Literatura científica selecionada sobre o tema "Speech comprehensibility"

Bilingualism is nowadays a worldwide phenomenon due to different factors such as migration, education or political and geographical reasons. These factors have led to both individual and social bilingualism, which favor an increase of communicative encounters between native and non-native speakers of a certain language. The above situation has contributed to a growth of studies on second language acquisition. Some of these studies have focused on native speakers’ perception and understanding of the non-native speech. More specifically, perception and effectiveness of communication through the analysis of three dimensions; accentedness, comprehensibility, and intelligibility. Although these three constructs have been intensively studied, little is known about the effect of semantic context on them. To our knowledge, just two studies have analyzed the effect of semantic context on native speakers’ judgments of the three dimensions by using read material. Therefore, this thesis extends the research on the influence of semantic context over these three dimension when the auditory stimulus corresponds to spontaneous non-native speech. In this thesis, the results on accentedness, comprehensibility, and intelligibility tasks of 40 native Swedish speakers are presented. The findings showed that listeners’ perceived comprehensibility of the non-native speech is affected by the use of additional contextual information. That is, the listeners who were provided with additional visual information perceived the non-native speech as significantly easier to understand than those listeners who did not receive the extra contextual support. Furthermore, the results showed that accentedness and comprehensibility perception of the listeners is influenced by their actual understanding of the non-native speech. Finally, this thesis proves the difficulty of studying the effect of semantic context on listeners’ response to accentedness, comprehensibility, and intelligibility when using spontaneous non-native speech. The results indicate that more research on how semantic context influences the perception of extemporaneous non-native speech is needed.

Teaching & Learning<br>Ph.D.<br>The main purpose of this study was to investigate the effects of form-focused instruction (FFI) on prosody with or without self-assessment on the prosodic and global aspects of L2 speech by Japanese EFL learners using a pre-post design. In addition, native English speaking (NS) and non-native English speaking (NNS) raters with high levels of English proficiency were compared to examine the influence of raters’ L1 backgrounds on their comprehensibility ratings. Sixty-one Japanese university students from four intact English presentation classes participated in the study. The comparison group (n = 16) practiced making one-minute speeches in class (45 minutes x 8 times) without explicit instruction on prosody, while the two experimental groups (n = 17 for the FFI-only group; n = 28 for the FFI + SA group) received FFI on word stress, rhythm, and intonation, practiced the target prosodic features in communicative contexts, and received metalinguistic feedback from the instructor. In total, the experimental groups received six-hours of instruction in class, which was comparable to the comparison group. Additionally, the experimental groups completed homework three times; only the FFI + SA group recorded their reading performance and self-assessed it in terms of word stress, rhythm, and intonation. Three oral tasks were employed to elicit the participants’ speech before and after the treatment: reading aloud, one-minute speech, and picture description. The speech samples were rated for comprehensibility by NS and NNS raters and were also analyzed with four prosodic measurements: word stress, rhythm, pitch contour, and pitch range. Instructional effects on prosody were observed clearly. The FFI-only group improved their controlled production of rhythm and pitch contour, while the FFI + SA group significantly improved all of the prosodic features except pitch range. Moreover, the instructional gains for the FFI + SA group were not limited to the controlled task but transferred to the less-controlled tasks. The results showed differential instructional effects on the four prosodic aspects. The FFI in this study did not help the participants widen their pitch range. The FFI on prosody, which was focused on the cross-linguistic differences between Japanese and English, tended to be more effective in terms of improving rhythm and pitch contour, which were categorized as rule-based, than an item-based feature, word stress. The study offered mixed results regarding instructional effects on comprehensibility. The FFI-only group did not significantly improve comprehensibility despite their significant prosodic improvements on the reading aloud task. Their significant comprehensibility growth on the picture description task was not because of the development of prosody, but of other linguistic variables that influence comprehensibility such as speech rate. The FFI + SA group made significant gains for comprehensibility on the three tasks, but the effect sizes were small. This finding indicated that the effects of FFI with self-assessment on comprehensibility were limited due to the multi-faceted nature of comprehensibility. The data elicited from the post-activity questionnaires and students’ interviews revealed that not all the participants in the FFI + SA group reacted positively to the self-assessment practice. Individual differences such as previous learning experience and self-efficacy appeared to influence the learners’ perceptions of the self-assessment practice and possibly their instructional gains. The two groups of raters, L1 English raters (n = 6) and L2 English raters with advanced or native-like English proficiency (n = 6) did not differ in terms of consistency and severity. These findings indicated that NNS raters with high English proficiency could function as reliably as NS raters; however, the qualitative data revealed that the NS raters tended to be more sensitive to pronunciation, especially at the segmental level, across the three tasks compared to the NNS raters. This study provides evidence that FFI, especially when it is reinforced by self-assessment, has pedagogical value; it can improve learners’ production of English prosody in controlled and less-controlled speech, and these gains can in turn contribute to enhanced L2 comprehensibility.<br>Temple University--Theses

