production of speech, facilitates learning of an alphabetic writing system. This study aimed to analyze the development of articulatory awareness skills in children with normal phonological development and to verify the articulatory awareness skills of children with speech disorders, comparing them with children with normal phonological development. One hundred and twenty subjects participated of this study. They were students of preschool to first grade of elementary school, and had their articulatory awareness evaluated. Initially, the data of only ninety participants with normal phonological development were analyzed, with the aim of obtaining data about their skills development in articulatory awareness with normal phonological development. Subsequently, data from thirty subjects with speech disorders to thirty subjects with normal phonological development were compared. Based on the analysis of the results, we conclude that the performance in articulatory awareness improves with age and schooling. Female children present better performance than male children. The better is the performance in articulatory gesture perception tasks, the better will be the performance in articulatory gesture production tasks. Children with speech disorders have bigger difficulty in phonological awareness when compared to children with normal phonological development. / A consciência fonoarticulatória é a parte da consciência fonológica responsável pela distinção dos diferentes pontos de articulação dos sons da fala. Essa habilidade, além de auxiliar a percepção e produção da fala, facilita a aprendizagem de um sistema de escrita alfabético. Os objetivos deste estudo foram analisar o desenvolvimento das habilidades em consciência fonoarticulatória de crianças com desenvolvimento fonológico normal e verificar as habilidades em consciência fonoarticulatória de crianças com desvio fonológico, comparando-as com crianças com desenvolvimento fonológico normal. Participaram desta pesquisa cento e vinte sujeitos da Educação Infantil e primeira série do Ensino Fundamental avaliados em consciência fonoarticulatória. Num primeiro momento foram analisados os dados de noventa participantes com desenvolvimento de fala normal, com o intuito de obter dados sobre o desenvolvimento das habilidades em consciência fonoarticulatória em condições normais de desenvolvimento de fala. Posteriormente, foram comparados os dados de trinta sujeitos com desvio fonológico com os de trinta sujeitos com desenvolvimento de fala normal. Com base na análise dos resultados obtidos, pôde-se concluir que a consciência fonoarticulatória é uma habilidade que se aprimora de acordo com a idade e com a escolaridade. Crianças do sexo feminino apresentaram melhor desempenho nessa habilidade se comparadas as do sexo masculino. Quanto melhor for o desempenho nas tarefas de percepção da consciência fonoarticulatória, melhor será o desempenho nas tarefas de produção da consciência fonoarticulatória. Crianças com desvio fonológico apresentam maior dificuldade em habilidades de consciência fonoarticulatória se comparadas a crianças com desenvolvimento fonológico normal.

