Doctor of Philosophy / Department of Psychological Sciences / Heather R. Bailey / Bilingualism refers to an ability to speak two or more languages and the daily experience involved in coordinating these two languages can have a strong effect on bilinguals’ cognition. For decades, research strongly supported the idea of bilingual advantage; however, recent studies have found no bilingual advantage. Not surprisingly, such conflicting findings raised concerns about the validity of previous research as well as several other methodological issues. For instance, simple cognitive tasks like the Simon task are commonly used in bilingualism research, but they may not best capture bilinguals’ daily experience using the two languages. Bilinguals are constantly suppressing one language while engaged in other tasks, which is better captured by complex working memory (WM) tasks. Most importantly, previous work has not empirically evaluated the effects of language mode on bilinguals’ cognitive performance. Language mode refers to the state of activation of each language. Bilinguals may be in monolingual mode if only one of their languages is activated, whereas they may be in bilingual mode if both of their languages are activated. Previous work has proposed that language mode can have an effect on performance. Thus, the main objective of this dissertation was to evaluate the possible effects of bilingualism on complex WM performance while controlling for language mode and various demographic variables. The Pretest provided initial evidence that language mode affected performance on a simple cognitive task like MPWI. Therefore, a similar language mode manipulation was used in the Main study while testing performance on several complex WM tasks (CSPAN, OSPAN, and RotSpan) and the Simon task for monolinguals, bilinguals in bilingual mode, and bilinguals in monolingual mode. No significant differences were observed between all bilinguals and all monolinguals on any of the measures. However, significant differences were observed once language mode was accounted for. That is, bilingual participants in bilingual mode outperformed both bilingual participants in monolingual mode and monolingual participants on measures of complex WM. Further, there were no differences between monolinguals and bilinguals in monolingual mode. Thus, being in monolingual mode and fully suppressing one language may require more inhibition resources than bilingual mode in which both languages are active, and as a result, there may be fewer resources left to complete the complex WM span tasks. Importantly, the current work shed light on the hotly debated issue of the existence of a bilingual advantage by identifying a third variable that may explain the conflicting results in the literature. That is, no bilingual advantage was observed, but the current data provide evidence of a bilingual mode advantage.

The research on bilingual language processing explores two main avenues of relevance to the present study: lexico-semantic access and cognitive control. Lexico-semantic access research investigates the manner in which bilingual individuals retrieve single words from their lexical system. Healthy bilingual individuals can manipulate their lexico-semantic access to accommodate settings in which code- or language-switching is expected. Alternatively, they can manipulate their lexico-semantic access to speak only their first (L1) or second (L2) languages. Cognitive control, also known as executive functioning, is closely related to lexico-semantic access. Specifically, bilingual individuals maintain and switch between their languages through a mechanism known as cognitive control. Both cognitive control and lexico-semantic access are important for language processing in healthy bilingual individuals as well as bilingual persons with aphasia (BPWA). However, the extent to which BPWA utilize each of these processes in the production of single words is still unknown. The present study used a method of verbal fluency in the form of a novel modified category generation task to assess the relative contributions of lexico-semantic access and cognitive control in bilingual healthy controls and BPWA.