Se faire comprendre en situation de communication, voire d'interaction orale, est essentiel au quotidien. La compréhensibilité est ainsi devenue un objectif important dans le domaine de l'apprentissage des langues, plus encore que d'avoir une parole sans accent étranger, proche d'un locuteur natif. Cependant, les enseignants et apprenants d'une langue étrangère (L2) ne disposent pas d'outils automatiques leur permettant d'évaluer de manière objective la compréhensibilité des productions orales. La compréhensibilité représente un concept linguistique influencé par des dimensions comme la phonologie/phonétique, la fluence, le lexique, la syntaxe et le discours. En plus de ces dimensions, elle peut également être influencée par le profil d'un apprenant (sa langue maternelle, ou L1, plus ou moins proche de la langue cible), le profil d'un auditeur (familiarisé ou non avec l'accent de l'apprenant) et la tâche de production orale pour les mettre en situation et collecter la parole des apprenants. Dans nos travaux de recherche nous nous sommes intéressés à la description de ces différentes dimensions dans la littérature. Nous avons ensuite implémenté différents paramètres considérés comme ayant une influence sur la compréhensibilité de la parole. Une première étape a été de valider leur adéquation lors d'une tâche de prédiction du niveau CECRL (Cadre Européen Commun de Référence pour les Langues) des apprenants du corpus CLIJAF. En se fondant sur ces paramètres linguistiques multi-niveaux, nous avons pu aborder la contribution principale de ce travail de thèse en proposant une méthode permettant de mesurer de manière automatique la compréhensibilité des apprenants. Afin d'évaluer la compréhensibilité, nous avons réalisé deux corpus : CAF-jp (Compréhensibilité d'Apprenants du Français - Japonais) et CAF-al (Compréhensibilité d'Apprenants du Français - Allemands). Ces corpus contiennent respectivement des productions orales de 40 apprenants japonais et 9 apprenants allemands de français. La mise en place d'un protocole de collecte a permis de collecter des productions orales. Ce protocole est basé sur une tâche de traduction orale, en L2, d'énoncés écrits en L1. Les énoncés ont été spécifiquement construits par des experts de FLE (Français Langue Étrangère) afin de contenir des difficultés typiques de traduction propres à chaque paire L1/français. Une fois la collecte des données effectuée, nous avons créé un protocole d'annotations nous permettant d'obtenir des évaluations subjectives de la compréhensibilité de la parole. Nous avons mené une campagne d'annotation auprès de 80 Français natifs et avons collecté 3920 scores de compréhensibilité, dont la moitié correspondent à la compréhensibilité a priori (compréhensibilité perçue) et l'autre moitié à la compréhensibilité a posteriori (compréhensibilité du sens du message véhiculé après prise en compte du réel sens du message à véhiculer). Afin de prédire automatiquement la compréhensibilité de la parole des apprenants, nous mettons en place une phase d'extraction de paramètres sur les productions orales. Ces paramètres sont d'ordre phonético-phonologique, lexical, syntaxique, discursif et sémantique. Nous obtenons d'excellents résultats de prédiction, aussi bien pour le corpus CAF-jp (r=0,97, MAE=0,15) que pour le corpus CAF-al (r=0,98, MAE=0,18), en utilisant l'algorithme Random Forest, une stratégie de fusion précoce et une validation croisée imbriquée de type leave-one-out. De plus, en entraînant un modèle sur la totalité des données du corpus CAF-jp et en testant sur les données du corpus CAF-al, nous obtenons également de bonnes performances (r=0,98, MAE=0,34), montrant ainsi la généricité de notre approche. Nos différents résultats montrent que notre méthodologie de prédiction de la compréhensibilité est tout à fait adaptée à l'évaluation de l'apprentissage du français L2, et pourrait même être appliquée à d'autres paires de langues L1/L2<br>Being understood in communication situations, even in oral interactions, is essential in everyday life. Comprehensibility has thus become a significant goal in the field of language learning, even more than having a speech without a foreign accent, close to that of a native speaker. However, teachers and learners of a foreign language (L2) lack automatic tools to objectively assess the comprehensibility of oral productions. Comprehensibility is a linguistic concept influenced by dimensions such as phonology/phonetics, fluency, lexis, syntax, and discourse. In addition to these dimensions, it can also be influenced by a learner's profile (native language, or L1, more or less similar to the target language), a listener's profile (familiarized or not with the learner's accent), and the oral production task to contextualize and collect learners' speech. In our research, we focused on describing these different dimensions in the literature. We then implemented various features considered to have an influence on the speech comprehensibility. A first step was to validate their adequacy during a task predicting the CEFRL (Common European Framework of Reference for Languages) level of learners in the CLIJAF corpus. Based on these multi-level linguistic features, we approached the main contribution of this thesis by proposing a method to automatically measure learners' comprehensibility. To assess comprehensibility, we created two corpora: CAF-jp (Comprehensibility of French Learners - Japanese) and CAF-al (Comprehensibility of French Learners - Germans). These corpora respectively contain oral productions from 40 Japanese learners and 9 German learners of French. The implementation of a collection protocol allowed us to gather oral productions. This protocol is based on an oral translation task, in L2, of statements written in L1. The statements were specifically constructed by experts in FFL (French as a Foreign Language) to contain typical translation difficulties for each L1/French pair. Once data collection was completed, we created an annotation protocol to obtain subjective evaluations of speech comprehensibility. We conducted an annotation campaign with 80 native French speakers and collected 3920 comprehensibility scores, half of which correspond to a priori comprehensibility (perceived comprehensibility) and the other half to a posteriori comprehensibility (comprehensibility of the message's meaning after considering the actual message's intended meaning). To automatically predict learners' speech comprehensibility, we implemented a feature extraction phase on oral productions. These features are phonetic-phonological, lexical, syntactic, discursive, and semantic in nature. We achieved excellent prediction results for both the CAF-jp corpus (r=0.97, MAE=0.15) and the CAF-al corpus (r=0.98, MAE=0.18), using the Random Forest algorithm, an early fusion strategy, and a nested leave-one-out cross-validation. Furthermore, by training a model on the entire CAF-jp corpus and testing it on the CAF-al corpus data, we also obtained good performance (r=0.98, MAE=0.34), demonstrating the generality of our approach. Our various results show that our methodology for predicting comprehensibility is well-suited for evaluating French L2 learning and could even be applied to other L1/L2 language pairs

