Sequence Of Utterances Research Articles

A new corpus of geolocated ASR transcripts from Germany

AbstractThis report describes the Corpus of German Speech (CoGS), a 56-million-word corpus of automatic speech recognition transcripts from YouTube channels of local government entities in Germany. Transcripts have been annotated with latitude and longitude coordinates, making the resource potentially useful for geospatial analyses of lexical, morpho-syntactic, and pragmatic variation; this is exemplified with an exploratory geospatial analysis of grammatical variation in the encoding of past temporal reference. Additional corpus metadata include video identifiers and timestamps on individual word tokens, making it possible to search for specific discourse content or utterance sequences in the corpus and download the underlying video and audio from the web, using open-source tools. The discourse content of the transcripts in CoGS touches upon a wide range of topics, making the resource potentially interesting as a data source for research in digital humanities and social science. The report also briefly discusses the permissibility of reuse of data sourced from German municipalities for corpus-building purposes in the context of EU, German, and American law, which clearly authorize such a use case.

CONVERSATIAONAL ANALYSIS AS A TOOL OF DRAMA DISCOURSE RESEARCH (BASED ON THE DRAMAS OF THE BRITISH PLAYWRIGHTS OF THE 20TH CENTURY)

The article examines the texts of dramas of British playwrights of the 20th century through the prism of the methodology of conversational analysis (CA). CA is a relatively new branch of Pragmatics and Sociolinguistics that investigates the features of speech interaction: the structure and formal properties of speech, which help interlocutors to interpret each other’s behaviour, language means at the empiric level, which put in order and control the sequence of expressions of interlocutors. CA developed from the theories of social organization of ethnomethodology by Harold Garfinkel. Sociologists and linguists Harvey Sacks, Emanuel A. Schegloff, and Gail Jefferson became the founders of the new research method of a dialogue (polilogue). Basic methodical principle of CA is collecting and fixing of the material in the situations of a real communication in everyday life. As a result of empiric data processing the categories of CA were developed: turn-taking, reparations, overlaps, sequences, adjacent pairs, and preferences. Turn takings are the base element of conversation organization. They are wellorganized utterances. A change of interlocuters is obligatory for turn-taking. They are also characterized by the presence of structural and allocational components. The first determines the volume and form of an utterance, the second is responsible for a choice of the next speaker. The sequence of utterances in a conversation is set on rules according to which a current interlocuter elects the following one. This is marked in a relevant transit point by means of one or a number of linguistic, paralinguistic or nonverbal facilities. It is also possible that another participant of a conversation starts talking on their own. If no one is willing to talk, a current interlocutor may continue talking. Nowadays conversationalists study not only real-life (domestic or institutional) but also written texts, that may be a representation of real-life talks (those which contain a direct address to the interlocutor or audience, have the system of turn-taking including more than one interlocutor). Such method of research can be applied to the analysis of drama discourse organization. The text of a drama is built, mainly, on the principle of exchange remarks between characters, and it is the interaction that is a global means of characters description, and action development, and also a means of the main idea expression. All the above mentioned makes it possible to apply CA to drama discourse research.

Is Discourse Role Important for Emotion Recognition in Conversation?

A conversation is a sequence of utterances, where each utterance plays a specific discourse role while expressing a particular emotion. This paper proposes a novel method to exploit latent discourse role information of an utterance to determine the emotion it conveys in a conversation. Specifically, we use a variant of the Variational-Autoencoder (VAE) to model the context-aware latent discourse roles of each utterance in an unsupervised way. The latent discourse role representation further equips the utterance representation with a salient clue for more accurate emotion recognition. Our experiments show that our proposed method beats the best-reported performances on three public Emotion Recognition in Conversation datasets. This proves that the discourse role information of an utterance plays an important role in the emotion recognition task, which no previous work has studied.