A presente tese se ocupa da percepção fônica de línguas não nativas e tem como objetivo repensar as premissas básicas do PERCEPTUAL ASSIMILATION MODEL-L2 [PAM-L2 (BEST; TYLER, 2007)] no que diz respeito aos seus eixos cognitivo, fônico e filosófico. De acordo com os proponentes do modelo, (i) a percepção da fala não nativa dispensa mecanismos cognitivos no que se refere a representações mentais ou processos inferenciais; (ii) a unidade de análise do evento perceptual, em termos de fala, é o gesto articulatório; e (iii) o acesso às informações disponíveis no mundo é direto, garantido pela atuação dos sentidos como nossos próprios sistemas perceptuais. Julgamos que tais premissas são limitadas e incoerentes com o objeto de investigação dos autores e argumentamos, portanto, que os eixos cognitivo, fônico e filosófico do modelo devam ser vislumbrados sobre um ponto de vista alternativo. Quanto ao primeiro eixo, defendemos que o evento perceptual seja concebido essencialmente como um fenômeno cognitivo, criado e gerenciado pelo encéfalo, que envolve abstrações, representações mentais e inferências acerca dos objetos do mundo. Em relação ao segundo eixo, julgamos que o tratamento acústico-articulatório (ALBANO, 2001) à unidade gestual seja mais apropriado à percepção de elementos fônicos não nativos, diferindo do tratamento articulatório (BROWMAN; GOLDSTEIN, 1989, 1992) que reside originalmente no modelo. No que diz respeito ao terceiro eixo, adotamos a posição realista indireta (JACKSON, 1977, 2010) como sendo aquela que abarca de modo mais adequado a percepção das unidades fônicas não nativas, em detrimento do realismo direto (J. GIBSON, 1966, 1986). Decorrentes de nosso refinamento teórico, questões adicionais à percepção fônica não nativa são endereçadas, as quais concernem à falsa dicotomia L2 VS. LE, à influência do sistema grafêmico sobre a percepção fônica e à formação de novas categorias fônicas a serviço do idioma-alvo. Discutimos, também, os aspectos metodológicos de alguns estudos perceptuais, assim como suas implicações para uma nova caracterização do modelo, e ponderamos sobre o delineamento, o tipo de conhecimento de base testado (língua materna ou não nativa) e o objetivo de cada tarefa perceptual a ser empregada em ambiente laboratorial. / This Doctoral Dissertation addresses the phonic perception of non-native languages and aims to rethink the central tenets of the PERCEPTUAL ASSIMILATION MODEL-L2 [PAM-L2 (BEST; TYLER, 2007)] with respect to its cognitive, phonic and philosophical spheres. According to the proponents of the model, (i) the perception of non-native speech disregards any cognitive mechanisms related to mental representations or inferential processes; (ii) the unit for analyzing perceptual events, in terms of speech, corresponds to the articulatory gesture; and (iii) we have direct access to the information available in the world, since our senses act as our own perceptual systems. We argue that these premises are limited and inconsistent with the research object of the authors and, thus, we assume that the cognitive, phonic and philosophical spheres of the model should be glimpsed from an alternative point of view. As for the first sphere, we argue that the perceptual event is essentially a cognitive phenomenon, created and managed by the brain, which involves abstractions, mental representations and inferences about the objects of the world. Regarding the second sphere, we state that an acoustic-articulatory treatment (ALBANO, 2001) to the gestural unit is more appropriate to the perception of non-native phonic elements, differing from the articulatory treatment (BROWMAN; GOLDSTEIN, 1989, 1992) that is originally conveyed by the model. With respect to the third sphere, opposing direct realism (J. GIBSON, 1966, 1986), we adopt the indirect realist position (JACKSON, 1977, 2010) as the one that encompasses the perception of non-native phonic units more adequately. Due to our theoretical refinement, we address additional issues to the phonic perception, which concern the false L2 vs. LE dichotomy, the impact of graphical systems on phonic perception, and the formation of new phonic categories at the service of the target language. We discuss methodological aspects of perceptual studies, as well as their implications for a new characterization of the model, and we also consider the design, the type of knowledge tested (native or nonnative language) and the purpose of each perceptual task to be used in a laboratory environment.

