Magister Artium - MA / Foetal Alcohol Spectrum Disorder (FASD) is a global problem that affects various communities. FASD denotes a pattern of abnormalities intermittently seen in children born to women who consume huge quantities of alcohol during pregnancy (Church & Kaltenbach, 1997). Church and Kaltenbach (1997) suggest that FAS may be one of the primary causes of hearing, speech and other language problems in children. The two main approaches used to determine the effects of FASD on language are standardised language test (using a statistical approach to test some or all four domains of language, namely, phonology, syntax, morphology and semantics) applied to close-ended questionnaire answers and, to some extent, narrative analysis (in the course of which researchers use wordless picture books to analyse narratives in order to determine the social-communicative characteristics of individuals with FASD). Although the use of standardized measures of language might be helpful to determine problematic areas in relation to the different language domains (Wyper & Rasmussen, 2011), they do not show the difficulty with social-communicative functions which these children might be facing (Coggins, Friet, & Morgan, 1998). On the other hand, while narrative analysis addresses an important level of language (discourse level), it does not foreground the inherently interactive nature of language use and the problems that may be associated with communicative interactions. These shortcomings, in turn, suggest possible limitations in the interventions intended to address the language needs of children with FASD. There is, therefore, a need for complementary approaches that offer a more rounded picture of language impairment in children with FASD. In this study, three approaches are used in identifying features of the speech of children with FASD against the backdrop of comparisons with features in the speech of normally developing children. Firstly, conversational analysis (applied to spontaneous, open-ended speech) is introduced as a means to determine the more social-interactive aspects of speech impairment in children with FASD. Secondly, measures of linguistic aspects of speech (the mean length of utterance, Index of Productive Syntax and the number of different word roots) designed specifically for spontaneous speech are employed (they are applied to the same spontaneous data as the conversational analysis data). Thirdly, the more traditional standardized language test measures applied to non-spontaneous speech are used (covering the four domains of syntax, phonology, semantics, and pragmatics). The study’s objectives are to (1) compare patterns in the interactive speech of FASD children and normally developing children; (2) explore the relationship between FASD children and normally developing children in relation to both spontaneous speech measures and standardized measures of language; and (3) compare the impact of the primary caregiver's level of education on testing through spontaneous measures versus standardised measures. Using data from 14 children in the Bellville suburb of Cape Town, South Africa, the study finds that, on the conversational analysis measures, children with FASD, in contrast to normally developing children, tend to obey fewer rules of turn-taking, to overlap less, to engage less in self-repair and to struggle with management and maintenance of topics. The study also finds that children whose scores on the standardized language tests (with non-spontaneous data) suggest they have no language difficulty, especially in terms of phonology, obtained scores in measures of spontaneous speech that indicated language difficulty. The study also found that the socio-economic status of caregivers was a credible explanation for certain features in the speech of children with FASD is very similar to features in the speech of normally developing children. This finding highlights the role of family setting in mitigating the effects of FASD. / National Research Foundation (NRF)

Le projet EPAC est le point de départ de nos travaux de recherche. Nous présentons ce contexte de travail dans notre premier chapitre.Dans un deuxième temps, nous nous intéressons à la tâche de transcription de la parole. Nous en exposerons quelques jalons, ainsi qu'un inventaire des corpus oraux disponibles aujourd'hui. Enfin, nous comparons deux méthodes de transcription : manuelle et assistée. Par la suite, nous réalisons une étude comparative de huit logiciels d'aide à la transcription. Cela afin de démontrer que, suivant les situations, certains sont plus indiqués que d'autres. Le codage des données est l'objet de notre quatrième chapitre. Peut-on facilement échanger des transcriptions? Nous démontrerons que l'interopérabilité est un domaine où beaucoup de travail reste à faire. Enfin, nous terminons par une analyse détaillée de ce que nous appelons la parole spontanée. Par différents angles, définitions et expériences, nous tentons de circonscrire ce que cette appellation recouvre. / Our researches are based upon the EPAC project. We develop this work context in our first chapter. Then, transcription task is presented. Some important dates and people are mentionned, as well as an inventory of available speech corpus. Also, assisted and manual transcription task are evaluated and compared. A comparative study of eight transcription tools is developed in the third chapter. It shows that depending on the transcription context (data size, type of annotations ... ), some are more useful than others. Encoding data is the next step of our work. Is it really easy to exchange some transcriptions? We will demonstrate that interoperability must be much more efficient than it currently is, in order to easily share transcribed data. At least, what we name spontaneous speech is precisely analysed. Thanks to several points of view, definitions and experiences, we try to get the precise meaning of this expression.

