Stuttering is a developmental disorder which may adversely affect the individual on many functional and emotional levels. Common sequelae of speech disfluency include powerful emotions such as anxiety, shame, and anger, as well as speech-avoidant behaviors. For males, the influence of gender role socialization may present an additional burden. From a traditional perspective of masculinity, emotional expression and exposing one's flaws are strongly discouraged in most forms and contexts and may be seen as signs of weakness. While expression of emotions is a common developmental milestone for many who stutter, it is unclear what impact awareness, repression, and avoidance of emotions have on the well-being of people who stutter. This study explored the effects of disruptions in emotional expression and the influence of masculinity on the impact of disfluency for adult males who stutter (n=65). It was hypothesized that masculine-type emotional restriction would mediate the relationship between disruptions in emotional expression and the perceived impact of stuttering. Regression analysis revealed disruptions in emotional expression accounted for 25% of the variance in self-reported perceptions of stuttering, and self-regulation of emotion was negatively correlated with perceived impact of stuttering. Contrary to hypothesis, masculine-type emotional restriction was not significantly correlated with perceived impact of stuttering and thus invalidated impetus for mediation analysis. Implications and suggestions for further exploration are discussed.

The purpose of this study is to investigate disfluency and gesture in the second language under specific conditions such as familiarity vs. non-familiarity, concrete topic type vs. abstract topic type and speaking with native speaker vs. speaker with non-native speaker. The sample of this study was sixteen students from the Department of Basic English in Middle East Technical University (DBE), three instructors from DBE and one instructor from Modern Language Department in Middle East Technical University. Two of the instructors are native and the rest of them are non-native speakers of English. With an assigned instructor each student spoke on the following topics: making spaghetti, giving directions, spring festival and clashing midterms, and effects of religion on our life. The conversations on each topic were recorded audio-visually. Then the audio-visual data was annotated in terms of linguistic and gestural elements. In order to explore the relation among disfluency, gesture and controlled variables, quantitative data analysis methods were used. Levelt&rsquo / s speech production and Krauss&rsquo / s gesture production model were used as a basic framework. Dual Coding theory and Metalinguistic Awareness Theory was used to explain intricate results of the present study. As a result of the study, it was found that in the concrete topic condition, learners speak more fluently because of time and topic effects. Similarly, in the condition of familiar addressee and native speaker, learners speak more fluently than they do when speaking with a non-familiar or a non- native speaker.

Orientador: Ester Mirian Scarpa / Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Estudos da Linguagem / Made available in DSpace on 2018-08-11T21:45:21Z (GMT). No. of bitstreams: 1 Ramos_Susy_M.pdf: 465198 bytes, checksum: 4bbfefc00139aaf15691874ac4aac39d (MD5) Previous issue date: 2008 / Resumo: Na maioria dos trabalhos sobre linguagem as disfluências são tratadas pelo viés patológico e há poucas pesquisas que procuram elucidar suas ocorrências, dada a complexidade do tema. Este trabalho tem por objetivo estudar as hesitações na fala de uma criança no período de 1;9 até os 4 anos para analisar as ocorrências de pausas e repetições, itens que são considerados indícios de disfluência. A análise visa corroborar a hipótese de que as disfluências são intrínsecas à linguagem e ocorrem devido à mudança da posição do sujeito com relação à língua. Os momentos de hesitação são aqueles em que as dominâncias se alternam: dominância da língua, dominância da fala do outro e a dominância do sujeito. Dessa forma, as hesitações propiciam a progressão e a ancoragem para a continuidade do discurso, o que demonstra que não são apenas ocorrências intrínsecas à linguagem, mas necessárias. / Abstract: Disfluencies are generally seen through the bias of pathology in a great deal of research on the subject. The aim of this dissertation is to study the instances of hesitation in the speech of a child from 1;9 to 4 years old and to analyze the occurrence of pauses and repetitions, items that have frequently been considered traces of disfluency. The analysis aims at confirm the hypothesis that disfluencies are intrinsic to language and occur due to the change of position of the subject in relation to language. The moments of hesitation are the ones in which the dominance in a structure of three poles alternate: dominance of the language, of the speech of the other, dominance of the subject. Hesitations, thus, yield to progression and anchorage to the continuity of discourse, which shows that they are not only intrinsec to language, but necessary to it. / Mestrado / Mestre em Linguística

