Rozprawy doktorskie na temat „Telephone Conversations”

Thesis (M.S.) -- University of Maryland, College Park, 2006.
Thesis research directed by: Dept. of Electrical and Computer Engineering. Title from t.p. of PDF. Includes bibliographical references. Published by UMI Dissertation Services, Ann Arbor, Mich. Also available in paper.

Previous research has examined the discourse features of legitimate phone interactions including commodity trading calls (Firth 1994), emergency phone calls (Zimmerman 1992), company-initiated sales calls (Freed 2010), and call center telephone calls (Hultgren and Cameron 2010). This research describes the structure and operation of phone fraud calls, specifically calls in which scammers pretend to provide Windows technical service and support. In contrast to expectations in the work on social engineering that indicates that it operates on trust, this research finds that this bulk phone scam follows a rigid discourse structure in which the calls pass quickly to a “proof” phase that convinces the target that they need the service and then a “sales” phase that is surprisingly varied in its length. An analysis of lexical chains in the proof phase show how those lexical chains create cohesion and thus texture, creating an overall linguistic meaning between the target and the scammer even when neither understands the technical topic under discussion and creating a situation in which the target is faced with an urgent problem for which a solution is being sold.

This thesis is a proposal for a new approach to the investigation of the language of telephone conversations. It strives to integrate the framework of analysis advocated by conversation analysts (originally the first to examine these conversations) and that of pragmaticists, in an attempt to provide description and explanation of telephone behaviour. It is based on the examination of telephone talk in Ecuadorian Spanish and British English. Such an approach requires the examination of these conversations at four levels - the level of their underlying structure which is analyzed in terms of skeletons, paths and moves, the linguistic realization of moves in relation to features of context, the politeness orientations behind moves and their linguistic realizations, and finally, the socio-cultural motivations behind linguistic choices and politeness orientations. In addition, this thesis constitutes a study in cross-cultural communication in that it attempts to determine basic similarities and differences in the organization of telephone conversations in two languages and cultures, which can provide insights into broader differences in the linguistic and politeness systems of those two languages.

In this thesis we train and evaluate an Automatic Speech Recognition system for phone communication in Icelandic. We use Kaldi, an open source toolkit, to build both GMM-HMM and Neural Network based models for general speech recognition in Icelandic. A simple telephone based dialogue system is built to test the speech recognition model in a real world scenario by calling users with a simple back and fourth dialogue between the user and the system. The resulting Speech Recognition models offer improved results compared to baseline systems in terms of Word Error Rate and are found to be successful for use in telephone communication.
I denna uppsats tränar och utvärderar vi ett automatiskt taligenkänningssystem för telefonkommunikation på isländska. Vi använder Kaldi, ett ramverk med öppen källkod, så tränas både GMM-HMM och neurala nätverksbaserade modeller för generell taligenkänning på isländska. Ett telefonbaserat system byggs för att testa modellerna i ett verklighetstroget scenario. Det bygger på en enkel dialog mellan användaren och systemet. De resulterande taligenkänningsmodellerna visar sig vara framgångsrika vid användning inom telefonkommunikation.

The purpose of this study was multifold: it aimed to investigate similarities and differences between Chinese and English telephone conversations, to test the validity of the theoretical distinction between transactional talk and interactional talk, and to examine L2 learners' use of the target language. Comparisons along four dimensions were conducted: (1) across languages--Chinese and English, (2) across functions--transactional and interactional talk, (3) across settings--China and the U.S., (4) L1 vs. L2--English as native language and English as a second language. The data consist of natural telephone conversations in Chinese and in English recorded by eighteen female participants (native speakers of Chinese and native speakers of American English) and interviews with the participants. Four sets of data were analyzed: Chinese telephone conversations recorded in China, Chinese telephone conversations recorded in the U.S., English telephone conversations recorded in the U.S., and English telephone conversations recorded in the U.S. by native speakers of Chinese. The findings suggest that primary differences between Chinese and English telephone conversations occur in identification, phatic talk, and leave-taking. Transactional calls and interactional calls display variation in greeting, phatic talk, initiation of closing, and register. The comparison of the use of language between the two settings reveals differences predominantly in transactional calls. The examination of L2 discourse suggests that learners' communicative competence will be further enhanced with the promotion of sociolinguistic knowledge and pragmatic awareness of the communicative event.

