This thesis investigates the relationship between speech production and speech perception in the early stages of phonological and lexical acquisition. Previous studies have mainly focused on independent investigations of speech production and perception abilities in language acquisition. This thesis connects the individual speech production capacities to the child's perception and is organized around three major studies: Study I explores methodological alternatives such as the combination of EEG and eye-tracking in different Swedish participant groups: adults, 17-month-olds, and 24-month-olds. Visual and auditory stimuli, as well as the connection between word production and word perception are explored. Study II investigates phonological capacities in terms of consonant inventory, percentage of correctly pronounced words, segmental errors, as well as phonological templates in relation to vocabulary size in a group of Swedish 18-month-olds. Study III studies the influence of the children's individual phonological and lexical capacities in speech production on their word recognition in a group of Swedish toddlers with a productive vocabulary size above 100 words. The general results show that children accept mispronounced word forms as appropriate word candidates when the word forms are related to their individual word production. The occurrence of segmental errors increases with vocabulary size, and phonological templates are more likely to be observed in children with a productive vocabulary size above 100 words. The results thus indicate an influence of the individual child's production on word recognition, and a relationship between phonological capacities and lexical knowledge. These insights contribute to theoretical debates in linguistics regarding the abstractness of phonological word form representations and reveal a closer relationship between production and perceptual abilities in toddlers than what has previously been shown.

Le travail présenté dans ce mémoire est sous-tendu par trois observations majeures. D'abord, de nombreux travaux ont mis en évidence que la parole n'était pas uniquement de nature auditive mais aussi visuelle. D'autre part, la prosodie, domaine de l'intonation, du rythme et du phrasé joue un rôle crucial en parole. Enfin, la deixis ou monstration est un phénomène au coeur de la communication parlée et de son acquisition par les jeunes enfants. Or celle-ci peut, entre autre, s'exprimer uniquement avec la parole : il est possible de « montrer de la voix » par la focalisation prosodique par exemple. Ces observations et constatations permettent d'émettre l'hypothèse que la focalisation contrastive prosodique se manifesterait non seulement par la modalité auditive ,comme il a déjà été largement exploré, mais aussi par la modalité visuelle. C'est la piste que les travaux de ce mémoire visent à explorer pour le cas particulier du français. Plusieurs analyses en production de la parole ont ainsi permis, grâce aux enregistrements de six locuteurs avec deux systèmes de mesure différents et complémentaires, de mettre en évidence les stratégies de signalisation visuelle de la focalisation. Il semble que les locuteurs produisent des indices articulatoires visibles selon deux stratégies principales : la stratégie de signalisation absolue et la stratégie de signalisation différentielle. Les analyses ont également permis de montrer que d'autres gestes faciaux non articulatoires (mouvements des sourcils et de la tête) pourraient être liés à la production de la focalisation mais de façon très variable non seulement d'un locuteur à l'autre mais aussi pour un même locuteur. Par ailleurs, des analyses parallèles en perception, ont permis de montrer que les indices visuels mis en évidence en production, étaient effectivement utilisés en perception et qu'ils permettent d'extraire l'information de focalisation quand la modalité auditive est indisponible ou dégradée. Il a été mis en évidence que les indices visuels identifiés en production correspondent au moins en partie à ceux utilisés en perception audiovisuelle. Ces travaux montrent ainsi que la focalisation contrastive en français est « visible » et est « vue ». Ces résultats permettent d'esquisser un modèle cognitif de la production et de la perception audiovisuelles de la focalisation contrastive en français.