Les méthodologies de constitution de corpus linguistiques ont été amplement étudiées, mais sont moins abondantes quand il s’agit de corpus oraux ; ces méthodologies sont encore plus rares en ce qui concerne l’interlangue orale. Le projet CIL (Corpus Inter Langue), en cours de finalisation à l’Université Rennes 2 et sous la supervision de l’équipe d’accueil LIDILE (EA 3874), vise à la constitution d’un corpus de productions écrites et orales d’apprenants en FLE et ALE. Cette thèse concerne le corpus oral de FLE du projet global (CIL-FLE). Partant du constat que l’intérêt des linguistes pour la langue orale a systématiquement été en retard par rapport à celui porté à la langue écrite, nous nous intéressons dans un premier temps à l’étude de l’oralité dans différents domaines de la linguistique d’un point de vue historique et épistémologique. Le second chapitre est consacré à la linguistique de corpus de manière générale et au corpus en tant qu’objet linguistique en particulier. En ce qui concerne la linguistique de corpus, nous tentons de présenter les différentes méthodologies auxquelles les linguistes ont recours lorsqu’il s’agit de consulter des données : introspection, élicitation ou consultation de données authentiques. Le concept de corpus est ensuite analysé selon un ensemble de critères définitoires que nous étudions en détail, afin de proposer une définition du corpus linguistique. Le troisième et dernier chapitre est la mise en application des constats théoriques dans la constitution du corpus CIL-FLE : nous détaillons lesconstituants du corpus, les protocoles de collecte et d’archivage. C’est au protocole de transcription que nous nous intéressons en particulier, en insistant sur les difficultés de la transcription de l’interlangue. Le corpus CILFLE, qui représente environ 105000 mots, représente le fruit de ce travail et sera ainsi détaillé. / The need to design linguistic corpora to support research in linguistics has triggered the development of numerous studies exploring various approaches and methodologies regarding good practices for written corpus building. Fewer studies are available when it comes to spoken data and those that concern the interlanguage of learners are even rarer. The CIL project (Corpus Inter Langue), under completion at the University of Rennes2 and supervised by a research team specialising in the fields of linguistics and pedagogy (LIDILE), aims at building a large corpus of written and spoken productions in EFL and in FFL. This phd dissertation mainly focuses on the FFL (French as a Foreign Language) corpus (CIL-FLE).The first chapter of the thesis is dedicated to the study of oral speech as a linguistic object from both a historical and an epistemological perspective. The second chapter tackles the question of corpus linguistics generally speaking as well as the concept/ notion of corpus as a linguistic object. Regarding corpus linguistics, we will review and explore the diverse approaches and methods that are used so as to carry out research enquiries: introspection, elicitation or consultation of authentic data. The concept of corpus is then analysed according to/following a series of criteria which we will closely examine in order to propose a definition of the linguistic corpus. The third and last chapter will implement the former theoretical findings through the description of the CIL corpus design. Thus, corpus constituents, transcription and archiving protocols will be described in detail. We are particularly interested in the transcription protocol and we will insist on the difficulties encountered when attempting to transcribe learners ‘data. Finally, the CIL-FLE corpus, which contains approximately 105 000 words and was developed all along this phd, will be described.