Au cours des dernières années, l’apprentissage profond est devenu l’approche privilégiée pour le développement d’une intelligence artificielle moderne (IA). L’augmentation importante de la puissance de calcul, ainsi que la quantité sans cesse croissante de données disponibles ont fait des réseaux de neurones profonds la solution la plus performante pour la resolution de problèmes complexes. Cependant, la capacité à parfaitement représenter la multidimensionalité des données réelles reste un défi majeur pour les architectures neuronales artificielles.Pour résoudre ce problème, les réseaux de neurones basés sur les algèbres des nombres complexes et hypercomplexes ont été développés. En particulier, les réseaux de neurones de quaternions (QNN) ont été proposés pour traiter les données tridi- mensionnelles et quadridimensionnelles, sur la base des quaternions représentant des rotations dans notre espace tridimensionnel. Malheureusement, et contrairement aux réseaux de neurones à valeurs complexes qui sont de nos jours acceptés comme une alternative aux réseaux de neurones réels, les QNNs souffrent de nombreuses lacunes qui sont en partie comblées par les différents travaux détaillés par ce manuscrit.Ainsi, la thèse se compose de trois parties qui introduisent progressivement les concepts manquants, afin de faire des QNNs une alternative aux réseaux neuronaux à valeurs réelles. La premiere partie présente et répertorie les précédentes découvertes relatives aux quaternions et aux réseaux de neurones de quaternions, afin de définir une base pour la construction des QNNs modernes.La deuxième partie introduit des réseaux neuronaux de quaternions état de l’art, afin de permettre une comparaison dans des contextes identiques avec les architectures modernes traditionnelles. Plus précisément, les QNNs étaient majoritairement limités par leurs architectures trop simples, souvent composées d’une seule couche cachée comportant peu de neurones. Premièrement, les paradigmes fondamentaux, tels que les autoencodeurs et les réseaux de neurones profonds sont présentés. Ensuite, les très répandus et étudiés réseaux de neurones convolutionnels et récurrents sont étendus à l’espace des quaternions. De nombreuses experiences sur différentes applications réelles, telles que la vision par ordinateur, la compréhension du langage parlé ainsi que la reconnaissance automatique de la parole sont menées pour comparer les modèles de quaternions introduits aux réseaux neuronaux conventionnels. Dans ces contextes bien spécifiques, les QNNs ont obtenus de meilleures performances ainsi qu’une réduction importante du nombre de paramètres neuronaux nécessaires à la phase d’apprentissage.Les QNNs sont ensuite étendus à des conditions d’entrainement permettant de traiter toutes les représentations en entrée des modèles de quaternions. Dans un scénario traditionnel impliquant des QNNs, les caractéristiques d’entrée sont manuellement segmentées en quatre composants, afin de correspondre à la representation induite par les quaternions. Malheureusement, il est difficile d’assurer qu’une telle segmentation est optimale pour résoudre le problème considéré. De plus, une segmentation manuelle réduit fondamentalement l’application des QNNs à des tâches naturellement définies dans un espace à au plus quatre dimensions. De ce fait, la troisième partie de cette thèse introduit un modèle supervisé et un modèle non supervisé permettant l’extraction de caractéristiques d’entrée désentrelacées et significatives dans l’espace des quaternions, à partir de n’importe quel type de signal réel uni-dimentionnel, permettant l’utilisation des QNNs indépendamment de la dimensionnalité des vecteurs d’entrée et de la tâche considérée. Les expériences menées sur la reconnaissance de la parole et la classification de documents parlés montrent que les approches proposées sont plus performantes que les représentations traditionnelles de quaternions
In the recent years, deep learning has become the leading approach to modern artificial intelligence (AI). The important improvement in terms of processing time required for learning AI based models alongside with the growing amount of available data made of deep neural networks (DNN) the strongest solution to solve complex real-world problems. However, a major challenge of artificial neural architectures lies on better considering the high-dimensionality of the data.To alleviate this issue, neural networks (NN) based on complex and hypercomplex algebras have been developped. The natural multidimensionality of the data is elegantly embedded within complex and hypercomplex neurons composing the model. In particular, quaternion neural networks (QNN) have been proposed to deal with up to four dimensional features, based on the quaternion representation of rotations and orientations. Unfortunately, and conversely to complex-valued neural networks that are nowadays known as a strong alternative to real-valued neural networks, QNNs suffer from numerous limitations that are carrefuly addressed in the different parts detailled in this thesis.The thesis consists in three parts that gradually introduce the missing concepts of QNNs, to make them a strong alternative to real-valued NNs. The first part introduces and list previous findings on quaternion numbers and quaternion neural networks to define the context and strong basics for building elaborated QNNs.The second part introduces state-of-the-art quaternion neural networks for a fair comparison with real-valued neural architectures. More precisely, QNNs were limited by their simple architectures that were mostly composed of a single and shallow hidden layer. In this part, we propose to bridge the gap between quaternion and real-valued models by presenting different quaternion architectures. First, basic paradigms such as autoencoders and deep fully-connected neural networks are introduced. Then, more elaborated convolutional and recurrent neural networks are extended to the quaternion domain. Experiments to compare QNNs over equivalents NNs have been conducted on real-world tasks across various domains, including computer vision, spoken language understanding and speech recognition. QNNs increase performances while reducing the needed number of neural parameters compared to real-valued neural networks.Then, QNNs are extended to unconventional settings. In a conventional QNN scenario, input features are manually segmented into three or four components, enabling further quaternion processing. Unfortunately, there is no evidence that such manual segmentation is the representation that suits the most to solve the considered task. Morevover, a manual segmentation drastically reduces the field of application of QNNs to four dimensional use-cases. Therefore the third part introduces a supervised and an unsupervised model to extract meaningful and disantengled quaternion input features, from any real-valued input signal, enabling the use of QNNs regardless of the dimensionality of the considered task. Conducted experiments on speech recognition and document classification show that the proposed approaches outperform traditional quaternion features

