Conversational Utterances Research Articles

Revisiting Conversation Discourse for Dialogue Disentanglement

Dialogue disentanglement aims to detach the chronologically ordered utterances into several independent sessions. Conversation utterances are essentially organized and described by the underlying discourse, and thus dialogue disentanglement requires the full understanding and harnessing of the intrinsic discourse attribute. In this paper, we propose enhancing dialogue disentanglement by taking full advantage of the dialogue discourse characteristics. First of all, in feature encoding stage , we construct the heterogeneous graph representations to model the various dialogue-specific discourse structural features, including the static speaker-role structures (i.e., speaker-utterance and speaker-mentioning structure) and the dynamic contextual structures (i.e., the utterance-distance and partial-replying structure). We then develop a structure-aware framework to integrate the rich structural features for better modeling the conversational semantic context. Second, in model learning stage , we perform optimization with a hierarchical ranking loss mechanism, which groups dialogue utterances into different discourse levels and carries training covering pair-wise and session-wise levels hierarchically. Third, in inference stage , we devise an easy-first decoding algorithm, which performs utterance pairing under the easy-to-hard manner with a global context, breaking the constraint of traditional sequential decoding order. On two benchmark datasets, our overall system achieves new state-of-the-art performances on all evaluations. In-depth analyses further demonstrate the efficacy of each proposed idea and also reveal how our methods help advance the task. Our work has great potential to facilitate broader multi-party multi-thread dialogue applications.

The Conversational Development Function of Repetition of Others’ Utterances in Japanese Conversation and Korean Conversation

The Conversational Development Function of Repetition of Others’ Utterances in Japanese Conversation and Korean Conversation

HindiPersonalityNet: Personality Detection in Hindi Conversational Data Using Deep Learning with Static Embedding

Personality detection along with other behavioral and cognitive assessment can essentially explain why people act the way they do and can be useful to various online applications such as recommender systems, job screening, matchmaking, and counseling. Additionally, psychometric natural language processing relying on textual cues and distinctive markers in writing style within conversational utterances reveals signs of individual personalities. This work demonstrates a text-based deep neural model, HindiPersonalityNet, of classifying conversations into three personality categories (ambivert, extrovert, introvert) for detecting personality in Hindi conversational data. The model utilizes a gated recurrent unit with BioWordVec embeddings for text classification and is trained/tested on a novel dataset, शख्सियत (pronounced as Shakhsiyat) curated using dialogues from an Indian crime-thriller drama series, Aarya . The model achieves an F1-score of 0.701 and shows the potential for leveraging conversational data from various sources to understand and predict a person's personality traits. It exhibits the ability to capture both semantic and long-distance dependencies in conversations and establishes the effectiveness of our dataset as a benchmark for personality detection in Hindi dialogue data. Further, a comprehensive comparison of various static and dynamic word embedding is done on our standardized dataset to ascertain the most suitable embedding method for personality detection.

Quantifying the emergence of moral foundational lexicon in child language development.

Theorists have argued that morality builds on several core modular foundations. When do different moral foundations emerge in life? Prior work has explored the conceptual development of different aspects of morality in childhood. Here, we offer an alternative approach to investigate the developmental emergence of moral foundations through the lexicon, namely the words used to talk about moral foundations. We develop a large-scale longitudinal analysis of the linguistic mentions of five moral foundations (in both virtuous and vicious forms) in naturalistic speech between English-speaking children with ages ranging from 1 to 6 and their caretakers. Using computational methods, we collect a dataset of 1,371 human-annotated moral utterances and automatically annotate around one million utterances in child-caretaker conversations. We discover that in childhood, words for expressing the individualizing moral foundations (i.e. Care/Harm, Fairness/Cheating) tend to emerge earlier and more frequently than words for expressing the binding moral foundations (i.e. Authority/Subversion, Loyalty/Betrayal, Purity/Degradation), and words for Care/Harm are expressed substantially more often than the other foundations. We find significant differences between children and caretakers in how often they talk about Fairness, Cheating, and Degradation. Furthermore, we show that the information embedded in childhood speech allows computational models to predict moral judgment of novel scenarios beyond the scope of child-caretaker conversations. Our work provides a large-scale documentation of the moral foundational lexicon in early linguistic communication in English and forges a new link between moral language development and computational studies of morality.