碩士<br>國立臺北護理健康大學<br>聽語障礙科學研究所<br>102<br>This study aims to investigate the speech performance on the continuous speech sample of Mandarin-speaking Taiwanese adults. Two hundred and thirty-five healthy adults (M: 120; F: 115) from the northern, middle, southern and eastern Taiwan, aged above 20 years old (mean age: 55.31±19.19), have been recruited as the study participants. Based on the confounding factors, the participants were then subgrouped by age, sex, and the language of habitual use. They all required to read eighteen sentences in varied word length and a short essay under the simultaneous recording procedure. Two qualified speech pathologist blind in the study as the analyzers took charge of transcribing speech samples. Data thereafter were calculated and analyzed based on the study parameters including speech intelligibility and its efficiency, comprehensibility and its efficiency, and speaking rate (both in sentences and a short essay). The result showed a significant difference across all parameters among subgroups in age and the language of habitual use (p< .05), except sex. In the within-group comparisons, the speaking rate of sentences is faster than those in the essay’s reading (p< .05). The intelligibility, comprehensibility, and speaking rate are all found to be free of the length effect by the sentences. The longer sentences (11-13-word sentences) speech performance is computationally selected as the best predictor for intelligibility. In addition, intelligibility is the best predictor for comprehensibility. To sum up, the performance of intelligibility and comprehensibility of the healthy adults varies depending on the age and/or the language used in proficiency, but not by the gender and the length of reading material. Details would be discussed in the chapters. The results will provide information for clinicians while treating adult patients in judgments of speech assessment and as for the research references.