Children with Tourette Syndrome (CWTS) frequently exhibit a high prevalence of disfluent speech behaviors which are often labeled stuttering. The present study analyzed the fluency characteristics of CWTS, in comparison to children who stutter (CWS) and typically developing peers (TDP). It was predicted that CWTS would be less fluent than TDP but more fluent than CWS. A related purpose was to explore whether differences existed in the pattern of disfluencies demonstrated by these groups. To this end, it was predicted that CWTS would demonstrate significantly lower proportions of stuttering-like disfluencies than CWS and significantly higher proportions of stuttering-like disfluencies than TDP. Participants included eight CWTS, eight CWS and eight TDP. Speech samples, collected during a narrative story telling task, were analyzed to determine whether significant differences in the type and frequency of disfluencies were evident between the groups. Results revealed that CWTS were significantly more fluent than CWS and that CWTS produced significantly lower proportions of stuttering-like disfluencies than CWS. Although not statistically significant, CWTS were twice as disfluent as TDP and CWTS produced significantly higher proportions of stuttering-like disfluencies than TDP. These findings confirmed that CWTS present with an atypical disfluency pattern which can be differentiated from that of CWS and TDP based on the total disfluency level and the proportion of stuttering-like disfluencies. / Communication Sciences

Given the increasing problem of fraud, crime, and national security threats, assessing credibility is a recurring research topic in Information Systems and in other disciplines. Decision support systems can help. But the success of the system depends on reliable cues that can distinguish deceptive/truthful behavior and on a proven classification algorithm. This investigation aims to identify linguistic cues that distinguish deceivers from truthtellers; and it aims to demonstrate how the cues can successfully classify deception and truth.Three new datasets were gathered: 202 fraudulent and nonfraudulent financial disclosures (10-Ks), a laboratory experiment that asked twelve questions of participants who answered deceptively to some questions and truthfully to others (Cultural Interviews), and a mock crime experiment where some participants stole a ring from an office and where all participants were interviewed as to their guilt or innocence (Mock Crime). Transcribed participant responses were investigated for distinguishing cues and used for classification testing.Disfluencies (e.g., um, uh, repeated phrases, etc.), hedging words (e.g., perhaps, may, etc.), and interjections (e.g., okay, like, etc.) are theoretically developed as potential cues to deception. Past research provides conflicting evidence regarding disfluency use and deception. Some researchers opine that deception increases cognitive load, which lowers attentional resources, which increases speech errors, and thereby increases disfluency use (i.e., Cognitive-Load Disfluency theory). Other researchers argue against the causal link between disfluencies and speech errors, positing that disfluencies are controllable and that deceivers strategically avoid disfluencies to avoid appearing hesitant or untruthful (i.e., Suppression-Disfluency theory). A series of t-tests, repeated measures GLMs, and nested-model design regressions disconfirm the Suppression-Disfluency theory. Um, uh, and interjections are used at an increased rate by deceivers in spontaneous speech. Reverse order questioning did not increase disfluency use. Fraudulent 10-Ks have a higher mean count of hedging words.Statistical classifiers and machine learning algorithms are demonstrated on the three datasets. A feature reduction by backward Wald stepwise with logistic regression had the highest classification accuracies (69%-87%). Accuracies are compared to professional interviewers and to previously researched classification models. In many cases the new models demonstrated improvements. 10-Ks are classified with 69% overall accuracy.