Orientador: Plínio Almeida Barbosa / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Estudos da Linguagem / Made available in DSpace on 2018-08-17T08:58:39Z (GMT). No. of bitstreams: 1 Lima-Gregio_AvelinyMantovan_D.pdf: 2533583 bytes, checksum: 7d9212f2ef11e1b2e829368236a2934d (MD5) Previous issue date: 2011 / Resumo: Os objetivos deste estudo foram: 1) caracterizar a produção de oclusiva glotal de sujeitos com fissura palatina, a partir da percepção deste fenômeno pelos profissionais da área; 2) comparar as produções de cinco sujeitos sem fissura palatina (Grupo Controle - GC) com as de cinco sujeitos com fissura palatina (Grupo Teste - GT) na produção de laringalização em relação às regiões de fronteira; e 3) avançar no entendimento das articulações compensatórias (AC) e da teoria dos sistemas dinâmicos. O corpus consistiu na leitura de um texto realizada por ambos os grupos. A análise da oclusiva glotal foi realizada após a identificação de sua presença por cinco fonoaudiólogos experientes no diagnóstico e avaliação das AC e/ou disfunção velofaríngea (DVF). Em seguida, para cada produção foi observada a presença de três parâmetros (burst, formantes de transição e laringalização). Foi realizado um experimento de percepção, a partir da manipulação do sinal áudio de um participante do GC. Então, a análise prosódica da laringalização foi realizada. Quanto à análise da oclusiva glotal, os resultados mostraram que a presença de formantes de transição e burst foi determinante para a dúvida dos profissionais. Na ausência destes dois parâmetros, os juízes foram significativamente unânimes em afirmar a presença de oclusiva glotal. A presença de laringalização ajudou os fonoaudiólogos a julgarem a ocorrência da oclusiva glotal. A ocorrência de laringalização para o GC foi maior nas regiões de fronteira prosódica, enquanto que para o GT foi dentro e fora destas regiões. O experimento de percepção revelou um estímulo ambíguo, embora com significância estatística para concordância entre os juízes. A discussão foi norteada pelos pressupostos teóricos da Fonologia Articulatória, derivada do modelo de Browman e Goldstein (1990) e da Fonética Cognitiva de Tatham (1994). Discutiu-se ainda o estado da glote, a partir do conceito de válvulas laríngeas (Edmondson e Esling, 2006). A análise da oclusiva glotal evidenciou dois aspectos: a utilização de diferentes válvulas laríngeas e a proposta de um gesto glotal tripartido. A análise da laringalização sugere falha na supervisão do gesto glotal, o acoplamento da oclusiva glotal e da laringalização, e ação das fronteiras como atratores para as laringalizações / Abstract: The objectives of this study were: 1) to characterize the production of glottal stop of individuals with cleft palate, from the perception of this phenomenon by professionals; 2) to compare the productions of five subjects without cleft palate (Control Group - GC) and five subjects with cleft palate (Test Group - GT) to produce laryngealization in relation to border regions; and 3) improve the understanding of compensatory articulations (AC) and the theory of dynamical systems. The corpus consisted of the reading of the same text by both groups. The analysis of the glottal stop was made after identifying the presence by five speech pathologists (judges) experienced in the diagnosis and assessment of AC and/or velopharyngeal dysfunction (DVF). Then, for each production up to three parameters (burst, formant transition and laryngealization) were observed. A speech perception experiment was conducted after manipulating the audio signal of a participant in the GC. Then, the prosodic analysis of laryngealization was carried out. As regards the glottal stops, the results showed that the presence of burst and formant transition was responsible for ambiguous judgments. When these two parameters were absent, the judges were unanimous in signaling the presence of a glottal stop. The presence of laryngealization helped speech pathologists to identify glottal stop. For GC, the occurrence of laryngealization was in the prosodic boundary while for the GT laryngealization occurred everywhere. The speech perception experiment revealed an ambiguous stimulus, although a statistically significant correlation between the judges. The discussion was guided by theoretical assumptions of Articulatory Phonology, derived from the Browman and Goldstein's (1990) model and Tatham's (1994) Cognitive Phonetics. It was also discussed the state of the glottis, from the concept of laryngeal valves (Edmondson and Esling, 2006). Analysis of the glottal stop emphasizing two aspects: the use of different laryngeal valves and a proposed tripartite glottal gesture. The laryngealization analysis suggests failure of the supervisor component of the glottal gesture, the glottal stop and laryngealization coupling and the boundary regions as attractors for laryngealization / Doutorado / Linguistica / Doutor em Linguística