Bakgrund och syfte. Ordflödestest undersöker en persons förmåga att generera så många ord som möjligt under en minut. Orden kan börja på en viss bokstav eller tillhöra en särskild kategori. FAS är ett vanligt ordflödestest där orden som ska genereras börjar på bokstäverna F, A och S. Exempel på kategoriska ordflödestest är djurflöde, som går ut på att säga så många djur som möjligt, samt verbflöde där instruktionen är att säga så många saker man kan göra (handlingar). Forskare har undersökt vilka mentala färdigheter som ligger till grund för ordflödesförmågan, framför allt planerande och reglerande (exekutiva) funktioner samt språklig förmåga. I forskning används ofta FAS och djurflöde för att undersöka olika delar av den språkliga förmågan. Även verbflöde har undersökts i dessa sammanhang, men inte i lika stor utsträckning. Däremot har verbflöde en större roll i forskning på planerande och reglerande funktioner hos äldre personer som har exempelvis Alzheimers eller Parkinson. Forskningen har lett till att ordflödestest används i kliniska sammanhang som en del i utredningen av dessa sjukdomar. I praktiken används även FAS och djurflöde vid utredningar av dyslexi, språkstörning och koncentrations-/hyperaktivitetssvårigheter (ADHD). Forskning har visat motstridiga resultat om vilka typer av ordflödestest som är nedsatta vid dyslexi och språkstörning, eller vilka mentala förmågor som är viktiga vid genomförande av ordflödestest. I denna studie undersöks FAS, djurflöde och verbflöde hos studenter inom högre utbildning med och utan dyslexi. Syftet är att utreda om ordflödesförmågan är nedsatt hos studenter med dyslexi. Studien undersöker om ordflöde kan bidra till att förklara spridningen i läsförmåga. Metod. I undersökningen deltog 42 studenter, varav 16 hade dyslexidiagnos och 26 kontroller utan dyslexidiagnos. Deltagarna genomförde test som undersöker läsförmåga, fonologisk förmåga, snabb benämning samt ordflöde av FAS, djur och verb. Resultat. Prestationen på ordflödesförmåga var signifikant nedsatt hos studenter med dyslexi jämfört med kontrollgruppen. En multipel regression med bakåteliminering genomfördes för att undersöka om FAS, djurflöde och verbflöde kunde förutsäga spridningen i läsförmåga när fonologisk medvetenhet och snabb benämning kontrollerades för. Regressionsanalysen visade att verbflöde, tillsammans med fonologisk medvetenhet, kunde förutsäga läsförmåga hos studenter med och utan dyslexi. Diskussion. Den nedsatta ordflödesförmågan hos studenter med dyslexi diskuteras utifrån faktorer som utbildning och andra mentala förmågor. Resultatet pekar på ett unikt samband mellan verbflöde och läsförmåga hos studenter med och utan dyslexi. De strukturer i hjärnan som aktiveras vid verbflöde är även strukturer som ligger till grund för andra mentala förmågor. Dessa mentala förmågor har även visats vara nedsatta hos personer med dyslexi. Generellt indikerar dessa nya fynd att verbflöde har en betydelse i förhållande till läsförmåga och dyslexi som behöver undersökas vidare. Resultatet diskuteras även utifrån ett kliniskt perspektiv. / Verbal fluency is commonly measured in cognitive assessments and has been shown to measure aspects of verbal ability and executive function, as well as to involve specific cortical areas during performance. Verbal fluency tasks, in which participants generate words during a given time limit, have been used in research and assessments of neurobiological disorders and impairments. Dyslexia is a neurobiologically based reading disorder that is characterized by difficulties in word decoding and spelling. Research on verbal fluency in individuals with dyslexia shows that semantic and letter fluency is impaired. However, studies show inconsistent results. This study examines performance on semantic fluency (animals), action fluency (verbs), and letter fluency (FAS) in 42 students with developmental dyslexia (DD, n = 16) and a control group with typical reading development (TD, n = 26). Participants also perform a test battery that measures reading and phonological abilities, amongst others. Additionally, it is examined if verbal fluency performance can contribute to predicting reading ability, when phonological awareness and rapid automatized naming (RAN) are taken into account. Results show that verbal fluency performance was impaired in the DD group, and that action and letter fluency were relatively more impaired than semantic fluency. A backward elimination regression showed that action fluency and phonological awareness were significant predictors of reading ability, together explaining 48 % of the variance. The impaired verbal fluency ability is discussed in relation to factors such as education and cognitive abilities. Further, the findings point to a possible unique connection between action fluency and reading ability in students, in addition to phonological awareness. The possibility that the relationship between action fluency and reading may be partly explained by common neurocognitive underpinnings is discussed. These novel findings indicate that action fluency has a pertinent role in reading ability and dyslexia, which should be further examined.