Bakgrund: En stor del av distriktssköterskors arbete inom primärvården består av telefonrådgivning. Det krävs stor kunskap inom omvårdnad, kommunikation, medicin och teknik för att kunna göra en korrekt bedömning av patientens hälsotillstånd. Inom telefonrådgivning finns det flertalet faktorer som försvårar bedömningen. Syfte: Syftet med denna studie var att beskriva distriktssköterskors upplevelser av svåra samtal vid telefonrådgivning inom primärvården. Metod: I studien användes en kvalitativ innehållsanalys med induktiv ansats. Intervjuer genomfördes med åtta distriktssköterskor. Resultat: I resultatet framkom fyra kategorier vilka var: Svårt att bedöma via telefonen, Svårt att inte kunna tillgodose allas önskemål, Organisationsrelaterade problem samt Stöd i bedömningen med tillhörande nio underkategorier. Distriktssköterskorna upplever att svåra samtal kan uppstå i många olika situationer samt att dessa påverkar distriktssköterskorna känslomässigt och deras arbete inom telefonrådgivning. Konklusion: Denna studie tydliggör vad distriktssköterskor upplever som svåra samtal och att dessa påverkar dem känslomässigt. Resultatet kan bidra till att öka medvetenheten av vad distriktssköterskor upplever sig behöva för att lättare kunna hantera svåra samtal.
Background: A large part of the district nurses’ work in primary health care consists of telephone counseling. This requires extensive knowledge of nursing, communications, medicine and technology to be able to make an accurate assessment of the patient's health. In the telephone counseling there are several factors that complicate the assessment. Aim: The purpose of this study was to describe district nurses’ experiences of difficult conversations in telephone counseling in primary health care. Method: The study used a qualitative content analysis with inductive approach. Interviews were conducted with eight district nurses. Results: The result shows four categories: Difficult to assess through the telephone, Difficult to not be able to accommodate everyone, Organizational problems and Support in the assessment. District nurses feel that difficult conversations can occur in many different situations and that they affect the district nurses emotionally and their work with telephone counseling. Conclusion: This study clarifies what the district nurses’ experience as difficult conversations and that they affect them emotionally. The results can help to raise awareness of what the district nurses feel they need to manage difficult conversations easier.