La parole est le moyen de communication le plus utilisé par l’Homme. Elle permet d’exprimer ses besoins, d’échanger ses pensées avec autrui et contribue à la construction de l’identité sociale. C’est aussi un canal de communication complexe impliquant un contrôle moteur élaboré en production et la capacité à analyser des séquences sonores produites par une grande variété de locuteurs en perception. Cette complexité fait qu’elle est souvent le mode de communication le plus altéré ou difficile à acquérir pour des personnes dont les systèmes sensori-moteurs impliqués sont perturbés. C’est en particulier le cas des personnes avec trisomie 21 (T21), syndrome d’origine génétique induisant des difficultés motrices orofaciales complexes et des altérations des sphères auditives et somatosensorielles. Si parler est possible pour la plupart de ces personnes, leur intelligibilité est toujours touchée. Améliorer leur communication orale est un enjeu clinique et d’intérêt social. L’étude de la production de la parole par des personnes avec T21 et de sa perception par des auditeurs tout-venant présente aussi un intérêt théorique, en particulier relativement aux questions fondamentales de la perception multimodale de la parole et de l’implication du système moteur de l’auditeur dans sa perception.Dans cette thèse, nous repositionnons le trouble de l’intelligibilité des personnes avec T21 dans un cadre qui conçoit la parole comme un acte coopératif entre locuteur et auditeur. En contre-pied de l’attention traditionnellement portée sur le locuteur dans la recherche appliquée, nous nous intéressons aux recours de l’auditeur pour mieux percevoir la parole en partant de deux observations : (1) la parole T21 est peu intelligible auditivement ; (2) son intelligibilité est meilleure pour des interlocuteurs familiers que non familiers. Ces observations sont mises en relation avec deux résultats importants de la recherche sur la perception de la parole. Primo, en situation de communication face-à-face, en plus de l’information auditive, l’auditeur utilise aussi l’information visuelle produite par le locuteur. Cette dernière permet notamment de mieux percevoir la parole quand l’information auditive est altérée. Deuxio, la familiarisation à un type de parole spécifique entraîne une meilleure perception de celle-ci. Cet effet est augmenté par l’imitation de la parole perçue, qui activerait davantage les représentations motrices internes de l’auditeur.Cette mise en relation des difficultés spécifiques des personnes avec T21 avec la recherche sur la perception de la parole nous amène à formuler les questions suivantes. Compte-tenu des spécificités anatomiques orofaciales du locuteur avec T21 impactant ses gestes moteurs articulatoires, l’auditeur tout-venant bénéficie-t-il de la présence de l’information visuelle ? L’implication du système moteur dans la familiarisation à cette parole spécifique peut-elle aider à mieux la percevoir ? Pour répondre à ces questions, nous avons mené deux études expérimentales. Dans la première, en utilisant un paradigme classique de perception audio-visuelle de la parole dans le bruit, nous montrons que voir le visage du locuteur avec T21 améliore l’intelligibilité de ses consonnes de manière comparable à des locuteurs tout-venant. L’information visuelle semble donc relativement préservée malgré les spécificités anatomiques et physiologiques. Dans une deuxième étude, nous adaptons un paradigme de familiarisation avec et sans imitation pour évaluer si l’imitation lors de la perception auditive de mots produits par un locuteur avec T21 peut aider à mieux les percevoir. Nos résultats suggèrent que c’est le cas. Ce travail ouvre des perspectives cliniques et théoriques : l’étude de la perception de la parole produite par des personnes avec un conduit vocal et des mécanismes de contrôle atypiques permet d’évaluer la généralité des mécanismes de perception mis en avant avec des locuteurs typiques et d’en délimiter les contours. / Speech is the most widely used means of communication by humans. It allows people to express their needs, exchange thoughts with others and contributes to the construction of social identity. It is also a complex communication channel involving elaborate motor control in production and the ability to analyze sound sequences produced by a wide variety of speakers in perception. This complexity results in speech being often the most altered or difficult to acquire mode of communication for people whose sensorimotor systems are impaired. This is particularly the case for people with trisomy 21 (T21), a genetic syndrome inducing complex orofacial motor difficulties and alterations in the auditory and somatosensory spheres. While speaking is possible for most of these people, their intelligibility is always affected. Improving their oral communication is a clinical and social issue. The study of speech production by people with T21 and its perception by typical listeners is also of theoretical interest, particularly with regard to the fundamental issues of multimodal perception of speech and the involvement of the auditor's motor system in this perception.In this thesis, we reposition the intelligibility disorder of people with T21 in a framework that conceives speech as a cooperative act between speaker and listener. In contrast to the traditional focus on the speaker in applied research, we are interested in the listener's means to better perceive speech, based on two observations: (1) T21 speech is not very intelligible auditorily; (2) its intelligibility is better for familiar than unfamiliar interlocutors. These observations are linked to two important research results on speech perception. First, in a situation of face-to-face communication, in addition to auditory information, the listener also uses the visual information produced by the speaker. In particular, the latter makes it possible to better perceive speech when auditory information is altered. Secondly, familiarization with a specific type of speech leads to a better perception of it. This effect is increased by the imitation of the speech perceived, which would further activate the listener’s internal motor representations.This connection between the specific difficulties of people with T21 and research on speech perception leads to the following questions. Given the anatomical orofacial specificities of the speaker with T21 impacting his articulatory motor gestures, does the typical listener benefit from the presence of visual information? Can the involvement of the motor system in familiarizing oneself with this specific speech help to better perceive it? To answer these questions, we conducted two experimental studies. In the first one, we show that seeing the face of the speaker with T21 improves the intelligibility of his consonants in a way comparable to typical speakers, using a classical paradigm of audio-visual perception of speech in noise. Visual information therefore seems to be relatively preserved despite anatomical and physiological specificities. In a second study, we adapt a familiarization paradigm with and without imitation to assess whether imitation during the auditory perception of words produced by a speaker with T21 can help improve their perception. Our results suggest that this is the case. This work opens up clinical and theoretical perspectives: the study of the perception of speech produced by people with atypical vocal tract and control mechanisms makes it possible to evaluate the generality of the perception mechanisms put forward with typical speakers and to delimit their contours.