xiii, 92 p. : ill. A print copy of this thesis is available through the UO Libraries. Search the library catalog for the location and call number. / In recent years, state and national policy created the need for higher accountability standards for student academic performance. This increased accountability creates an imperative to have a formative assessment system reflecting validity in inferences about the effectiveness of instruction and performance on statewide large-scale assessments. Curriculum-based measurement (CBM) satisfies both functions. However, research shows the predictive power of oral passage reading fluency (PRF) diminishes in middle and high school. Because of the decreased predictive validity of PRF in the upper grade levels, additional reading CBMs should be explored. This study compares PRF and Vocabulary CBM data for all sixth grade students in a school district using two statistical procedures: correlation and regression. The correlation coefficients were moderately high among PRF, Vocabulary CBM, and the Reading test in Oregon Assessment of Knowledge and Skills (OAKS). A regression analysis indicated that the Vocabulary CBM explained more variance than PRF in predicting reading performance on OAKS. A second multiple regression analysis introduced three non-performance indicators (Gender, Attendance, and NCLB At-Risk), along with the two CBMs (Vocabulary and PRF). The second regression results revealed that Vocabulary again was more predictive than PRF, Gender, Attendance, or NCLB At-Risk. At-Risk status was the only non-performance indicator that was significant. All the findings have been discussed within the context of understanding reading skills using CBMs and their relation to performance on a large-scale test used for accountability. The findings have been framed as part of an information system that allows schools and districts to better tailor staffing, instruction, and schedules to student needs. Suggestions for future research also have been discussed, particularly in enhancing the predictions on large-scale test outcomes using a variety of CBMs. / Committee in charge: Gerald Tindal, Chairperson, Educational Methodology, Policy, and Leadership; Paul Yovanoff, Member, Educational Methodology, Policy, and Leadership; Keith Hollenbeck, Member, Educational Methodology, Policy, and Leadership; Jean Stockard, Outside Member, Planning Public Policy & Mgmt

xv, 145 p. : ill. (some col.) / This dissertation examines the role of affect in consumer judgments in two essays. The first essay explores the use of affect as a heuristic basis for judgments of the risks and benefits associated with new products. Current perspectives regarding the processes by which consumers make decisions about the adoption of innovations maintain that it is largely a cognitive process. However, the four studies that make up the first essay suggest that consumer assessments of the risks and benefits associated with product innovations are often inversely related and affectively congruent with evaluations of those innovations. The results support and extend previous research that has investigated the affect heuristic in the context of social hazards. The findings further indicate that more affectively extreme evaluations are associated with increasingly disparate assessments of risk and benefit. The results indicate that this relationship is consistent across a variety of products and product categories. Together, these findings challenge traditional conceptualizations of innovation adoption decision making and suggest that cognitive models alone are insufficient to explain innovation adoption decisions. The second essay investigates if processing fluency - the difficulty associated with processing information - may serve as an input to the affect heuristic and subsequent judgments of risk and benefit. Recently, Song and Schwarz investigated the relationship between differences in fluency and perceptions of risk. Their results suggested that fluency experiences influence risk perception through differences in familiarity and not as the result of fluency-elicited affect. The three studies included in the second essay re-examine those results in an effort to clarify the role of affect as a basis for perceptions of risk. The findings document a previously unreported reversal in preference for less fluent stimuli and suggest that fluency-elicited affect can explain the relationship between processing experiences and perceptions of risk. The results have important theoretical implications for our understanding of how people derive meaning from fluency experiences and for the role of fluency-elicited affect as a basis for judgments of risk and benefit. / Committee in charge: David Boush, Chairperson; Robert Madrigal, Member; Joan Giese, Member; Paul Slovic, Outside Member

The purpose of this study was to analyze the relationship between oral reading fluency (ORF) and Maze, two common Curriculum-Based Measures (CBMs), and the statewide large-scale assessment of reading in Oregon, the Oregon Assessment of Knowledge and Skills- Reading (OAKS-R). A sample of three cohorts of eighth-grade students in an Oregon school district was used to examine concurrent validity, predictive validity, and the relation between demographic characteristics, disability status, and socio-economic status and the ability to predict performance on the OAKS-R. Findings of the concurrent validity analysis revealed a moderately strong positive correlation between the OAKS-R and both ORF and Maze measures, with ORF demonstrating a slightly stronger correlation with the OAKS-R. Multiple regression analyses were used to analyze the predictive relationship between ORF and Maze and scores on the eighth-grade OAKS-R. Both ORF and Maze were statistically significant predictors of OAKS-R, demonstrating moderately positive relationships with scores on the state reading test. Although no interaction effect was found between disability classification or eligibility for free or reduced-price meals and the different CBMs, in relation to the OAKS-R, student disability status was negatively related to performance on the OAKS-R. The relationship between OAKS-R performance and low socio-economic status, as measured by eligibility for free or reduced-price lunch status, was not consistent across the cohorts. This finding is promising, as it indicates that there may be factors that schools can take advantage of to ameliorate the relationship between poverty and reading outcome measures for eighth-grade students. In light of study results, suggestions for future research, as well as implications for the field, are discussed. This study adds to the research literature documenting that ORF and Maze assessments provide schools with valuable information to predict student performance on statewide large-scale assessments of reading. With CBM data available early in the school year, schools can provide additional intervention as needed, potentially leading to improved end-of-year student performance on the OAKS-R.