The presented diploma thesis maps the situation of contemporary child's reading. It deals with the factors which form the child reader externally (especially family, school and libraries, or the projects supporting child's reading). The main part of the thesis presents an analysis of the contemporary poetry for children. It examines how the books can influence the readers and motivate them for further reading. For the text analysis, the books of poetry which were awarded in competitions of Zlatá stuha or Magnesia Litera in 2013-2017 were selected as representative samples, namely Poetický slovníček dětem v příkladech, Všelijaké řečičky pro kluky a holčičky and Moře slané vody by Radek Malý, Tetovaná teta by Daniela Fischerová, Vynálezárium by Robin Král and Hlava v hlavě by David Böhm and Ondřej Buddeus. When analyzing the illustrations, attention was paid to the selected books that won the art category of Zlatá stuha and the competition Nejkrásnější česká kniha roku (The Most Beautiful Czech Book of the Year) and to the books representing the model production of the publishing houses of Běžíliška, Meander and series Raketa in publishing house Labyrint. An aspect of illustration was analyzed in the concertina books Rekomando and Ferdinande! by Robin Král and concertina book O čem sní by Petr...

Submitted by Fabio Sobreira Campos da Costa (fabio.sobreira@ufpe.br) on 2016-08-01T12:02:48Z No. of bitstreams: 2 license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5) Dissert_RafaelAlves-BC.pdf: 4054416 bytes, checksum: b14eecff58d266c6c5fcccdab9b7bb1b (MD5) / Made available in DSpace on 2016-08-01T12:02:48Z (GMT). No. of bitstreams: 2 license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5) Dissert_RafaelAlves-BC.pdf: 4054416 bytes, checksum: b14eecff58d266c6c5fcccdab9b7bb1b (MD5) Previous issue date: 2016-02-25 / CNPq / Esta pesquisa investiga, sob a perspectiva das Teorias de Percepção da Fala em L2 (FLEGE, 1995; KUHL; IVERSON, 1995; BEST, 1995; BEST; TYLER, 2007) e da Teoria dos Sistemas Dinâmicos (LARSEN-FREEMAN, 1997; DE BOT; LOWER; VERSPOOR, 2007; ELLIS, 2008), os efeitos da instrução fonética explícita na produção dos fonemas fricativos sonoros labiodentais, alveolares e alveopalatais, /v,z,3/, e indícios de sua projeção na percepção do contraste de vozeamento entre suas homorgânicas surdas, /f,sj/, que formam pares mínimos em português, por hispanofalantes aprendizes de Português como Língua Estrangeira (PLE). Para tanto, analisamos duas turmas de PLE da Universidade Federal da Paraíba (UFPB) e da Universidade Federal de Pernambuco (UFPE), dos níveis de proficiência básico e avançado, que se encontravam em processo de imersão linguística. O protocolo de coleta de dados consistiu-se na leitura em voz alta de frases foneticamente balanceadas em que os fonemas / f,v,s,zj,3/ estivessem em posição de onset silábico inicial ou medial (CV) em três posições de tonicidade e de atonicidade diante, quando possível, das sete vogais orais do PB: [i,e,£,a,o,o,u]. Também utilizamos um Teste Auditivo-Perceptivo de múltipla escolha (Multiple Forced Choice), em que os informantes tiveram de assinalar uma das opções do par mínimo exposto a partir do estímulo auditivo recebido. O período de coleta de dados e a instrução fonética explícita ocorreram no período de 3 meses para apenas uma turma, a da UFPB, denominada Turma de Intervenção; a turma da UFPE foi a que não recebeu nenhum tipo de instrução, denominada Turma de Controle. Os dados coletados em ambas as turmas receberam tratamento acústico, através do programa Praat (BOERSMA; WEENINCK, 2006), que nos permitiu investigar as principais estratégias de pronúncia utilizadas pelos informantes