People with motor function disorders that cause dysarthric speech find difficulty using state-of- the-art automatic speech recognition (ASR) systems. These systems are developed based on non- dysarthric speech models, which explains the poor performance when used by individuals with dysarthria. Thus, a solution is needed to compensate for the poor performance of these systems. This thesis examines the possibility of quantifying vowels of dysarthric and non-dysarthric speech into codewords regardless of inter-speaker variability and possible to be implemented on limited- processing-capability machines. I show that it is possible to model all possible vowels and vowel- like sounds that a North American speaker can produce if the frequencies of the first and second formants are used to encode these sounds. The proposed solution is aligned with the use of neural networks and hidden Markov models to build an acoustic model in conventional ASR systems. A secondary finding of this study includes the feasibility of reducing the set of ten most common vowels in North American English to eight vowels only. / Graduate / 2021-05-11

京都大学 / 0048 / 新制・論文博士 / 博士(医学) / 乙第13325号 / 論医博第2193号 / 新制||医||1043(附属図書館) / (主査)教授 黒田 知宏, 教授 藤渕 航, 教授 別所 和久 / 学位規則第4条第2項該当 / Doctor of Medical Science / Kyoto University / DFAM

The current study examined the effects of bite blocks on articulatory kinematics when producing /r/ within a phrase. Participants consisted of 20 young adults (10 males, 10 females) with no speech, language or hearing disorders. Participants produced the carrier phrase, I say __ with the nonsense words /əri/ (high front vowel), /əræ/ (low front vowel), /əru/ (high back vowel), /ərɑ/ (low back vowel). A Northern Digital Instruments Wave electromagnetic articulograph measured the articulatory movements while the speaker produced the stimuli in two conditions (Pre bite block and post bite block). Bilateral bite blocks were made using Express dental putty, which is a silicone impression material, in order to create an inter-incisal gap of 10 mm. The hull area (i.e., a boundary enclosing the total distance the sensor traveled during the target phrase) of the data for each sensor (i.e., tongue back, tongue mid, tongue front, lower lip, mandibular central incisor) was calculated for the individual nonsense words /eɪərɑ/, /eɪəræ/, /eɪəri/, and /eɪəru/. Results revealed kinematic differences across vowel phrases and between genders. The hull area of the tongue and jaw were significantly different for the vowel phrases /eɪəræ/, /eɪəri/, and /eɪəru/ compared to /eɪərɑ/. The hull area for the jaw for /eɪərɑ/ was significantly larger than the other vowel phrases. The between-gender analyses showed larger hull areas for males than females. Different motor equivalent strategies for tongue movements were observed when speakers produced /eɪərɑ/ and there were individual differences in compensating for the presence of the bite block

京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第21391号 / 情博第677号 / 新制||情||117(附属図書館) / 京都大学大学院情報学研究科知能情報学専攻 / (主査)教授 河原 達也, 教授 黒橋 禎夫, 教授 壇辻 正剛, 准教授 南條 浩輝 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

The current study examined the effects of bite blocks on articulatory kinematics when producing diphthongs /ɑɪ/ and /ɑʊ/ within a phrase. Participants consisted of 20 young adults (10 males, 10 females) with no speech, language or hearing disorders. Participants produced the diphthongs in the carrier phrase Im an owl that hoots. A Northern Digital Instruments Wave electromagnetic articulograph measured the articulatory movements while the speaker produced the stimuli in two conditions (pre bite block insertion and post bite block insertion). Bilateral bite blocks were made using Express dental putty, which is a silicone impression material, in order to create a 10 mm inter-incisal gap. Marker distance, maximum speed, and jaw contribution to tongue movement for three sensors (tongue back, tongue mid, tongue front) were calculated for the diphthongs segmented from the carrier phrase. F1 and F2 transitions and rate were also calculated for each diphthong. Results revealed kinematic differences during diphthong production after the bite block was inserted. Tongue movements independent from the jaw increased after the bite block was inserted, especially during production of the diphthong /ɑʊ/. Bite block by gender interactions during production of the diphthong /ɑɪ/ revealed larger and faster initial movements for males. The results did not reveal any significant acoustic changes other than a longer transition duration. Kinematic adjustments were sufficient to maintain overall similar acoustic output before and after bite block insertion.

No description available.