The complexity of text has a number of implications for educators in the areas of instruction and assessment. Text complexity is particularly important in formative assessments, which utilize repeated, alternate, equivalent forms to capture student growth towards a general outcome. A key assumption of such tools is that alternate forms of the assessment are of equal complexity. Consequently, there is a need to better understand what variables contribute to text complexity and how they impact student performance. This study was designed to evaluate features of text that are not typically included in readability estimates but may contribute to the text complexity: text cohesion and genre. Currently, text complexity of oral reading fluency measures is often quantified using readability estimates. It is hypothesized that a factor generally excluded from readability estimates, text cohesion&mdashthe extent to which the text functions as a cohesive, meaningful whole&mdashcontributes to text variability and variability in student performance. This research evaluated the role of a type of text cohesion (referential cohesion) in text complexity by manipulating the cohesion of passages otherwise assumed to be of equal difficulty. Genre was also considered, as research suggests that genre may impact complexity ratings of texts. Passages were strategically selecting to capture four conditions&mdash1) informational text/low cohesion, 2) informational text/high cohesion, 3) narrative text/low cohesion, and 4) narrative text/high cohesion. Data were collected on reading rate, accuracy, and passage-specific reading comprehension Results were analyzed using two-way, univariate ANOVA with dependent observations. Results indicate effects for each of the dependent variables included in the design. For rate and accuracy, results indicate significant interactions between genre and referential cohesion; scores were significantly higher for high cohesion narrative text than low cohesion narrative text and high cohesion informational text. There was a significant main effect of genre on comprehension, with students performing significantly better on the comprehension measure for narrative texts than informational texts. Altogether, these results indicate direct effects of genre and referential cohesion on student reading performance and provide evidence that text cohesion may be a meaningful component of text complexity. / 2015-10-10

This study uses a predominantly quantitative approach to explore the effect of different error correction feedback mechanisms on students’ English as a Second Language writing (narrative and descriptive) amongst high school students in Botswana. A longitudinal, quasi-experimental design is used, with a control group that received no correction feedback while the experimental groups received direct, coded and uncoded feedback. Three hypotheses define the study in terms of fluency, correction success and accuracy development over time. No significant increases in fluency were found between the pretests and posttests. Correction success achieved by the three treatment groups when rewriting texts reflected the explicitness of the feedback, with the direct group highest, followed by the coded and uncoded groups. Findings were mixed on the important issue of accuracy development, although they strongly suggest that for spelling, any type of feedback is significantly better than none and that coded feedback is better than direct feedback despite the latter being more explicit. Students from all the treatment groups expressed similarly positive opinions on correction feedback. / Applied Language / M.A. (Spec. in Applied Linguistics)

Submitted by MAIRA ANELLI MARTINS null (maira.anelli@hotmail.com.br) on 2018-03-11T19:59:39Z No. of bitstreams: 1 TESE_MairaMartins_repositório.pdf: 4695898 bytes, checksum: 9d70a9ff88173b5ffc3de74205b0517e (MD5) / Approved for entry into archive by Satie Tagara (satie@marilia.unesp.br) on 2018-03-12T16:23:50Z (GMT) No. of bitstreams: 1 martins_ma_dr_mar.pdf: 4695898 bytes, checksum: 9d70a9ff88173b5ffc3de74205b0517e (MD5) / Made available in DSpace on 2018-03-12T16:23:50Z (GMT). No. of bitstreams: 1 martins_ma_dr_mar.pdf: 4695898 bytes, checksum: 9d70a9ff88173b5ffc3de74205b0517e (MD5) Previous issue date: 2018-02-28 / Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) / Observando que uma grande