nas três coletas de dados realizadas (Teste 1, Teste 2 e Teste 3). Os resultados mostraram, para a Turma de Intervenção, que mesmo em processo naturalístico (cf. ELLIS, 2008), altamente favorável à aprendizagem de L2, o informante tende a não perceber algumas características fonético-fonológicas presentes no input linguístico disponível e que o processo de aprendizagem de uma L2 é complexo, dinâmico e não linear, na medida em que os informantes oscilaram algumas vezes entre as formas “corretas”, que seriam aquelas produzidas por um nativo do PB, e entre os “erros gradientes” (ZIMMER; ALVES, 2011), que caracterizam gestos articulatórios próximos à L2. A mesma gradiência foi observada nos Testes Auditivo-Perceptivos, aplicados ao longo dos três testes. Contudo, os resultados mostraram que, em se comparando a TI e a TC, a instrução fonética explícita funciona como uma excelente ferramenta para o ensino de PLE na produção; indícios de sua projeção na percepção também foram constatados. / This research investigates, from the perspective of the L2 Speech Perception Theories (FLEGE, 1995; KUHL; IVERSON, 1995; BEST, 1995; BEST; TYLER, 2007) and the Dynamical Systems Theory (LARSEN-FREEMAN, 1997; DE BOT; LOWER; VERSPOOR, 2007; ELLIS, 2008) approaches, the effects of explicit phonetics instruction in the production of labiodental, alveolar and alveopalatal voiced fricatives, /v,z,3/, and its evidence on perception of the voicing contrast between the homorganic unvoiced /f,sj/, which forms minimal pairs in Portuguese, by Spanish leaners of Portuguese as Foreign Language (PFL). Thereby, we analyzed two groups of PFL at the UFPB and at the UFPE, in two kinds of proficiency levels (basic and advanced), which were at linguistic input process. The data collection protocol it was constituted by reading aloud phases phonetically balanced where the phonemes /f,v,s,zj,3/ were at the medial or initial syllabic onset (CV) in three stressed and unstressed positions against, when was possible, the seven oral Portuguese vowels: [i,e,£,a,o,o,u]. We used an Auditory Perception Test of multiple choices (Multiple Forced Choice), where the listeners had to point out, between the minimal pairs exposed, the hearing stimulus received. As well as the data collection period, the explicit phonetics instruction occurred in the period of three months just for one group, from the UFPB, named as Intervention Group; the UFPE group was the one that did not receive any phonetic instruction, named as Control Group. The data collected in both groups received an acoustic treatment, through the software Praat (BOERSMA; WEENINCK, 2006), allowing us to investigate the main pronouncing strategies used by learners at the three data collection conducted (Test 1, Teste 2 and Test 3). The results of Test 1 showed for both groups (Intervention Group and Control Group) that even in the naturalistic context (ELLIS, 2008) highly favorable to the L2 learning, the learners tends to not recognize some phonetics and phonologic characteristics being in the available linguistic input and that the learning process of the L2 seems to be complex, dynamic and nonlinear, because the learners sometimes oscillated between the correct forms, corresponding that one produced by a Brazilian native, and the called “gradient errors” (ZIMMER; ALVES, 2011), sharing articulatory gestures near to the L2. The same gradience was observed in the Auditory Perception Tests, applied along the three tests. However, the results showed that, comparing both Intervention and Control Groups, explicit phonetics instruction works as an excellent tool to the PFL teaching in the production; evidences in the perception also were found.