Thesis (S.M.)--Harvard-MIT Division of Health Sciences and Technology, 2009.
Cataloged from PDF version of thesis.
Includes bibliographical references (p. 12).
Telephone linked care (TLC) is a telehealth intervention that has been shown to be effective in a variety of clinical settings. TLC is an interactive computerized telephone system. The system 'speaks' to patients, asking them questions and understanding their responses. There is logic built into the calls, so that a patient's response to a question will dictate the next question that is asked. This serves to personalize the call for each patient, and makes the conversation more realistic. All of the patients' responses are stored in a database. This database provides much opportunity for analysis because a single phone call contains many responses. Visualization is an important way of gaining insight into data. Visualization can make data easier to understand and process. Different aspects of data can be encoded in a visualization. The TLC data lends itself to visualization. By viewing each of the questions that the system asks as nodes, and connecting the nodes by the chronological order in which these questions are asked, a tree structure will reveal the conversational paths that are taken in the calls. By combining data from multiple calls and encoding them in this tree structure, new insights can be gained into the TLC data. For example, the frequency with which questions are answered in a particular way can be encoded to reveal the most common pathways through the tree. This paper describes a visualization application of TLC data which will allow researchers to gain new insights into the TLC conversations and into medical interviews in general.
by Penelope H. Cuevas.
S.M.

This thesis focuses on the social organisation of advice, as it unfolds in interactions between mothers and their young adult daughters on the telephone. The analysis is based on a corpus of 51 telephone calls from 5 different families. Advice giving is studied here using the methods of conversation analysis and discursive psychology. The main interest has been to consider the dimensions that are relevant to the potentially tricky action of advice giving, building on the dimensions of normativity and knowledge asymmetry that have already been identified in the literature. The less strictly institutionalised context studied here provides a relatively new arena for considering the array of issues that are relevant to advice giving. Indeed, this has provided a broad scope for specifying how recipiency is brought off in advice giving sequences and how the position of advice recipient is managed. The analysis begins by considering the different forms of advice that were found in the data and their affordances in terms of the recipient s next turn. Contingency is identified as an important dimension in advice giving and a range of resources are identified which build contingency into the advice in various ways and which provide the recipient with different degrees of optionality when responding to advice. The thesis then goes on to consider how recipients respond to advice and the sorts of issues that make relevant one response type over another. The analysis identifies the importance of affiliation and alignment when considering different types of advice response. Furthermore, it is shown that morality, activity type, and alignment to the recipient s position, are important features of why a particular response type is chosen over another. The final analytic chapter then considers how the potentially tricky action of advice giving is made relevant in the first place. It is shown that the choice between different forms of advice is related to local issues of entitlement and contingency. In considering these different components to advice giving, the analysis explicates an array of important issues in advice giving sequences including: knowledge asymmetry, normativity, entitlement, contingency, affiliation, alignment and morality as well as considering evidence to suggest that advice is a dispreferred action. The findings are discussed in terms of their implications for studying advice and promoting advice acceptance, as well as considering how we can begin to see relationality being constituted.