Difficulties in speech production are often marked by disfluency; fillers, hesitations, prolongations, repetitions and repairs. In recent years a body of work has emerged that demonstrates that listeners are sensitive to disfluency, and that this affects their expectations for upcoming speech, as well as their attention to the speech stream. This thesis investigates the extent to which delay may be responsible for triggering these effects. The experiments reported in this thesis build on an Event Related Potential (ERP) paradigm developed by Corley et al., (2007), in which participants listened to sentences manipulated by both fluency and predictability. Corley et al. reported an attenuated N400 effect for words following disfluent ers, and interpreted this as indicating that the extent to which listeners made predictions was reduced following an er. In the current set of experiments, various noisy interruptions were added to Corley et al.,'s paradigm, time matched to the disfluent fillers. These manipulations allowed investigation of whether the same effects could be triggered by delay alone, in the absence of a cue indicating that the speaker was experiencing difficulty. The first experiment, which contrasted disfluent ers with artificial beeps, revealed a small but significant reduction in N400 effect amplitude for words affected by ers but not by beeps. The second experiment, in which ers were contrasted with speaker generated coughs, revealed no fluency effects on the N400 effect. A third experiment combined the designs of Experiments 1 and 2 to verify whether the difference between them could be characterised as a context effect; one potential explanation for the difference between the outcomes of Experiments 1 and 2 is that the interpretation of an er is affected by the surrounding stimuli. However, in Experiment 3, once again no effect of fluency on the magnitude of the N400 effect was found. Taken together, the results of these three studies lead to the question of whether er's attenuation effect on the N400 is robust. In a second part to each study, listeners took part in a surprise recognition memory test, comprising words which had been the critical words in the previous task intermixed with new words which had not appeared anywhere in the sentences previously heard. Participants were significantly more successful at recognising words which had been unpredictable in their contexts, and, importantly, for Experiments 1 and 2, were significantly more successful at recognising words which had featured in disfluent or interrupted sentences. There was no difference between the recognition rates of words which had been disfluent and those which were affected by a noisy interruption. Collard et al., (2008) demonstrated that disfluency could raise attention to the speech stream, and the finding that interrupted words are equally well remembered leads to the suggestion that any noisy interruption can raise attention. Overall, the finding of memory benefits in response to disfluency, in the absence of attenuated N400 effects leads to the suggestion that different elements of disfluencies may be responsible for triggering these effects. The studies in this thesis also extend previous work by being designed to yield enough trials in the memory test portion of each experiment to permit ERP analysis of the memory data. Whilst clear ERP memory effects remained elusive, important progress was made in that memory ERPs were generated from a disfluency paradigm, and this provided a testing ground on which to demonstrate the use of linear mixed-effects modelling as an alternative to ANOVA analysis for ERPs. Mixed-effects models allow the analysis of unbalanced datasets, such as those generated in many memory experiments. Additionally, we demonstrate the ability to include crossed random effects for subjects and items, and when this is applied to the ERPs from the listening section of Experiment 1, the effect of fluency on N400 amplitude is no longer significant. Taken together, the results from the studies reported in this thesis suggest that temporal delay or disruption in speech can trigger raised attention, but do not necessarily trigger changes in listeners' expectations.

This report explored the effects of stereotype threat—i.e., the apprehension associated with the possibility of confirming a self-relevant negative stereotype—on the stigmatized group Latinos as they were interviewed about their academic achievements and career aspirations. Latino participants were exposed to a self-relevant negative stereotype in the news, an illegal immigrant crossing the Mexican-American border smuggling drugs, as a stimulus activating stereotype threat. The study used deception as participants were unaware of the connection between the news article and the interview thus ensuring stereotype threat activation. Latino participants in the illegal immigrant/criminal condition displayed more verbal disfluency and tentative language than those in the control condition demonstrating evidence for media’s ability to stereotype threat. / text