The opening sequences in the radio phone call-in conversations in Behdini-Kurdish

Verbal interaction is organized into sequences of utterances which are understood according to their sequential context. One of them is the opening sequences which are the initiation chunks of speech in order to start a conversation. Much work has been done on the opening sequences in other languages, whereas these sequences have not received attention by researchers in Kurdish. The study is the first attempt to identify the structure of the opening sequences in one variety of Kurdish, namely, Behdini. Besides, it will show the forms of each sequence identified in the structure and how the structures and sequences vary according to gender. The study is based on the analysis of the opening sequences of 77 radio phone call-in conversations. The data has been analysed qualitatively and quantitatively using the conversation analysis approach. Two structures of opening sequences in Kurdish are identified depending on whether the call is known or unknown. For the caller-unknown conversations, the structure is summons answer, greeting, how are you sequence, identification, and greetings sending/ compliments/ guest welcoming sequence. In the caller-known, identification sequence is not realized. The study identified the common forms of each sequence realized in the structure of the opening sequences. Some gender variations are realized in the structure of the opening sequences, their sequences and forms.
 The study is important because it will help identify language-specific conventions of opening sequences and to which extent they vary according to social factors. Besides, it will enrich the typological studies of opening sequences by adding another language-specific convention of opening sequences to the ones already investigated and to pave the way to formulate general or even universal principles of organisation of talk in interaction.

“Ana Malie” The Traditional Buginise Song Lyrics in Discourse Analysis

This study aimed to find out discourse analysis of traditional Buginese songs through elements used in the song.By analysing the lyrics of the song and focuses on the discourse widely used in the song.The study uses qualitative descriptive method.Some discourse devices that used are cohesion, coherence, intentionality, acceptability, informativeness, situationality, and intertextuality. In this song the cohesion of each part of the sentence does not appear, it is because the sentences depend on each other.The coherence in this song is clear enough because the meaning of each part of the sentences can be easily understood. The writer’s intentionality stated in a clear word that the writer used in each of the sentences.In acceptable devices the song is considered acceptable because the song uses phrases or words that are commonly used and known by the society. The informativeness we can gain fromthis song is the story about a mother who feels sorry about her son’s fate in this world, it is measured by constituting some sequence of utterances that contain information. In this song, situationality can be measured fromthe level of coherence and cohesiveness. Some cohesive markers in the song were not available.The sentences in the song are standalone but whether it is coherence or not, for the researcher its coherence. Intertextuality is the relation between texts that are inflicted by means of quotations and allusion The intention of the song writer was explicitly stated. The moral value that is immersed in the song gives a huge lesson of real life situations.

A Human–Machine Interaction Scheme Based on Background Knowledge in 6G-Enabled IoT Environment

6G-Enabled Internet of Things (IoT) is about to open a new era of Internet of Everything (IoE). It creates favorable conditions for new application services. The human-machine dialogue system, one of the most important forms of human–machine interaction, is expected to replace mobile applications in the future. This article proposes a dialogue generation scheme named background knowledge-aware dialogue generation model with pretrained encoders (BKADGPE). Dialogue generation, which takes the context as input and response as output, is a sequence-to-sequence (Seq2Seq) task. Instead of only generating the response based on the previous sequence of utterances, background knowledge-aware dialogue generation is also relying on background knowledge documents. This is because people often communicate based on their background knowledge. This article divides it into two tasks: 1) a knowledge selection task and 2) a response generation task. One of the latest language pretraining models, a lite bidirectional encoder representations from transformers (ALBERT), is applied as the encoder. In the knowledge selection task, ALBERT adds the linear layer and softmax layer to predict the content-related knowledge span. In the response generation task, the ALBERT after fine-tuning through the knowledge selection task adds the left-context-only transformer with a copy mechanism to incorporate background knowledge span into the generated response. Empirical studies on the HOLL-E dataset show that the result of BKADGPE is better than the related works.