Cette thèse porte sur les systèmes de reconnaissance automatique des émotions dans la parole, dans un contexte d'urgence médicale. Elle aborde certains des défis rencontrés lors de l'étude des émotions dans les interactions sociales et est ancrée dans les théories modernes des émotions, en particulier celles de Lisa Feldman Barrett sur la construction des émotions. En effet, la manifestation des émotions spontanées dans les interactions humaines est complexe et souvent caractérisée par des nuances, des mélanges et étroitement liée au contexte. Cette étude est fondée sur le corpus CEMO, composé de conversations téléphoniques entre appelants et Agents de Régulation Médicale (ARM) d'un centre d'appels d'urgence français. Ce corpus fournit un ensemble riche de données pour évaluer la capacité des systèmes d'apprentissage profond, tels que les Transformers et les modèles pré-entraînés, à reconnaître les émotions spontanées dans les interactions parlées. Les applications pourraient être de fournir des indices émotionnels susceptibles d'améliorer la gestion des appels et la prise de décision des ARM ou encore de faire des synthèses des appels. Les travaux menés dans ma thèse ont porté sur différentes techniques liées à la reconnaissance des émotions vocales, notamment l'apprentissage par transfert à partir de modèles pré-entraînés, les stratégies de fusion multimodale, l'intégration du contexte dialogique et la détection d'émotions mélangées. Un système acoustique initial basé sur des convolutions temporelles et des réseaux récurrents a été développé et validé sur un corpus émotionnel connu de la communauté affective, appelé IEMOCAP puis sur le corpus CEMO. Des recherches approfondies sur des systèmes multimodaux, pré-entraînés en acoustique et linguistique et adaptés à la reconnaissance des émotions, sont présentées. En outre, l'intégration du contexte dialogique dans la détection des émotions a été explorée, mettant en lumière la dynamique complexe des émotions dans les interactions sociales. Enfin, des travaux ont été initiés sur des systèmes multi-étiquettes multimodaux capables de traiter les subtilités des émotions mélangées dues à l'ambiguïté de la perception des annotateurs et du contexte social. Nos recherches mettent en évidence certaines solutions et défis liés à la reconnaissance des émotions dans des situations "in the wild". Cette thèse est financée par la Chaire CNRS AI HUMAAINE : HUman-MAchine Interaction Affective & Ethique
This thesis explores automatic speech-emotion recognition systems in a medical emergency context. It addresses some of the challenges encountered when studying emotions in social interactions. It is rooted in modern theories of emotions, particularly those of Lisa Feldman Barrett on the construction of emotions. Indeed, the manifestation of emotions in human interactions is complex and often characterized by nuanced, mixed, and is highly linked to the context. This study is based on the CEMO corpus, which is composed of telephone conversations between callers and emergency medical dispatchers (EMD) from a French emergency call center. This corpus provides a rich dataset to explore the capacity of deep learning systems, such as Transformers and pre-trained models, to recognize spontaneous emotions in spoken interactions. The applications could be to provide emotional cues that could improve call handling and decision-making by EMD, or to summarize calls. The work carried out in my thesis focused on different techniques related to speech emotion recognition, including transfer learning from pre-trained models, multimodal fusion strategies, dialogic context integration, and mixed emotion detection. An initial acoustic system based on temporal convolutions and recurrent networks was developed and validated on an emotional corpus widely used by the affective community, called IEMOCAP, and then on the CEMO corpus. Extensive research on multimodal systems, pre-trained in acoustics and linguistics and adapted to emotion recognition, is presented. In addition, the integration of dialog context in emotion recognition was explored, underlining the complex dynamics of emotions in social interactions. Finally, research has been initiated towards developing multi-label, multimodal systems capable of handling the subtleties of mixed emotions, often due to the annotator's perception and social context. Our research highlights some solutions and challenges in recognizing emotions in the wild. The CNRS AI HUMAAINE Chair: HUman-MAchine Affective Interaction & Ethics funded this thesis

This study contributed to study how Chinese and Pakistani interactants in 10 telephone conversations, despite their different cultural backgrounds and different speech habits, managed miscommunications and misunderstandings in their communications. It focused on various aspects that hurdle the inter-subjectivity and progressivity of talk between the interlocutors during the telephone interactions. The purpose behind the data collection for this research on miscommunication and (mis)understanding was to obtain audio recordings of real life telephonic conversation in which misunderstanding or reduced understanding as well as not understanding occurred. Out of 10 audio-recordings, the researcher chose the ones in which the participants were Chinese and Pakistani interactants in such a way that the telephone communication always took place cross culturally i.e. between a Pakistani person on one side and a Chinese person on the other. They used English as medium of communication. The participants of the study were introduced with their initials for the purpose of anonymity. Although the conversational corpus had a small size, yet suitably designed analyses aiming at finding the misunderstandings with the help of clear-cut proofs and solid evidences in the transcripts; could yield the required results. The recordings were then transcribed using the Ian Hutchby and Robin Wooffitt’s transcription glossary.The following research questions were designed for the study:1. What is the background of the miscommunication in the recording?2. What proof do the transcripts provide for the miscommunication?3. How does miscommunication emerge?4. How is miscommunication negotiated by the speaker and the recipient?5. What makes an interactional unit a trouble source?Different softwares like Audacity, CLAN, ELAN and Sound Forge will be used for repeated listening and then transcribing purpose. Finally, the data will be analyzed using procedures used specifically in the field of CA.

Families of children with life-limiting prognoses are often their primary care providers. The child’s variable condition and changing care requirements can result in families’ uncertainty while managing their child’s care at home. This conversation analytic investigation of a paediatric palliative telephone-support line reveals how specialist clinicians care for the changing needs of children, while also caring for parents’ myriad practical, moral, and emotional needs by providing reassurance. Analysis of clinicians’ delivery of reassurance show that when parents report differing dimensions of uncertainty, clinicians recurrently respond by prioritising parents’ emotional support so that parents’ delivery of care can continue.