The move towards larger vocabulary Automatic Speech Recognition (ASR) systems places greater demands on language models. In a large vocabulary system, acoustic confusion is greater, thus there is more reliance placed on the language model for disambiguation. In addition to this, ASR systems are increasingly being deployed in situations where the speaker is not conscious of their interaction with the system, such as in recorded meetings and surveillance scenarios. This results in more natural speech, which contains many false starts and disfluencies. In this thesis we investigate a novel approach to the modelling of speech corrections. We propose a syntactic model of speech corrections, and seek to determine if this model can improve on the performance of standard language modelling approaches when applied to conversational speech. We investigate a number of related variations to our basic approach and compare these approaches against the class-based N-gram. We also investigate the modelling of styles of speech. Specifically, we investigate whether the incorporation of prior knowledge about sentence types can improve the performance of language models. We propose a sentence mixture model based on word-class N-grams, in which the sentence mixture models and the word-class membership probabilities are jointly trained. We compare this approach with word-based sentence mixture models.

Dans la première partie de cette thèse, nous présentons une nouvelle méthode de production de variantes de prononciations qui adapte des prononciations standards, c'est-à-dire issues d'un dictionnaire, à un style spontané. Cette méthode utilise une vaste gamme d'informations linguistiques, articulatoires et acoustiques, ainsi qu'un cadre probabiliste d'apprentissage automatique, à savoir les champs aléatoires conditionnels (CAC) et les modèles de langage. Nos expériences poussées sur le corpus Buckeye démontrent l'efficacité de l'approche à travers des évaluations objectives et perceptives. Des tests d'écoutes sur de la parole synthétisée montrent que les prononciations adaptées sont jugées plus spontanées que les prononciations standards, et même que celle réalisées par les locuteurs du corpus étudié. Par ailleurs, nous montrons que notre méthode peut être étendue à d'autres tâches d'adaptation, par exemple pour résoudre des problèmes d'incohérences entre les différentes séquences de phonèmes manipulées par un système de synthèse. La seconde partie de la thèse explore une nouvelle approche de production automatique de disfluences dans les énoncés en entrée d'un système de synthèse de la parole. L'approche proposée offre l'avantage de considérer plusieurs types de disfluences, à savoir des pauses, des répétitions et des révisions. Pour cela, nous présentons une formalisation novatrice du processus de production de disfluences à travers un mécanisme de composition de ces disfluences. Nous présentons une première implémentation de notre processus, elle aussi fondée sur des CAC et des modèles de langage, puis conduisons des évaluations objectives et perceptives. Celles-ci nous permettent de conclure à la bonne fonctionnalité de notre proposition et d'en discuter les pistes principales d'amélioration. / In numerous domains, the usage of synthetic speech is conditioned upon the ability of speech synthesis systems to generate natural and expressive speech. In this frame, we address the problem of expressivity in TTS by incorporating two phenomena with a high impact on speech: pronunciation variants and speech disfluencies. In the first part of this thesis, we present a new pronunciation variant generation method which works by adapting standard i.e., dictionary-based, pronunciations to a spontaneous style. Its strength and originality lie in exploiting a wide range of linguistic, articulatory and acoustic features and to use a probabilistic machine learning framework, namely conditional random fields (CRFs) and language models. Extensive experiments on the Buckeye corpus demonstrate the effectiveness of this approach through objective and subjective evaluations. Listening tests on synthetic speech show that adapted pronunciations are judged as more spontaneous than standard ones, as well as those realized by real speakers. Furthermore, we show that the method can be extended to other adaptation tasks, for instance, to solve the problem of inconsistency between phoneme sequences handled in TTS systems. The second part of this thesis explores a novel approach to automatic generation of speech disfluencies for TTS. Speech disfluencies are one of the most pervasive phenomena in spontaneous speech, therefore being able to automatically generate them is crucial to have more expressive synthetic speech. The proposed approach provides the advantage of generating several types of disfluencies: pauses, repetitions and revisions. To achieve this task, we formalize the problem as a theoretical process, where transformation functions are iteratively composed. We present a first implementation of the proposed process using CRFs and language models, before conducting objective and perceptual evaluations. These experiments lead to the conclusion that our proposition is effective to generate disfluencies, and highlights perspectives for future improvements.

No description available.