Literary Stylistics Features in the Musical Drama NURBAYA by Indonesia Kaya

This study examines the literary stylistic features in NURBAYA, a contemporary Indonesian musical drama produced by Indonesia Kaya. Employing Leech and Short's (1981) framework, the research focuses on sound patterns and figures of speech, investigating how these elements enhance narrative cohesion and emotional impact while reflecting cultural values. Using qualitative content analysis, the study systematically categorizes and interprets stylistic features in conversation utterances and song lyrics. The analysis reveals a prevalence of sound patterns, particularly rhyme, which contributes to the musical's rhythmic quality and cohesion. Figures of speech, notably litotes, are employed to reflect cultural values of politeness and subtlety. For instance, the frequent use of litotes (found in 35% of character dialogues) underscores adherence to traditional norms, enhancing the play's cultural resonance. NURBAYA, as a bridge between traditional and contemporary Indonesian theatre, offers a unique lens for examining the evolution of literary style in performance art. This study demonstrates how specific literary devices enrich artistic expression and reinforce cultural themes in contemporary Indonesian theatre. These findings contribute to our understanding of the interplay between linguistics, literature, and cultural expression in performing arts, offering insights into how stylistic analysis can illuminate broader sociocultural dynamics in theatrical works.

Breathy Voice and the Oxytonic or the Paroxytonic Rhythm in the Reproduction of Thought Creations in German-Language Conversations

This research is based on the assumption that the breathy voice within a set of linguistic-prosodic features, in addition to the reproduction of already realized utterances, marking utterances that the speaker assumes participants of possible future events could utter and marking added fictive utterances in German-language conversations is also used for reproduction of other thought creations. The corpus of this research consists of clips from conversations conducted within German phone-in radio shows, with each of them containing a single use of breathy voice. The analysis shows that the breathy voice is realized in one part of the reported speech and then remains without breathiness, leaving the modal voice. In the analysis, reported speech partly represented the verbalization of assumed utterances or concrete utterances realized multiple times in the past. Thought creations whose reconstructions were verbalized included, in addition to utterances, previous thoughts or assumed thoughts of others.

Violations Of Grice’s Maxim

Violations in conversation are found not only in everyday life when people interact with other people, but also, in drama and films performed by characters, in novels, interviews, debates, and other forms of social interaction performed for various reasons. This study examines the violations of Grice's Maxim in the movie Jack Reacher: Never Go Back, focusing on the characters' conversational utterances. Using observation method and note-taking technique, the descriptive qualitative method was applied, analyzing data using the Theory of Cooperative Principles and Context of Situation. The results showed twenty-nine data of conversation were violated by the characters in the movie. Maxim of Quality (51,7%), Maxim of Quantity (13,8%), Maxim of Relation (10,4%), and Maxim of Manner (24,1%).

Cooking with Conversation: Enhancing User Engagement and Learning with a Knowledge-Enhancing Assistant

We present two empirical studies to investigate users’ expectations and behaviours when using digital assistants, such as Alexa and Google Home, in a kitchen context: First, a survey (N = 200) queries participants on their expectations for the kinds of information that such systems should be able to provide. While consensus exists on expecting information about cooking steps and processes, younger participants who enjoy cooking express a higher likelihood of expecting details on food history or the science of cooking. In a follow-up Wizard-of-Oz study (N = 48), users were guided through the steps of a recipe either by an active wizard that alerted participants to information it could provide or a passive wizard who only answered questions that were provided by the user. The active policy led to almost double the number of conversational utterances and 1.5 times more knowledge-related user questions compared to the passive policy. Also, it resulted in 1.7 times more knowledge communicated than the passive policy. We discuss the findings in the context of related work and reveal implications for the design and use of such assistants for cooking and other purposes such as DIY and craft tasks, as well as the lessons we learned for evaluating such systems.

Multi-Modal Attentive Prompt Learning for Few-shot Emotion Recognition in Conversations

Emotion recognition in conversations (ERC) has emerged as an important research area in Natural Language Processing and Affective Computing, focusing on accurately identifying emotions within the conversational utterance. Conventional approaches typically rely on labeled training samples for fine-tuning pre-trained language models (PLMs) to enhance classification performance. However, the limited availability of labeled data in real-world scenarios poses a significant challenge, potentially resulting in diminished model performance. In response to this challenge, we present the Multi-modal Attentive Prompt (MAP) learning framework, tailored specifically for few-shot emotion recognition in conversations. The MAP framework consists of four integral modules: multi-modal feature extraction for the sequential embedding of text, visual, and acoustic inputs; a multi-modal prompt generation module that creates six manually-designed multi-modal prompts; an attention mechanism for prompt aggregation; and an emotion inference module for emotion prediction. To evaluate our proposed model’s efficacy, we conducted extensive experiments on two widely recognized benchmark datasets, MELD and IEMOCAP. Our results demonstrate that the MAP framework outperforms state-of-the-art ERC models, yielding notable improvements of 3.5% and 0.4% in micro F1 scores. These findings highlight the MAP learning framework’s ability to effectively address the challenge of limited labeled data in emotion recognition, offering a promising strategy for improving ERC model performance.