The present paper focuses on exploring the usage of communicative strategies (CSs) by business professionals who use business English as a lingua franca (BELF) in telephone negotiations. The purpose of the study is also to analyze the occurrence of misunderstanding and non-understandings. The data consists of nine naturally-occurring spoken ‘buy-sell’ negotiations between business professionals from seven different countries around Europe and Asia.  The negotiations summed up to 1 hour and 30 minutes of recorded material. The transcripts of these negotiations were analyzed in detail by using the qualitative method ‘conversation analysis’ (CA). The results show that business professionals use CSs as pre-empting measures to prevent misunderstanding. In addition, CSs in the present data appear to be used to create cooperativeness and subsequently reach successful communication. The present study also found two scenarios where communicative breakdown occurred but were subsequently repaired through ‘clarity requests’, ‘confirmation checks’ and repetition. Lastly, the present paper recognizes that previous knowledge of the other party is an effective aspect in keeping upbeat negotiations and reaching mutual understanding.

Submitted by Renata Lopes (renatasil82@gmail.com) on 2017-05-22T11:14:29Z No. of bitstreams: 1 robertoperobellideoliveira.pdf: 1998429 bytes, checksum: 203bdda37697c874ffb79787afe4ca75 (MD5)
Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2017-05-22T17:18:59Z (GMT) No. of bitstreams: 1 robertoperobellideoliveira.pdf: 1998429 bytes, checksum: 203bdda37697c874ffb79787afe4ca75 (MD5)
Made available in DSpace on 2017-05-22T17:18:59Z (GMT). No. of bitstreams: 1 robertoperobellideoliveira.pdf: 1998429 bytes, checksum: 203bdda37697c874ffb79787afe4ca75 (MD5) Previous issue date: 2006-09-25
A presente dissertação enfoca os encerramentos de conversa telefônica cotidiana, na perspectiva da Análise da Conversa etnometodológica. Levando em consideração as contribuições de Schegloff e Sacks (1973), Button (1987) e Button (1990), dentre outras concepções teórico-metodológicas, este trabalho tem os seguintes objetivos: (1) mapear algumas formas lingüísticas que desempenham a função de encerrar a conversa telefônica cotidiana em contexto brasileiro e (2) evidenciar as estratégias de negociação que os participantes utilizam para encerrar ou para continuar o telefonema. O corpus se constitui de dados reais de fala espontânea, gravados em uma família juizforana, e traduzem um pouco como os participantes, em contexto brasileiro, constroem as relações sociais micro-seqüencialmente em conversas desse gênero. Enfim, esta pesquisa ratifica a noção de senso comum de que encerrar uma conversa é uma questão delicada para os participantes, uma vez que eles podem lançar mão de várias estratégias de negociação para encerrar ou continuar a conversa entre o primeiro “então ta” e o último “tchau”.
This study looks at the everyday telephone conversation closings, under the perspective of ethnomethodological Conversation Analysis. By the contributions of Schegloff & Sacks (1973), Button (1987) and Button (1990), among others theoretical-methodological conceptions, this work has the following aims: (1) to look into some linguistic forms that play the role of finish the everyday telephone conversation in Brazilian context and (2) to evidence the negotiation strategies used by the participants for closing or for continuing the call. The corpus is composed of data of talking occurring naturally, recorded in a family located at the Brazilian city of Juiz de Fora, and translates a bit how the parts, in Brazilian context, build the social relationship microsequentially in conversations like that. Thus, this study confirms the commonsense notion about finishing a conversation as a delicate issue for the participants, because they can use many negotiation strategies to closing or to continuing the conversation between the first “então tá” and the last “tchau”.

Este trabalho investiga a polidez lingüística nas negociações de telemarketing. Visa a demonstrar como se dá este funcionamento e pretende contribuir para a melhor compreensão do processo de polidez em ambientes muitas vezes hostis e sempre inesperados, como é o caso do telemarketing ativo. O corpus é composto pela transcrição de ligações telefônicas dos departamentos de telemarketing do BankBoston e SOS Computadores. No discurso informativo e de interesse comercial, característico do telemarketing, a polidez facilita a interação, quanto mais interesse comercial houver no contexto, mais importante se tornará usar de polidez. Já que a unilateralidade da situação não permite a abordagem face a face, a entonação da voz e outros índices serão fundamentais para demonstrar cortesia ou polidez.
This research aims to investigate linguistic politeness during telemarketing conversations and negotiations. The central concern is to demonstrate how this process works and contributes to a better understanding of linguistic politeness in hostile and unexpected environment - such is the case of active telemarketing approaches. The used corpus is composed by conversations among BankBoston Telemarketing Department and its juridical clients, as well as SOS Computadores and its prospect clients. Politeness helps commercial interaction, especially when the situation doesn’t allow face-to-face interaction. Voice intonation and other language resources will be fundamental to better promote the idea of linguistic politeness during telemarketing conversation.