Memory deficits in Alzheimer’s disease (AD) and mild cognitive impairment (MCI) can be reflected in language-based tests, especially spontaneous speech tasks. Three spontaneous speech tests were developed in this study, including Thai Picture description (TPD), Thai Story Recall (TSR), and Semi-structured Interview for Thai (SIT) Ninety-eight Thai older adults underwent screening tests and three spontaneous speech tests. Then they were classified into three groups, including healthy control (HC), MCI, and AD. Their verbal responses were extracted into the content variables and acoustic features. Then the discriminant ability and accuracy in differentiating HC, MCI, and AD were examined with by Multivariate Discriminant Analysis (MDA) and analysis of the ROC curve and AUC. Two content variables showed significant differences among three groups of participants, i.e., correct information unit (CIU) of the TPD and delayed recall scores of the TSR. For acoustic features, ANOVAs revealed that three variables were significantly different among the three experimental groups, i.e., total utterance time in delayed recall, number of voice breaks in the TPD, and the SIT. The result of a stepwise estimation in MDA presented that the best combination of predictive model was CIU and backward digit span (BDS), in which provides 61.1% of correct classification. This discriminant function showed AUC of .81 in differentiating HC and MCI, AUC of .91 in distinguishing HC and AD, and AUC of .86 in detecting persons with cognitive impairments (MCI and AD) from HC. In conclusion, the combination of CIU of TPD and BDS is suitable for differentiating AD and persons with cognitive impairments from HC. However, there is no appropriate predictor in distinguishing MCI and AD.

Cette thèse porte sur des analyses acoustiques et prosodiques du français à partir de grandes masses de données orales illustrant différents styles de parole (préparée et spontanée). Nous nous sommes intéressées aux attributs acoustiques et prosodiques qui pourraient caractériser la prononciation. En français, de nombreuses erreurs de reconnaissance automatique de la parole (RAP) sont dues à des mots fréquents homophones. Pour ces mots, la solution correcte dépend du modèle de langage. Une classification automatique (CA) a été effectuée pour discriminer deux paires homophones (‘et/est’ et ‘à/a’) par des propriétés acoustiques et prosodiques. Les résultats de la CA ont montré que le paire ‘et/est’ était plus dissociable. La CA par des attributs prosodiques et inter-segmentaux (15 attributs) s’est avérée aussi performante que celle utilisant la totalité des 62 attributs. Un test perceptif a été également effectué pour vérifier si les humains utilisaient eux aussi ces paramètres. Les résultats ont suggéré que des informations acoustiques et prosodiques pourraient être utiles pour effectuer un choix correct de mots dans des structures syntaxiquement ambigües. Ensuite, nous avons examiné des propriétés prosodiques globales aux niveaux du nom et du syntagme nominal. La comparaison entre mots lexicaux et grammaticaux a montré que la fréquence fondamentale (F0) montante et l’allongement vocalique de la dernière syllabe caractérisent les mots lexicaux, par opposition aux mots grammaticaux. Ainsi, le profil de F0 moyenne d’un syntagme nominal de longueur n pourrait être différent de celui du nom avec une valeur de F0 basse au début du syntagme. Les profils prosodiques peuvent être utiles pour localiser frontières de mots. Les résultats de ce travail pourront servir à localiser le focus et les entités-nommées par des classifieurs discriminants, et de manière plus générale à améliorer les techniques de localisation des frontières des mots pour la RAP. / This thesis focuses on acoustic and prosodic (fundamental frequency (F0), duration, intensity) analyses of French from large-scale audio corpora portraying different speaking styles: prepared and spontaneous speech. We are interested in particularities of segmental phonetics and prosody that may characterize pronunciation. In French, many errors caused by automatic speech recognition (ASR) systems arise from frequent homophone words, for which ASR systems depend on language model weights. Automatic classification (AC) was conducted to discriminate homophones by only acoustic and prosodic properties depending on their part-of-speech function or their position within prosodic words. Results from AC of two homophone pairs, et/est (and/is) and à/a (ton/has), revealed that the et/est pair was more discriminable. A selection of prosodic and inter-phoneme attributes, that is 15 attributes, performed as good results as with 62 attributes. Then corresponding perceptual tests have been conducted to verify if humans also use acoustico-prosodic parameters for the discrimination. Results suggested that acoustic and prosodic information might help in operating the correct choice in similar ambiguous syntactic structures. From the hypothesis that pronunciation variants were due to varying prosodic constraints, we examined overall prosodic properties of French on a lexical and phrase level. The comparison between lexical and grammatical words revealed F0 rise and lengthening at the end of final syllable on lexical words, while these phenomena were not observed for grammatical words. Analyses also revealed that the mean profile of a n length noun phrase could be different from that of a n length noun with a low F0 at the beginning of a noun phrase. The prosodic profiles can be helpful to locate word boundaries. Findings in this thesis will lead to localize focus and named-entity using discriminative classifiers, and to improve word boundary locations by an ASR post-processing step.