parte das dificuldades dos escolares se relaciona com a difícil tarefa da leitura, este estudo tem como objetivo realizar a tradução e adaptação cultural do Helping Early Literacy with Practice Strategies (HELPS), programa de intervenção com a fluência de leitura. A pesquisa foi realizada em dois estudos. O Estudo 1 teve por objetivo a tradução e adaptação cultural do Programa HELPS. O procedimento metodológico seguiu técnicas e literatura especializada na adaptação de instrumentos psicológicos de uma cultura para outra. O Estudo 2 teve por objetivo verificar a eficácia do Programa HELPS-PB em escolares do 3º ao 5º ano. Participaram desse estudo 30 escolares de ambos os gêneros, na faixa etária de 8 anos a 12 anos de idade – distribuídos em dois grupos, o grupo experimental, com 15 escolares que foram submetidos à intervenção e o grupo controle, com 15 escolares que não foram submetidos à intervenção (cinco escolares para cada ano, em cada grupo). Esses escolares foram submetidos a todos os procedimentos do Programa HELPS-PB, seguindo protocolo específico do Programa, com oito estratégias baseadas em evidências científicas: leituras cronometradas repetidas, modelação, correção de erros na frase, estímulos verbais para o escolar ler com fluência e para compreender o que leu, definição de metas, feedback do desempenho e o sistema motivacional por recompensa. Os escolares também foram submetidos a pré e pós-testagem com avaliação da fluência de leitura oral (número de palavras lidas corretamente e incorretamente por minuto) e compreensão leitora (teste com perguntas de múltipla escolha). A coleta de dados foi realizada na própria escola, com a aplicação de 30 sessões do Programa HELPS-PB em 10 minutos por dia, com a frequência de 3 dias por semana. Os resultados foram analisados estatisticamente e foi possível verificar que a média do desempenho dos escolares do grupo experimental nos escores de palavras lidas corretamente por minuto, após serem submetidos ao Programa HELPS-PB, foi superior na pós-testagem para todos os escolares. Os resultados também demonstraram que apesar dos escolares do grupo experimental não terem diminuído os escores de palavras lidas incorretamente por minuto, todos os anos escolares obtiveram no pós-teste escores dentro da meta, ou seja, não realizaram escores acima de 3 erros, que foi o critério máximo considerado para erros em uma leitura. Em relação a compreensão leitora os dados demonstram que somente o 3º ano do grupo experimental apresentou diferença estatisticamente significante entre a pré e a pós-testagem, enquanto que para o 4º e 5º ano observou-se apenas uma discreta diminuição dos erros. A hipótese do estudo foi confirmada; a partir dos resultados – com a aplicação do Programa HELPS-PB – os escolares não fluentes melhoraram o seu desempenho em leitura e – por consequência – o desempenho na compreensão leitora. Conclui-se que os procedimentos de adaptação e tradução do programa HELPS-PB foram adequados para aplicação com a população escolar brasileira a medida que levaram em consideração a cultura e a realidade desse país. / Observing that a large amount of problems of the students is related to the difficult task of reading, this study aims to perform the translation and cultural adaptation of Helping Early Literacy with Practice Strategies (HELPS), an intervention program towards reading fluency. The research was conducted in two studies. Study 1 aimed at the translation and cultural adaptation of HELPS Program. The methodological procedure followed techniques and literature specialized in the adaptation of the psychological instruments from one culture to another. Study 2 aimed to verify the effectiveness of HELPS-PB Program including students from 3rd to 5th grade. A total of 30 students, both genders, aged from 8 to 12 years old, were divided into two groups: the experimental group included 15 students who were submitted to intervention, and the control group, included 15 students who were not submitted to intervention (five students of each grade, in each group). These students were submitted to all the procedures of HELPS-PB Program, following a specific protocol of the Program, with eight strategies based on scientific evidence: repeated timed readings, modeling, correction of errors in the sentence, verbal stimuli for the students to read fluently and understand what they have