This thesis is a conversation analytic investigation of the social organisation of talk in telephone and computer-mediated calls to NHS Direct, a telephone health helpline in England. The data represent fifty-six routinely audio recorded telephone consultations between nurses and callers between June 2003 and June 2004 at one NHS Direct call centre. Data were transcribed using the Jefferson (2004) transcription system. Data analysis follows the broad trajectory of the call. Chapter three illustrates the overall structural organisation of the call as mediated by the Clinical Assessment System (CAS); Chapter four examines how CAS prompted history taking questions are tailoured and delivered by the nurse; Chapter five examines the delivery by the nurse, of the CAS output in the form of the 'disposition' or course of action the caller may take to manage their concern, and Chapter six examines caller's responses to the disposition. The results draw attention to the complexities of telephone and computer-mediated help in which nurses and callers must design their talk to take account of the CAS as a 'third party'. Analysis reveals that nurses typically orient to the CAS output as potentially troublesome. First nurses regularly deviate from and modify CAS prompted questions which works to 'cushion' the system and build rapport between the nurse and the caller. Second nurses regularly simultaneously produce and labour to deny hearably candidate diagnoses. Third callers regularly respond to the CAS produced disposition as dispreferred. In conclusion, this research has revealed how nurses and callers employ a range of interactional practices which work to skilfully tailor and fashion 'embodied help' from an otherwise disembodied CAS technical system. Thus, we can observe nurses and callers artfully displaying through talk the ordinary practical methods for accomplishing telephone and computer-mediated help in this setting.

UNE ANALYSE DE L’ENTAME CONVERSATIONNELLE DE COMMUNICATIONS ORALES ET ECRITES (SUR RÉPONDEUR TÉLÉPHONIQUE ET INTERNET)

MIREILLE FALESSE

ULB - FACULTÉ DE PHILOSOPHIE ET LETTRES - FÉVRIER 2005

La trame de base de l’étude est essentiellement linguistique et la plupart des catégorisations des éléments relevés sont de cet ordre également, l’étude prenant en compte le langage sous son aspect pragmatique dans les limites de la présentation des situations particulières ainsi que du collationnement des données du corpus.

Deux types de messages ont été choisis :des messages oraux laissés par des appelants sur répondeur téléphonique et des messages – provenant de nouveaux utilisateurs – recueillis dans des forums de discussion sur Internet.

L’analyse permet de préciser les souhaits et intentions communicationnels des émetteurs ;d’autre part le relevé des éléments constitutifs du corpus auquel nous avons procédé dans la seconde partie nous a permis d’entrer plus avant dans sa description.

L’énonciation est à l’énoncé ce que le processus de fabrication est à l’objet produit ;l'énoncé est le résultat alors que l'énonciation est l'acte de création du locuteur. C’est cet acte, la procédure de construction du message, les intentions du locuteur, les marques de son intervention en tant que sujet parlant – ses pensées, ses intentions, ses émotions au moment de la « prise de parole » (orale ou écrite) – qui ont fait l’objet de notre propos. Dès lors, les éléments de base du schéma de la communication ont été posés et les particularités de notre corpus explicitées à la suite d’un double choix :celui des outils d’analyse réellement utiles à la démarche et celui des éléments essentiels et nécessaires constitutifs des énoncés retenus et à retenir.

L’énonciation et l’énoncé

Le travail porte sur l’analyse d’un certain type de discours à l’intérieur d’actes de communication sur base d’énoncés, produits d’un acte d'énonciation, qui comportent des marques énonciatives faisant référence à la fois au locuteur et à l'allocutaire.

Il en est tenu compte lors de la description du corpus car les éléments retenus portent non seulement sur la structuration phrastique de l’énoncé mais également sur le sens qui lui est donné ainsi que ses utilisations caractéristiques en fonction des intentions, choix et motivations des destinateurs.