Значение говорящего: проблема участия людей с речевыми нарушениями в коммуникативном взаимодействии

The article considers the problem of participation in social activities of people with disabilities from a perspective of face-to-face interaction. The analysis focuses on episodes of interaction between people with speech impairments and staff members of a daycare center during regular group meetings. Communication in such cases may be difficult because of impairments that affect capacities of speech production (e.g., inability to speak, extremely limited vocabulary, problems with articulation) by disabled participants. These impairments lead to the ambiguity of some of their utterances and actions. However, as studies of “atypical interaction” demonstrate, participants can successfully overcome such obstacles relying on the context of interaction—previous sequences of utterances and actions, bodily orientation, posture, gaze direction, interactional environment, current activity, and common background knowledge. Tying current potentially ambiguous communicative act (e.g., gesture or phonetically distorted utterance) to one or another aspect of the context, fully competent speakers demonstrate their interpretation of that act by their subsequent responses. Due to such coconstruction of meaning, their partners with speech impairments get an opportunity to make contributions to the interaction, in other words, to participate in it. However, such contextualizations and interpretations may not only enhance participation but restrict it as well. The article demonstrates how staff members ascribe meaning to some of those ambiguous actions and utterances by tying them to the prior turns in the interaction. However, there is evidence that staff members’ ascriptions probably do not quite agree with communicative intentions of their partners. As analysis of the video recording of these interactions demonstrates, there are aspects of the context that may suggest different interpretations. But these aspects go unnoticed, and this probably leads the people with speech impairments to the loss of partial control of meaning of their actions and reduction of their participation. Article in Russian DOI: 10.25285/2078-1938-2020-12-3-100-125

Multi-Fusion Residual Memory Network for Multimodal Human Sentiment Comprehension

Multimodal human sentiment comprehension refers to recognizing human affection from multiple modalities. There exist two key issues for this problem. First, it is difficult to explore time-dependent interactions between modalities and focus on the important time steps. Second, processing the long fused sequence of utterances is susceptible to the forgetting problem due to the long-term temporal dependency. In this article, we introduce a hierarchical learning architecture to classify utterance-level sentiment. To address the first issue, we perform time-step level fusion to generate fused features for each time step, which explicitly models time-restricted interactions by incorporating information across modalities at the same time step. Furthermore, based on the assumption that acoustic features directly reflect emotional intensity, we pioneer emotion intensity attention to focus on the time steps where emotion changes or intense affections take place. To handle the second issue, we propose Residual Memory Network (RMN) to process the fused sequence. RMN utilizes some techniques such as directly passing the previous state into the next time step, which helps to retain the information from many time steps ago. We show that our method achieves state-of-the-art performance on multiple datasets. Results also suggest that RMN yields competitive performance on sequence modeling tasks.

Utterance-to-Utterance Interactive Matching Network for Multi-Turn Response Selection in Retrieval-Based Chatbots

This article proposes an utterance-to-utterance interactive matching network (U2U-IMN) for multi-turn response selection in retrieval-based chatbots. Different from previous methods following context-to-response matching or utterance-to-response matching frameworks, this model treats both contexts and responses as sequences of utterances when calculating the matching degrees between them. For a context-response pair, the U2U-IMN model first encodes each utterance separately using recurrent and self-attention layers. Then, a global and bidirectional interaction between the context and the response is conducted using the attention mechanism to collect the matching information between them. The distances between context and response utterances are employed as a prior component when calculating the attention weights. Finally, sentence-level aggregation and context-response-level aggregation are executed in turn to obtain the feature vector for matching degree prediction. Experiments on four public datasets showed that our proposed method outperformed baseline methods on all metrics, achieving a new state-of-the-art performance and demonstrating compatibility across domains for multi-turn response selection.

Reading Scripted Dialogue: Pretending to Take Turns

ABSTRACTThis study examined conversational turn-taking when reading scripted dialogue versus having a spontaneous conversation. When reading, listeners waited until speakers’ turns had been completed before moving to start their own turns, and gap lengths between turns were longer than those of spontaneous, natural conversation. Theatrical performance training and studying the scripts worked to reduce readers’ gap lengths but not to the level of natural conversation, because all readers persisted in waiting for the ends of speakers’ scripted turns instead of coming in earlier as they had done in natural conversation. Participants were therefore not negotiating for the conversational floor when reading but were instead working to realize the prescribed sequence of utterances. These results advise consideration when interpreting turn-taking findings drawn from participants reading scripted materials, because readers did not engage in the collaborative, interactive turn taking that is characteristic of spontaneous conversation. Instead, readers were essentially pretending to take turns.