read, definition of goals, performance feedback, and the motivational reward system. The students were also submitted to pre- and post-test with oral reading fluency assessment (number of words read correctly and incorrectly, per minute) and reading comprehension (multiplechoice test). Data collection was performed at the school, itself, with the application of 30 sessions of the HELPS-PB Program, 10 minutes per day, 3 days per week. The results were statistically analyzed and it was possible to verify that the mean performance of the students in the experimental group, concerning the scores of the correctly read words per minute, after being submitted to HELPS-PB Program, was higher during post-test for all students. The results also showed that even though the students in the experimental group did not decrease the scores of words read incorrectly per minute, all the school grades had post-test scores within the goal, that is, they did not score above 3 errors, which was the maximum criterion considered for errors in reading. Regarding reading comprehension, the data show that only the 3rd grade of the experimental group presented statistically significant difference between the pre- and posttest, while for the 4th and 5th grade, there was only a slight decrease in errors. The hypothesis of the study was confirmed; from the results - with the application of HELPS-PB Program - non-fluent students improved their performance in reading and - as consequence - the performance in reading comprehension. It can be concluded that the adaptation and translation procedures of HELPS-PB program was appropriate for applying to the Brazilian school population as they took into account the culture and reality of that country. / 140304/2014-2

Submitted by Mariane Carvalho Vischi (mazicarvalho@bol.com.br) on 2017-03-31T02:44:30Z No. of bitstreams: 1 DEFESA - VERSÃO FINAL MARIANE CARVALHO VISCHI.pdf: 6742229 bytes, checksum: 05af4a4bb57ab550fefe7fac1686e50e (MD5) / Approved for entry into archive by Juliano Benedito Ferreira (julianoferreira@reitoria.unesp.br) on 2017-04-06T17:14:36Z (GMT) No. of bitstreams: 1 vischi_mc_dr_arafcl.pdf: 6742229 bytes, checksum: 05af4a4bb57ab550fefe7fac1686e50e (MD5) / Made available in DSpace on 2017-04-06T17:14:36Z (GMT). No. of bitstreams: 1 vischi_mc_dr_arafcl.pdf: 6742229 bytes, checksum: 05af4a4bb57ab550fefe7fac1686e50e (MD5) Previous issue date: 2017-03-03 / Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) / O presente trabalho investiga e compara a disfluência na fala de pessoas gagas e não gagas. O primeiro tipo foi denominado de disfluência gaga e o segundo, de disfluência comum. De uma maneira geral, as disfluências são caracterizadas pelo não fluir normal dos segmentos sonoros da fala, devido à presença de hesitações e de interrupções durante o processo comunicativo. Inicialmente, foi feita uma revisão bibliográfica sobre a relação da fluência com a disfluência a partir de uma compilação dos principais estudos sobre o assunto. Na sequência, apresentamos os trabalhos da literatura especializada e da linguística para a disfluência gaga, bem como dos aspectos que tratam da articulação da fala e dos elementos suprassegmentais. A metodologia consistiu na gravação semiespontânea de seis informantes gagos e seis não gagos. Como parte da metodologia para a formação do corpus, foi pedido a eles que narrassem algum fato que tivesse marcado suas vidas. Não foi definida de uma temática específica. Os dados foram analisados auditivamente e acusticamente. Para a análise acústica utilizou-se o software PRAAT. Para análise e interpretação dos dados foram usados modelos teóricos referentes à fonologia autossegmental (Pierrehumbert,1980; Ladd, 1996, entre outros) e, como apoio teórico, o da fonologia funcional (Halliday, 1970; Cagliari, 2007). Também foram utilizados os trabalhos relacionados ao estudo da fluência e disfluência da fala de uma maneira geral (Scarpa, 1995; Cruz, 2009; Koch, 2009; Scarpa; Fernandes-Svartman, 2012; Merlo, 2015, entre outros), bem como da disfluência gaga (Van Riper, 1971; Friedman 1986; Azevedo, 2000, entre outros). A investigação acústica estuda os segmentos fonéticos, as pausas, as repetições, as hesitações, os alongamentos e os bloqueios de fala. A interpretação fonológica investiga o comportamento dos fonemas nos dois tipos de disfluências. Os resultados mostraram, em relação à análise