Lexical Alignment is Pervasive Across Contexts in Non-WEIRD Adult-Child Interactions.

Lexical alignment, a communication phenomenon where conversational partners adapt their word choices to become more similar, plays an important role in the development of language and social communication skills. While this has been studied extensively in the conversations of preschool-aged children and their parents in Western, Educated, Industrialized, Rich, and Democratic (WEIRD) communities, research in other pediatric populations is sparse. This study makes significant expansions on the existing literature by focusing on alignment in naturalistic conversations of school-aged children from a non-WEIRD population across multiple conversational tasks and with different types of adult partners. Typically developing children aged 5 to 8 years (n = 45) engaged in four semi-structured conversations that differed by task (problem-solving vs. play-based) and by partner (parent vs. university student), resulting in a corpus of 180 conversations. Lexical alignment scores were calculated and compared to sham conversations, representing alignment occurring at the level of chance. Both children and adults coordinated their conversational utterances by re-using or aligning each other's word choices. This alignment behavior persisted across conversational tasks and partners, although the degree of alignment was moderated by the conversational context. These findings suggest that lexical alignment is a robust phenomenon in conversations between school-age children and adults. Furthermore, this study extends lexical alignment findings to a non-WEIRD culture, suggesting that alignment may be a coordination strategy employed by adults and children across diverse linguistic and cultural backgrounds.

The Potential of Chatbots for Emotional Support and Promoting Mental Well-Being in Different Cultures: Mixed Methods Study.

Artificial intelligence chatbot research has focused on technical advances in natural language processing and validating the effectiveness of human-machine conversations in specific settings. However, real-world chat data remain proprietary and unexplored despite their growing popularity, and new analyses of chatbot uses and their effects on mitigating negative moods are urgently needed. In this study, we investigated whether and how artificial intelligence chatbots facilitate the expression of user emotions, specifically sadness and depression. We also examined cultural differences in the expression of depressive moods among users in Western and Eastern countries. This study used SimSimi, a global open-domain social chatbot, to analyze 152,783 conversation utterances containing the terms "depress" and "sad" in 3 Western countries (Canada, the United Kingdom, and the United States) and 5 Eastern countries (Indonesia, India, Malaysia, the Philippines, and Thailand). Study 1 reports new findings on the cultural differences in how people talk about depression and sadness to chatbots based on Linguistic Inquiry and Word Count and n-gram analyses. In study 2, we classified chat conversations into predefined topics using semisupervised classification techniques to better understand the types of depressive moods prevalent in chats. We then identified the distinguishing features of chat-based depressive discourse data and the disparity between Eastern and Western users. Our data revealed intriguing cultural differences. Chatbot users in Eastern countries indicated stronger emotions about depression than users in Western countries (positive: P<.001; negative: P=.01); for example, Eastern users used more words associated with sadness (P=.01). However, Western users were more likely to share vulnerable topics such as mental health (P<.001), and this group also had a greater tendency to discuss sensitive topics such as swear words (P<.001) and death (P<.001). In addition, when talking to chatbots, people expressed their depressive moods differently than on other platforms. Users were more open to expressing emotional vulnerability related to depressive or sad moods to chatbots (74,045/148,590, 49.83%) than on social media (149/1978, 7.53%). Chatbot conversations tended not to broach topics that require social support from others, such as seeking advice on daily life difficulties, unlike on social media. However, chatbot users acted in anticipation of conversational agents that exhibit active listening skills and foster a safe space where they can openly share emotional states such as sadness or depression. The findings highlight the potential of chatbot-assisted mental health support, emphasizing the importance of continued technical and policy-wise efforts to improve chatbot interactions for those in need of emotional assistance. Our data indicate the possibility of chatbots providing helpful information about depressive moods, especially for users who have difficulty communicating emotions to other humans.