Object-Label-Order Effect When Learning From an Inconsistent Source.

Learning in natural environments is often characterized by a degree of inconsistency from an input. These inconsistencies occur, for example, when learning from more than one source, or when the presence of environmental noise distorts incoming information; as a result, the task faced by the learner becomes ambiguous. In this study, we investigate how learners handle such situations. We focus on the setting where a learner receives and processes a sequence of utterances to master associations between objects and their labels, where the source is inconsistent by design: It uses both "correct" and "incorrect" object-label pairings. We hypothesize that depending on the order of presentation, the result of the learning may be different. To this end, we consider two types of symbolic learning procedures: the Object-Label (OL) and the Label-Object (LO) process. In the OL process, the learner is first exposed to the object, and then the label. In the LO process, this order is reversed. We perform experiments with human subjects, and also construct a computational model that is based on a nonlinear stochastic reinforcement learning algorithm. It is observed experimentally that OL learners are generally better at processing inconsistent input compared to LO learners. We show that the patterns observed in the learning experiments can be reproduced in the simulations if the model includes (a) an ability to regularize the input (and also to do the opposite, i.e., undermatch) and (b) an ability to take account of implicit negative evidence (i.e., interactions among different objects/labels). The model suggests that while both types of learners utilize implicit negative evidence in a similar way, there is a difference in regularization patterns: OL learners regularize the input, whereas LO learners undermatch. As a result, OL learners are able to form a more consistent system of image-utterance associations, despite the ambiguous learning task.

A Corpus-Based Study on Comparing Discourse Markers by Francis Younghusband and Nicholas Roerich to Tibet

A discourse marker is a word or a phrase that plays a role in managing the flow and structure of discourse. Since their main function is at the level of discourse (sequences of utterances) rather than at the level of utterances or sentences, discourse markers are relatively syntax-independent and usually do not change the truth conditional meaning of the sentence. Regarding this point, the surrent study aimed to do a corpus-based study to compare the discourse markers byFrancis Younghusband and Nicholas Roerichto Tibet. A study of the British expedition to Tibet by the famous orientalist Colonel F. Younghusband allows us to better understand the place that Tibet occupied in the international politics of various states, in particular, Russia and Great Britain, and the role that the investigator played in this. F. Younghusband left a noticeable mark in the history of the so-called “Big Game”. In addition to the political result, it is difficult to underestimate the important travel notes of F. Younghusband, numerous monographs describing not only the political process of the task assigned to him but also the region of Asia as a whole, nature and people. N. Roerich with his sons, one of whom was an orientalist and another was an artist, implemented many cultural projects, however, in historiography, there are opinions about the political nature of his mission. A number of implications of the results are discussed and relevant recommendations are presented.

Interactive Component in Oral Narratives of Preschool Children: on the Formulation of Hypothesis

The article presents an effort to elaborate the hypothesis for the experimental research devoted to interactive component in the oral narratives of preschool children. It is vital for psycholinguistics to reach a close approach to understanding of the essence of narrative processes which have in ontogenesis their own specifics, and according to some researches, these specifics are presented in indirect character of the process of narrative creation by child in the form of his verbal interaction with an adult.
 We assume that in order to reveal the ontogenetic cognitive discursive mechanisms involved in the process of narration, it is necessary to correlate the regularities of the verbal interaction of the participants in the narrative situation with the overall narrative tasks coming up in narrative process.
 Mеthodology, purpose and tasks of the research are determined by experimentally-genetic method (L.S. Vygotsky) put forward as main requirements «differentiation of the analysis of a thing and analysis of the process», «genetic restoration of all moments of the process development». The following of these requirements means a close following of the process of narrative development in its dynamics and the peculiarities of its formation at different stages of child’s development, paying attention to frequency facts as well as facts overcoming the limits of «narrow» phenomenon – narrative text.
 To work out the hypothesis the narrative material of the earlier made experiment with a group of 3 to 7 age children was used in which an adult took a passive part. The task was to tell the story. The material was not subjected to any analysis for the reason of the adult’s presence during the storytelling. The analysis of oral histories with interactive component included the following tasks: a) choose regular sequences of utterances in the dialogic interaction of child and adult; b) define the functional purpose of the adult’s utterance in the course in the child’s storytelling.
 The obtained results show that the adult practically at every step of the child’s age is oriented on the child’s comprehension and analyzing episodes of the history, that is on child’s formation of his mental image of structural and event sequence in history. The results of the analysis enabled us to propose a hypothesis for the next research step in the field. Along the line of this hypothesis, in the verbal ontogenesis the very nature of oral story as a multifaceted phenomenon of the discursive activity of man brings up to the surface the inner mechanisms of narration which have dialogue basis.