Master of Arts / Department of Modern Languages / Janice McGregor / The current study investigates phrase-initial parts of speech as found in intermediate German textbooks and compares these findings to utterance-initial parts of speech as found in spontaneous speech in German-language interactions. This is important, because learning and using German word order appears to be a struggle for German learners whose first language is English. Research has shown that possible word order realizations in a language are partly restricted by the parts of speech system of that language (Hengeveld, Rijkhoff, & Siewierska, 2004; Vulanovic & Köhler, 2009). This is important because English and German have different parts of speech systems (Hengeveld et. al., 2004; Hengeveld & van Lier, 2010). Doherty (2005) analyzed English to German translations of an international science magazine and found that almost every second sentence begins differently. Instead, this study looks at talk in contexts of use and compares these findings with textbook language because, in recent years, communicative approaches to language teaching have been adopted by a large number of US German language programs. One would thus expect that textbooks used in these classrooms would contain at least some input with constructions that are typical to contexts of use. The results of the study indicate that construction-initial parts of speech in textbooks and in contexts of use are quite different. These differences imply that if it is a communicative approach that is being promoted, textbook authors and German educators would do well to expose students to actual talk from contexts of use so that they might learn to make meaning based on considerations of context.

Orientador: Plinio Almeida Barbosa / Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Estudos da Linguagem / Made available in DSpace on 2018-08-11T08:38:59Z (GMT). No. of bitstreams: 1 Lucente_Luciana_M.pdf: 1562872 bytes, checksum: cecd73e055b7a0b7fcc5741425c84c3e (MD5) Previous issue date: 2008 / Resumo: O sistema de notação entoacional proposto para o português brasileiro (PB) inserido inicialmente no programa de investigação científica da fonologia entoacional, utilizava para a descrição da entoação do PB uma adaptação da notação que propõe este modelo, que consiste em uma notação linear, baseada em aspectos formais da curva entoacional. Este programa adota como pressuposto fundamental a representação fonológica de parâmetros acústicos contínuos, preservada pela ação de primitivos melódicos como o alvo/tom a ser atingido, o registro funcional dos tons e o alinhamento abstrato entre curva entoacional e material lingüístico. O emprego de tal sistema para a notação de um corpus de fala espontânea do PB, em substituição a um corpus de fala semi-espontânea, se mostrou pouco produtivo e apontou para a necessidade de um sistema que combinasse aspectos formais e funcionais em sua descrição. Esse fato justificou a inserção do sistema de notação entoacional do PB em um programa de investigação científica dinâmico-entoacional, que propõe uma notação dinâmica e funcional, tendo como pressuposto fundamental a ausência de representações fonológicas, apresentando como primitivos melódicos os contornos entoacionais, a gama de variação tonal e o alinhamento específico entre curva e material lingüístico. A notação entoacional definida para o PB nesse novo paradigma por meio do sistema DaTo ¿ Dynamical Tones of Brazilian Portuguese - propõe a descrição do foco na entoação segundo a noção de contornos/tons dinâmicos, ao invés de utilizar o conceito de tons isolados para tanto, assumindo que a entoação e o foco no PB são produzidos por um sistema dinâmico integrado, que alia produção e articulação / Abstract: The intonational annotation system proposed to Brazilian Portuguese (henceforth BP) was initially inserted in the intonational phonology research program. The BP annotation system used an adaptation of the phonological annotation, which consists in a linear annotation, based on formal aspects of the intonational curve. This research program adopts as fundamental basis the phonological representation of continuum acoustic