Made available in DSpace on 2016-04-28T18:24:13Z (GMT). No. of bitstreams: 1 Fabiola Dagostini Peleias.pdf: 3457328 bytes, checksum: 7dd5e2d586af21d038754992b5d71372 (MD5) Previous issue date: 2009-11-06 / Conselho Nacional de Desenvolvimento Científico e Tecnológico / This work is part of the sound acquisition area in second language (L2), specifically in English, and it aimed to investigate whether there was a relationship between production and perception of the voiceless interdental fricative pronunciation [θ] or not in a group of Brazilian native speakers (BNS), based on acoustic and phonetic analysis. It was stated that there is such a relationship on account of the BNS in the sample. According to the literature, Brazilian learners of English tend to produce the voiceless interdental fricative [θ] as [s], [f] and [t] ((LIEFF & NUNES, 1993). This is because they consider such sound to be similar to [s], [f] and [t], and as stated by Flege (1999) and Meador, Flege and MacKay (2000), similar sounds of L1 (native language, in this case, Brazilian Portuguese - BP) and L2 may cause difficulties for the learner. Three studies have been made: an analysis of the production of the 3 BNS as well as of the AMS (American Native speaker) concerning the sound [θ]; an analysis of the perception of the sound [θ] by the 3 BNS and, finally, an evaluation of the 3 BNS productions of the sound [θ] by 31 American Native judges. The results of the three analyses have been crossed, and it can be posed that they conform to the literature about the link between production and perception and the assimilation of sounds from L2 to L1. Moreover, with such results, it was possible to achieve the goal of this work. It was verified that, regarding the perception analysis, the three BNS tend to perceive the sound [θ] as [θ] or [t]. Regarding the production analysis, one of the BNS approached the ANS. Therefore, it can be stated that there is a relationship between this BNS perception and production. The two remaining BNS did not approach the ANS. Most of the time, those speakers perceived the sound [θ] as [t], and the same happened with their productions of the sound [θ]. Taking into account those figures, it can be stated that there is a relationship among their production and perception, even if they have not perceived the sound [θ] as it should have been done. The perception and production analyses results were compared with the evaluation of the 31 judges, and the same can be said concerning all BNS: one of the BNS approached the ANS concerning the sound [θ], and the two remaining ones did not approach the ANS. To sum up, it was observed that, in this particular sample, there is a relationship between their production and perception, and that two out of the three BNS assimilated the sound [θ] as [t] / Este trabalho situa-se na área de aquisição de sons de L2 (segunda língua), mais especificamente língua inglesa, e seu objetivo foi investigar se existe uma relação entre a produção e a percepção da pronúncia da fricativa interdental surda [θ] em um grupo de sujeitos nativos brasileiros, a partir da análise fonético-acústica. Defendia-se que há uma relação entre a percepção e a produção da pronúncia da fricativa interdental surda pelos sujeitos nativos brasileiros da amostra. Segundo a literatura, aprendizes brasileiros de inglês tendem a produzir a fricativa interdental surda da língua inglesa [θ] como [s], [f] e [t] (LIEFF e NUNES, 1993). Isso porque consideram que esse som é similar aos citados e, de acordo com Flege (1999) e Meador, Flege e MacKay (2000), sons similares de L1 (língua materna Português Brasileiro - PB)e L2 causam dificuldades para o aprendiz. Foram feitos três estudos: uma análise das produções dos 3 SNB (sujeitos nativos do PB) e do SNA (sujeito nativo americano) envolvendo a [θ]; uma análise de percepção do [θ] pelos 3 SNB e uma avaliação, por juízes nativo-americanos, das produções dos 3 SNB e do SNA envolvendo o [θ]. Os resultados dos estudos foram cruzados, e constatou-se que estão em consonância com a literatura sobre o vínculo entre produção e percepção e assimilação de sons em L2 para a L1, atingiram o objetivo e validaram a hipótese defendida. Os 3 SNB, no teste de percepção, tendem a perceber o som [θ], em sua maioria, como [θ] ou como [t]. No teste de produção, a SNB1 aproximou-se do SNA; para esse sujeito, há relação entre sua percepção e sua produção. Os SNB2 e SNB3 se distanciaram do SNA. Em sua maioria, perceberam os estímulos como plosiva alveolar surda [t], comportamento que também se refletiu em sua produção. Portanto, pode-se afirmar que há relação entre produção-percepção, mesmo não tendo percebido e produzido o som [θ], em sua maioria, de maneira correta. Os testes foram comparados às avaliações dos juízes norte-americanos, e o mesmo pode ser afirmado para todos os sujeitos: os estímulos do SNB1, em sua maioria, foram avaliados como fricativa interdental surda [θ], do SNB2 como plosiva alveolar surda [t] e do SNB3 também como plosiva alveolar surda [t]. Observou-se, nesta amostra, que houve a relação entre produção e percepção e que 2 entre 3 dos SNB assimilaram [θ] como [t]