A Qualitative Interview With Young Children: What Encourages or Inhibits Young Children’s Participation?

The goal of every qualitative interview is to produce rich data. Inducing storytelling is a challenge in every interview. Interviews with young children (ages 3–6) present an additional challenge because of perceived power differences between children and adults. This research examines how interviewers’ questions and expressions encourage or inhibit children from telling their stories. We extracted 1,339 child interviewee–adult interviewer turn exchanges from a national study on children’s perspectives on risk and protection ( N = 420) and analyzed them in two steps. First, we categorized the interviewers’ questions and expressions and children’s responses. Seven categories were found for interviewer expressions and five for children’s responses. We then examined the relationship between interviewer categories and children’s responses. The categories that produced the richest data were encouragement, open-ended questions, and question request. Sequence of utterances and closed-ended questions produced the least storytelling. We did not find significant differences based on a child’s gender with regard to the interviewer categories. The results and implications for researching young children are addressed.

Graph Convolutional Network with Sequential Attention for Goal-Oriented Dialogue Systems

Domain-specific goal-oriented dialogue systems typically require modeling three types of inputs, namely, (i) the knowledge-base associated with the domain, (ii) the history of the conversation, which is a sequence of utterances, and (iii) the current utterance for which the response needs to be generated. While modeling these inputs, current state-of-the-art models such as Mem2Seq typically ignore the rich structure inherent in the knowledge graph and the sentences in the conversation context. Inspired by the recent success of structure-aware Graph Convolutional Networks (GCNs) for various NLP tasks such as machine translation, semantic role labeling, and document dating, we propose a memory-augmented GCN for goal-oriented dialogues. Our model exploits (i) the entity relation graph in a knowledge-base and (ii) the dependency graph associated with an utterance to compute richer representations for words and entities. Further, we take cognizance of the fact that in certain situations, such as when the conversation is in a code-mixed language, dependency parsers may not be available. We show that in such situations we could use the global word co-occurrence graph to enrich the representations of utterances. We experiment with four datasets: (i) the modified DSTC2 dataset, (ii) recently released code-mixed versions of DSTC2 dataset in four languages, (iii) Wizard-of-Oz style CAM676 dataset, and (iv) Wizard-of-Oz style MultiWOZ dataset. On all four datasets our method outperforms existing methods, on a wide range of evaluation metrics.

Context formulation and the invocation of membership categories in an L2 classroom setting

The purpose of this article is to examine the practical work that participants in a second language (L2) classroom are doing in the course of their actions. These actions are observed through the accounts produced by members, which in turn represent their rationalities, or ethno-methods (Garfinkel, 2002), that are in use during the ongoing sequence of utterances. Based on the extracts presented for analysis, this article concentrates on how participants produce and recognise the context in which they are in, and to which they orient their work; and how this context becomes an inspectable phenomenon of inquiry through ethno-methods, such as the invocation of membership categories (‘teacher’ and ‘student’) produced along the course of the interaction. The materials on which I shall focus my discussions were gathered from a first grade Portuguese as an L2 classroom in a primary school in Macau, China. The results indicated an ongoing production of categorial aspects that enabled participants to have no problems in following the sequential organisation of the talk and other conduct in a way that it was ‘just enough’ to make L2 lesson events happen.

Maternal discourse continuity and infants' actions organize 12-month-olds' language exposure during object play.