parameters, preserved by the action of melodical primitives as targets, the functional register and the abstract alignment between the intonational curve and the linguistics material. The application of these parameters to the annotation of spontaneous speech in BP was not productive, and pointed out the necessity of a new annotation that combines formal and functional features on the description. This necessity justified the adoption of a new annotation in a new research program, the dynamical program. This paradigm adopts as fundamental basis the absence of phonological representations, presenting as melodic primitives the intonational contours, the speech range and the specific alignment between the intonational curve and the linguistic material. The BP dynamical intonational annotation, called DaTo ¿ Dynamic Tones of Brazilian Portuguese ¿, proposes the focus description following the concept of dynamic contours, rather than the isolated tones concept, assuming that the intonation and focus in BP are produced by a dynamical system that ally phonation and articulation / Mestrado / Mestre em Linguística

Orientador: Plínio Almeida Barbosa / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Estudos da Linguagem / Made available in DSpace on 2018-08-19T23:34:39Z (GMT). No. of bitstreams: 1 Merlo_Sandra_D.pdf: 6723815 bytes, checksum: b6ff391ab121e4b75d42eef04f5a7d28 (MD5) Previous issue date: 2012 / Resumo: Premissa: esta pesquisa partiu da premissa de que pausas demarcativas estão relacionadas ao planejamento conceitual e hesitações, à formulação linguística. O planejamento conceitual refere-se a um esquema abstrato do texto falado, composto pelas informações que o falante julga relevantes de acordo com sua meta comunicativa. A formulação linguística refere-se à seleção de lemas e sua organização em estruturas sintáticas e fonológicas. Se pausas demarcativas e hesitações estão relacionadas a processos tão cruciais para a produção falada, sua ocorrência não deve ser marginal e sua distribuição não deve ser aleatória ao longo do texto falado. Método: participaram da pesquisa dez sujeitos do sexo masculino, entre 20 e 34 anos, falantes nativos do português brasileiro, com alto grau de escolaridade e sem distúrbios de comunicação. Foram realizados cinco experimentos de fala semi-espontânea com as seguintes variáveis independentes: memória declarativa, memória operacional, macroplanejamento textual, tipos textuais e taxa de elocução. As variáveis dependentes (pausas demarcativas e hesitações) foram examinadas através de três medidas: proporção, duração e ciclos periódicos (p < 0,05). A variabilidade individual na manifestação das variáveis dependentes também foi avaliada. Resultados: em média, 24% do texto falado é composto por pausas e 21% por hesitações. Dois terços das pausas duram entre 0,5 e 1,5 s, enquanto dois terços das hesitações duram até 1 s. Todos os textos falados apresentam ciclos de pausas e de hesitações, sendo que dois terços dos ciclos de pausa apresentam períodos até 5 s, enquanto dois terços dos ciclos de hesitações apresentam períodos até 10 s. As séries temporais de pausas e de hesitações estão correlacionadas, de forma que mudanças nas séries de pausas precedem em 300 ms mudanças nas séries de hesitações. Apenas 15% dos ciclos de pausas e hesitações são sincronizados e a grande maioria está em oposição de fase. Todos os cinco experimentos modificaram a dinâmica temporal das pausas demarcativas: textos que exigem elaboração conceitual, análise de novas informações e decisões mais conscientes sobre o sequenciamento de informações aumentam a proporção, a duração e/ou o período dos ciclos de pausas. Dois dos cinco experimentos modificaram a dinâmica temporal das hesitações: textos novos e pouco familiares aumentam a duração das hesitações em relação a textos previamente conhecidos. A variabilidade individual também interfere na dinâmica das pausas e das hesitações, existindo sujeitos que produzem esses fenômenos em abundância, enquanto outros os produzem com parcimônia. Conclusões: os resultados obtidos confirmam a hipótese de que as pausas demarcativas estão relacionadas ao planejamento conceitual e as