Les salles de classe sont des milieux bruyants auditivement et visuellement. Au Québec, le bruit auditif dans les écoles s’avère une problématique qui règne depuis des dizaines d’années et qui reste encore d’actualité aujourd’hui. Le bruit environnant peut impacter négativement le bien-être et les apprentissages des élèves et le bien-être des enseignant·es. D’autant plus, souvent, la source principale de bruit dans les écoles provient de la parole, un comportement qui peut être contrôlé. Les élèves et surtout les enseignant·es surchargent leur système vocal pour pouvoir faire passer leur message oral à travers le bruit auditif, augmentant leur niveau d’effort vocal. Dans les salles de classe, les distractions visuelles telles que les affiches murales contribuent à l’encombrement visuel, qui peut être considéré comme du « bruit visuel ». Celui-ci peut également impacter négativement le bien-être des élèves et leurs apprentissages. Considérant la nature multimodale, ou audiovisuelle, de la parole, nous nous questionnions sur l’influence de l’encombrement visuel sur la parole. Ce mémoire de maîtrise, composé de deux études, avait pour but de 1) étudier la perception subjective des élèves québécois de leur salle de classe en relation avec des données acoustiques objectives, et 2) étudier l’effet de l’encombrement visuel sur l’effort vocal en termes de paramètres acoustiques (SPL, f0 et CPP) et de perception auto-rapportée de locutrices. Les deux études se sont basées sur des méthodes mixtes, intégrant des méthodes quantitatives similaires (ex. : collecte de mesures objectives acoustiques, analyses acoustiques) et des méthodes qualitatives similaires (ex. : collecte de données subjectives perceptuelles, analyse inductive de questions ouvertes). Les analyses quantitatives et qualitatives des deux études révèlent que bien que les résultats relatifs aux données objectives (mesures acoustiques) n’étaient pas statistiquement significatifs, les résultats relatifs aux données subjectives (perception) l’étaient. Par exemple, dans la première étude, les niveaux d’appréciation du son des élèves (EAE) n’étaient pas tous corrélés avec les SPLs, et dans la deuxième étude, le niveau d’encombrement visuel n’était pas corrélé aux paramètres acoustiques de SPL, f0 et CPP. Ces résultats mettent en lumière l’importance de considérer et d’inclure des données subjectives et perceptuelles dans les études portant sur les paysages sonores éducatifs et sur l’effort vocal, car ce sont des expériences vécues par des humains. / Classrooms are acoustically and visually noisy environments. In Quebec, auditory noise in schools is a problem that has existed for decades and is still relevant today. Background noise can negatively impact the well-being and learning of students and the well-being of teachers. Moreover, the main source of noise in schools comes from speech, a behavior that can be controlled. Students and especially teachers put pressure on their vocal system to be able to convey their oral message through auditory noise, increasing their level of vocal effort. In classrooms, visual distractions such as wall posters can be considered as visual clutter, or visual noise, an element that can also negatively impact students’ well-being and learning. Considering the multimodal, or audiovisual, nature of speech, we question the influence of visual clutter on speech. This master's thesis, composed of two studies, aimed to 1) study the subjective perception of Quebec students of their classroom in relation to objective acoustic data, and 2) study the effect of visual clutter on vocal effort in terms of acoustic parameters (SPL, f0 and CPP) and self-reported perception of female speakers. Both studies were based on mixed methods, integrating similar quantitative methods (e.g., collection of objective acoustic measurements, acoustic analyses) and similar qualitative methods (e.g., collection of subjective perceptual data, inductive analysis of open-ended questions). Quantitative and qualitative analyses of the two studies reveal that although the results relating to the objective data (acoustic measurements) were not statistically significant, the results relating to the subjective data (perception) were. For example, in the first study, students' Evaluation of the Acoustic Environment (EAE) were not all correlated with measured SPLs, and in the second study, the level of visual clutter was not correlated with acoustic parameters of SPL, f0 and CPP. These findings highlight the importance of considering and including subjective and perceptual data in studies of educational soundscapes and vocal effort, as the human experience is subjective.