Infant language learning depends on the distribution of co-occurrences within language-between words and other words-and between language content and events in the world. Yet infant-directed speech is not limited to words that refer to perceivable objects and actions. Rather, caregivers' utterances contain a range of syntactic forms and expressions with diverse attentional, regulatory, social, and referential functions. We conducted a distributional analysis of linguistic content types at the utterance level, and demonstrated that a wide range of content types in maternal speech can be distinguished by their distribution in sequences of utterances and by their patterns of co-occurrence with infants' actions. We observed free-play sessions of 38 12-month-old infants and their mothers, annotated maternal utterances for 10 content types, and coded infants' gaze target and object handling. Results show that all content types tended to repeat in consecutive utterances, whereas preferred transitions between different content types reflected sequences from attention-capturing to directing and then descriptive utterances. Specific content types were associated with infants' engagement with objects (declaratives, descriptions, object names), with disengagement from objects (talk about attention, infant's name), and with infants' gaze at the mother (affirmations). We discuss how structured discourse might facilitate language acquisition by making speech input more predictable and/or by providing clues about high-level form-function mappings.

REPRESENTATIONS OF INTERSUBJECTIVITY IN THE MEDIA DISCOURSE OF THE FRG

The article touches upon the verbal means of cognitive strategies and techniques used by participants of a personal interview, which is the most frequent pattern of modern media discourse, online media. The ways of overcoming conflicts and reaching consensus are described in the framework of the discourse cognitive phenomenon of intersubjectivity.The article contains a description of an integrated methodology developed for the study of one of the hybrid forms of modern communication, a personal interview, which combines a thoughtful cognitive plan characterizing the interviewer’s behavior and the spontaneous response of the interviewer, the condition of which becomes the topic of communication. The overall goal of the participants is to reach an agreement, which is possible only if the interviewer refuses to dominate and trust him.In the course of the research, the hypothesis of existence of modern hybrid practices of algorithms for verbal behavior was verified, with the help of which the regulation of communication process is regulated and the possibility of reaching the partners’ consent is provided. The complex methodology, which included the research methods of conversion analysis and discourse analysis, made it possible to identify the phases ofthe interview, characterized by standard formulas implemented with the help of information request tactics, and highlight transactions in which the field of interaction of participants is created. Key to the success ofthe interview phase of the communicative process involves, as the study shows, an equal distribution of the partners’ initiative, achieved in the event of the interviewer’s refusal to dominate.The selection of common questions, pronominal pick-ups, development of the topic chosen by the interviewer, implement strategies for motivating the response, which allows the partner to generate a detailed monologue statement about himself. In the transcript of the interview, this phase of the communicative process is reflected in the emergence of a sequence of utterances with high information density. At the same time, the speed of communicative exchange changes significantly, replicas of participants alternate with each communicative move.Verbal algorithms that organize communicative interaction in a personal interview are located, as the study shows, in certain phases of the communication process, and can be presented as a sequence oftransactions. The possibilities of conversion analysis, allowing to demonstrate the structure of transactions, receive meaningful content by describing communicative strategies and tactics that implement them.The results of the research are considered in the article in the perspective of the «grammar of the language game», declared by universal pragmatics and intended to form an idea of how communicative interaction proceeds in various discursive practices.

Smart Enough to Talk With Us? Foundations and Challenges for Dialogue Capable AI Systems

Smart Enough to Talk With Us? Foundations and Challenges for Dialogue Capable AI Systems

Exploring Digital Conversation Corpora with Process Mining

Understanding how sentences constitute conversations is still a matter of disagreement. The automatic analysis of corpora can bring new perspectives on this issue. We show that process mining, a computer science approach for analysing human behaviour, enables the inductive identification of within-turn regularities in digital conversations. We use speech acts as a way to discretise conversations and enable the analysis. To this end, we develop a fine-grained, corpus-independent classification of speech acts that we describe and validate. We perform a manual sentence-by-sentence multi-label contextual annotation of a corpus of digital conversations and we explore it with process mining. Our results suggest that speech act theory could be extended to include typicality conditions for sequences of utterances.