da F0 e padrão da intensitade, diferenças pouco significativas para a comparação entre os dois tipos de disfluências. A duração foi o evento que mais distanciou os dois tipos de fala, sendo maior para os falantes com disfluência gaga. Em relação aos tipos de disfluências (repetições, inserções, hesitações, entre outros), tanto os informantes gagos quanto os não gagos realizaram o mesmo tipo de disfluência. A diferença entre eles deve-se à quantidade e à complexidade das disfluências. Nesse caso, os informantes gagos tendem a apresentar um maior número de disfluências por grupo tonal, bem como uma quantidade maior de disfluências complexas. Outra diferença entre os dois grupos está na duração e quantidade de pausas, que tende a ser maior para os informantes com disfluência gaga. A análise fonológica, por sua vez, mostrou que gagos e não gagos tendem a manter o padrão entoacional para as repetições, inserções e hesitações. Esses resultados confirmam a hipótese de que a gagueira estaria mais relacionada a uma questão articulatória que leva, por meio de uma empatia fonética negativa, a quebra na expectativa do interlocutor. Já a disfluência comum, por estar mais relacionada a fenômenos estilísticos, não é sentida da mesma maneira. Mesmo assim, o caminho mais adequado não é o de distanciar negativamente esses dois tipos de fala, uma vez que os eventos são os mesmos, o que muda é o modo como eles são refletidos e sentidos. / The present work investigates and compares the dysfluency speech of stuttering and non-stuttering people. The first type was called stutterer dysfluency and the second common disfluency. In general, dysfluencies are characterized by no expected flow of the phonetic segments of speech due to the presence of hesitations and interruptions during the communication process. A general review of the literature on the relationship between fluency and dysfluency is presented with comments. A special review presents works of linguists, discussing stutterer dysfluency in particular in relation to segmental and suprasegmental articulatory problems. The methodology consists of recordings of semi spontaneous speech by six stutterers and non-stutterers informants. As part of the methodology for the formation of the corpus, the informants were asked to report any fact that marked their lives. Data were analyzed auditorily and acoustically. For acoustic analysis, the PRAAT software was used. Theoretical models regarding autosegmental phonology (Pierrehumbert, 1980; Ladd, 1996 among others) and functional phonology (Halliday, 1970, Cagliari, 2007) were used for theoretical support. The study of fluency and dysfluency discusses in particular the works from Scarpa (1995); Cruz (2009); Koch (2009) and Merlo (2015). On the other hand, stutterer dysfluency reminds the works of Van Riper (1971); Friedman (1986); Azevedo (2000), among others. The acoustic analyses take into consideration the phonetic segments, pauses, repetitions, hesitations, speech stretching and blocking. The phonological interpretation investigates the phoneme behavior in both types of disfluency. The results revealed, in relation to the analysis of F0 and intensity standard, minor differences in comparing the two types of disfluency. The duration was the event that more distanced the two types of speech behavior, being higher for speakers with stutterer dysfluency. The difference between them is due to the amount of occurrences and the types of complex disfluencies. Stuttering informants tend to have a larger number of dysfluency by tonal group as well as a greater amount of complex dysfluency. The phonological analyses, in turn, showed that stuttering and non-stuttering tend to maintain the intonation pattern for repeats, insertions and hesitations. These results confirm the hypothesis that stuttering is more related to an articulation process which leads, through a negative phonetic empathy, to break the expectation of the interlocutor. On the other hand, common dysfluency which is more related to stylistic phenomena is also involved with the same kind of speech problems but in a different dimension. Indeed, the most appropriate conclusion is to set a scale for the events with the two types of disfluency in opposite positions. This scale is interpreted as such by the persons involved in the problem, being framed by different expectations. / CNPq: 141913/2013-4