Le sens et le son

La considération du langage sous sa double articulation favorise une analyse appariant les points de vue sémantique et phonologique :le sens et le son.

\
Doctorat en philosophie et lettres, Orientation linguistique
info:eu-repo/semantics/nonPublished

When crises develop, people are confronted with difficulties beyond those experienced in normal everyday activities.  Due to the perceived threats inherent to such situations, familiar behaviors may prove ineffective, and such attempts can pose dangerous and unpredictable risks. Crises are extreme situations, occurring at the very edges of human experience. Oral communication in such situations cannot be casual; the seriousness of the situation demands exceptional communicative performance on the part of the participants. Therefore, certainties about everyday communication conventions are called into question. The following work examines conversations during which the participants were involved in an extreme situation. In this particular crisis, a politically motivated kidnapping, the personal involvement of the interlocutors is substantial. A clear and present fear of the situation escalating and the possibility of a failure to anticipate the resulting reactions from the other party(ies) characterize the communicative acts of those involved. Recorded telephone calls during the occupation of the West German Embassy in Stockholm by members of the Red Army Faction (RAF) on April 24, 1975 comprise the basis for this analysis. One of the occupiers speaks with various interlocutors located in an adjacent embassy building. These interlocutors are relatives of the hostages, the Swedish Minister of Justice, and a German official charged with leading the negotiations. In this study, the communicative processes of the crisis are reconstructed. In order to show how the interlocutors attempt to reach their goals in this tense situation with the resources available to them, as well as what they in fact achieve, ethnographic methods of analysis have been employed. This study shows how, despite strong conflicting interests and motives, a shared reality is built through the actions of the interlocutors. The interaction between two key figures in the early stages of the crisis can even be characterized as a form of coalition building. An explanation as to why this collaboration is not retained in the subsequent course of the events, however, leading to an escalation of the situation, is also presented. Furthermore, the following work sets forth qualities needed to interactively build a coalition in a precarious crisis situation, which has arisen between parties characterized by diametrically opposed aims.

博士
國立成功大學
資訊工程研究所
88
The goal of communication is to achieve a shared understanding of the message and speech communication is one of the basis and most essential capabilities possessed for human being. Speech is the most natural way of human-human communication and can be efficiently transmitted over a computer or telephone network. Information exchange by speech clearly plays a very significant role in our live. However, for the application and delivery of speech recognition technology, the performance degrades due to many human and environmental factors. Since the naive user cannot be expected, in general, to know how to talk to a machine, the keyword recognition and verification has been designed to be flexible in accepting a wide range of user response and behavior. It is able to spot the keywords and reject the speech that does not include any of valid keywords. Moreover, Mandarin is a tonal language and prosodic information especially plays an important role. In Mandarin speech, the unvoiced part of a syllable does not contain well prosodic information such as pitch period. Under the disadvantage environment of telephone network, there are many a variety of interference sources such as the channel distortion, handset distortion, and ambient background noise. Our research purpose is to design and develop an intelligent call manager system, which can provide a friendly user-interface and facilitate to accept a wide range of user response and behavior. More especially, the aims of this research are focused on: 1). by analyzing Mandarin speech, four prosodic features are extracted for utterance verification and further discrimination; 2). a fuzzy search algorithm is also developed to detect more reliable keyword candidates and deal with problems that the unvoiced part of a syllable is difficult to be correctly recognized; 3) due to channel distortion problems in telephony environment, the advanced model adaptation methods, named keyword-driven two-level codebook based stochastic matching, are proposed to eliminate the acoustic mismatching phenomenon; 4). an innovative statistical language modeling based on partial pattern tree is investigated and aimed to recover the undetected/misdetected errors in speech recognition; 5). incorporating with our approaches, a robust speech recognition system with friendly human-machine interface is developed and applied to computer and telephony integration. Experimental results show that the proposed approaches give an encouraging improvement and an intelligent call manager is implemented by integrating telephone speech recognition, text-to-speech conversion, Internet, and call pager. The proposed system can enhance the key phone switch to an innovative voice processing, and provide an efficient speech query interface. The related speech technologies can also make our live more productive and efficient. Recommendations pertaining to this research for future studies consist: coarticulation, segmentation, speaker variation problems, and insufficient linguistic knowledge.