In the last two decades, second language (L2) pronunciation education and research have emphasised the need for intelligibility and comprehensibility over the traditional goal of achieving native-like fluency. However, it remains unclear what instructors should focus on to promote intelligibility and comprehensibility (e.g., segmentals vs. suprasegmentals) among beginning-level learners in the Saudi EFL context. Another uncertainty is how long their pronunciation gains might last after instruction. Although motivation has been significantly related to L2 pronunciation development, it is unclear how this factor is related to developing speech intelligibility and comprehensibility among beginning-level EFL learners. Learners' perspectives about pronunciation instruction have also been largely neglected, and studies have not elicited in-depth comments from native English-speaking raters about these learners' difficulties. The present study recruited 45 participants from a beginning-level EFL classroom at a Saudi university to complete a battery of pre-, immediate post-, and delayed post-tests. Participants were divided into three groups. The first experimental group (the segmental group) of 15 participants received explicit pronunciation instruction on segmentals. The second experimental group (the suprasegmental group) of 15 participants received explicit pronunciation instruction on suprasegmentals. The control group of 15 participants was given non-explicit pronunciation instruction. A series of mixed-effects models was fit in R to predict significant differences within and between the groups across tests. The segmental group's results for the immediate post-test revealed that intelligibility and comprehensibility improved significantly over the pre-test and control group, with the exception of the segment /ch/. However, the group failed to attain intelligibility and comprehensibility improvement over time, except for segment /v/. Regarding the suprasegmental group, the immediate post-test showed significant improvement in intelligibility and comprehensibility over the pre-test and control group. In addition, the suprasegmental group showed improvement over time, with the exception of rhythmic patterns in terms of intelligibility and stress in terms of comprehensibility. The suprasegmental group was more motivated than the segmental and control groups in the "intended efforts" and "ideal L2 self" dimensions, while participants in the control group were more motivated to learn English than the two experimental groups in the "ought-to L2 self" dimension. In addition, the suprasegmental group showed more positive attitudes toward learning English than the segmental and control groups. Finally, the experimental groups found the intervention helpful, despite it being relatively short. Taken together, this study contributes to the L2 pronunciation field by demonstrating that pronunciation instruction could be successfully incorporated into a beginning-level Saudi EFL classroom, according to the segmental and suprasegmental results. The delayed post-test results for the suprasegmental group suggested the effect of pronunciation instruction could be sustained over time. In addition, the findings showed L2 motivation to be significantly related to greater improvement in pronunciation. Based on these empirical results, the study provides a pedagogical description of how to improve pronunciation teaching for beginning-level learners in a Saudi EFL context.

This diploma thesis deals with the role of rhoticity in English pronunciation teaching. The theoretical part focuses on the characterisation of rhoticity as a distinctive feature of English accents. It describes and compares approaches to pronunciation teaching and the factors which influence the choice of pronunciation models in English language teaching. It also introduces the development of English as a lingua franca and its impact on pronunciation. Finally, it characterises the processes of speech perception. The practical part describes the creation and implementation of research which aimed at the discovery of the degree of intelligibility and comprehensibility of rhotic and non-rhotic samples of speech. The results imply that Czech learners of English find non-rhotic speech less intelligible and more difficult to understand. Key words rhoticity, non-rhoticity, pronunciation teaching, speech perception, intelligibility, comprehensibility

The purpose of the present thesis is to shed light on the role of pauses in simultaneous interpreting between French and Czech, mainly from the point of view of the listeners of the source speeches in French as well as both students of interpretation and professional interpreters and from the viewpoint of the listeners of the speeches interpreted into Czech. It is divided into a theoretical and an empirical part. The theoretical part briefly outlines the process of simultaneous interpreting and Daniel Gile's Effort Model and then it focuses on prosody and individual prosodic (suprasegmental) features, especially pauses and their use in simultaneous interpreting. The empirical part describes the experiment and presents its results. The experiment itself examines the effect that unnaturally long or unnaturally placed pauses in the speaker's quasi-authentic source speeches have on the choice of strategy and the simultaneous interpreters' performances (students and professionnals) and on the subsequent assessment of their performances made by the listeners of the interpreted speeches. Keywords simultaneous interpreting, pause, listener, speech, speaker, interpreter, comprehensibility