hesitações, à formulação linguística. Também confirmam que a ocorrência desses fenômenos é significativa e que apresentam distribuição periódica no texto falado. Adicionalmente, os resultados indicam que pausas e hesitações são fenômenos dinâmicos da língua, que emergem de acordo com as necessidades da tarefa e o estilo do sujeito / Abstract: Background: this investigation assumed that demarcation pauses are related to conceptual planning, while hesitation phenomena are related to language formulation. "Conceptual planning" refers to an abstract scheme of spoken text, constituted by the information that the speaker consider relevant to his/her communication intent. "Language formulation" refers to lemma selection and its organization in syntactic and phonological structures. Considering that pauses and hesitations are related to essential processes to spoken language production, their occurrence may not be insignificant and their distribution may not be random in spoken text. Method: subjects of this research were ten males, from 20 to 34 years old, native speakers of Brazilian Portuguese, highly educated, and free from communication disorders. Five experiments of semispontaneous speech were done; each one addressed one of the following independent variables: declarative memory, working memory, text macroplanning, text types, and speech rate. Dependent variables (pauses and hesitations) were analyzed according the following three measures: proportion, duration, and periodic cycles (p < 0.05). Individual variability was also analyzed. Results: on average, 24% of spoken texts are composed by pauses and 21% by hesitations. Two thirds of pauses last from 0.5 to 1.5 s, while two thirds of hesitations last until 1 s. Pauses and hesitations are periodically distributed in all spoken texts; two thirds of pauses cycles exhibit periods until 5 s, whereas two thirds of hesitation cycles exhibit periods until 10 s. Time series of pauses and hesitations are correlated; changes in time series of pauses occur 300 ms before changes in time series of hesitations, on average. Just 15% of pauses and hesitations cycles are synchronized and the big majority is in phase opposition. All five experiments affect temporal dynamics of pauses: texts that demand conceptual elaboration, analysis of new information, and active decisions about information sequencing increase pauses proportion, durations and/or period of cycles. Two of five experiments affect temporal dynamics of hesitations: less familiar texts increase hesitations' durations compared to more familiar texts. Individual variability also affects temporal dynamics of pauses and hesitations; there are subjects that produce a lot of pauses and hesitations, while others produce them in small quantity. Conclusions: results support the initial assumption that demarcation pauses are related to conceptual planning and hesitations to language formulation. Results also indicate that the occurrence of pauses and hesitations is significant and that they are periodically distributed in spoken texts. Besides, results indicate that pauses and hesitations are dynamic components of spoken language, arising according to tasks needs and to subject style / Doutorado / Linguistica / Doutor em Linguística

This study investigated the possibility of pitch and tone register differences between native speakers of Taiwan and Beijing Mandarin by means of an acoustic analysis of three speech styles. Speech styles included spontaneous interview, spontaneous descriptive, and controlled read sentential speech. Data analysis included long segments of recorded speech in order to discern any statistically significant pitch register differences between the two dialects. Speech style and read tones were also analyzed. Results suggest that tones produced in Taiwan Mandarin are in a slightly lower register than those produced in Beijing Mandarin. Surprisingly, speech style was not a significant predictor of pitch register in long segments of recorded speech. Despite a limited sample size, this research effectively promotes the inclusion of sociolinguistic variables such as dialect in the field of tone research.

No description available.