ECPEC: Emotion-Cause Pair Extraction in Conversations

Conversational sentiment analysis (CSA) and emotion-cause pair extraction (ECPE) tasks have attracted increasing attention in recent years. The former aims to predict the sentiment states of speakers in a conversation, and the latter is about extracting emotion-cause clauses in a document. However, one drawback of CSA is that it cannot model the causal reasoning among emotion and neutral utterances from different speakers. In this work, we propose a new task: emotion-cause pair extraction in conversations (ECPEC), which aims to extract pairs of emotional utterances and corresponding cause utterances in conversations. The utterance-level ECPEC task is more challenging since the distance between emotion and cause utterances is larger than that of the clause-level ECPE task. To this end, we build a novel dataset ConvECPE and propose a specifically designed two-step framework for the new ECPEC task. Experimental results on ConvECPE dataset demonstrate the feasibility of the ECPEC task as well as the effectiveness of our framework.

TYPES OF BARRIERS TO COMMUNICATION IN EMERGENCY ROOM BETWEEN NURSE AND PATIENT’S FAMILY IN DOLOKSANGGUL HOSPITAL

The objective of this study was to describe types of barriers to communication in emergency room between nurse and patient’s family in Doloksanggul hospital. This research was conducted by using descriptive qualitative research. The data of this study were utterances in conversation between nurse and patient’s family in Doloksanggul hospital. Participant observation technique was performed to collect the data. The results showed that (1) The types of communication barriers most frequently encountered in emergency room of the Doloksanggul Hospital 4% related to process barriers, 32 % related to semantic barriers, 1 % related to physical and cultural barriers, 12% related to psychological barriers, and 16 % related to perception. Then, the dominant type is related to psychological barriers. From the analyzed realization types of barriers to communication it was found that Psychological barriers was the largest number, it showed that unstable mental condition or panic because of the patient's family situation who worried about the patient create barrier to communication.

Father and Toddler Language During Shared Book Reading with Text‐Based and Wordless Picture Books

AbstractWhile the association between shared book reading (SBR) and child language development is well documented, there has been less focus on how book characteristics may differentially elicit parents' language input and hence differentially relate to children's language skills during this activity. Moreover, despite the positive and unique role that fathers have been shown to play for children's language, the father‐child SBR evidence base is small. Accordingly, the present study examined variation in father and child language quantity, quality, complexity, and function during SBR with text‐based and wordless picture books and explored associations between fathers' language and children's language during these episodes. Participants were 46 father‐child dyads (20 females). Fathers' mean age was 38.74 years (SD = 5.74), and children's mean age was 37.89 months (SD = 2.71). Findings indicate that fathers in the text‐book condition produced significantly greater language quality, extended the topic more, and produced more conversational utterances but described and labeled pictures less than fathers in the wordless picture book condition. Fathers' language during both SBR episodes showed differential links with child's language and may suggest that fathers were fine‐tuning their language in accordance with their child's language skills (effect sizes ranged from r = .45 to r = .78). It is possible that variation in book content may encourage different types of learning with distinct effects on different domains of child language. Should future research replicate such differential relations, findings may inform caretakers and early childhood interventions with the aim of promoting children's books tailored to specific language areas of interest.

Exploring Bi-Directional Context for Improved Chatbot Response Generation Using Deep Reinforcement Learning

The development of conversational agents that can generate relevant and meaningful replies is a challenging task in the field of natural language processing. Context and predictive capabilities are crucial factors that humans rely on for effective communication. Prior studies have had a significant limitation in that they do not adequately consider the relationship between utterances in conversations when generating responses. This study aims to address this limitation by proposing a novel method that comprehensively models the contextual information of the current utterance for response generation. A commonly used approach is to rely on the information of the current utterance to generate the corresponding response, and as such it does not take advantage of the context of a multi-turn conversation. In our proposal, different from other studies, we will use a bi-directional context in which the historical direction helps the model remember information from the past in the conversation, while the future direction enables the model to anticipate its impact afterward. We combine a Transformer-based sequence-to-sequence model and the reinforcement learning algorithm to achieve our goal. Experimental results demonstrate the effectiveness of the proposed model through qualitative evaluation of some generated samples, in which the proposed model increases 24% average BLEU score and 29% average ROUGE score compared to the baseline model. This result also shows that our proposed model improves from 5% to 151% for the average BLEU score compared with previous related studies.

Top-down effect of dialogue coherence on perceived speaker identity

A key mechanism in the comprehension of conversation is the ability for listeners to recognize who is speaking and when a speaker switch occurs. Some authors suggest that speaker change detection is accomplished through bottom-up mechanisms in which listeners draw on changes in the acoustic features of the auditory signal. Other accounts propose that speaker change detection involves drawing on top-down linguistic representations to identify who is speaking. The present study investigates these hypotheses experimentally by manipulating the pragmatic coherence of conversational utterances. In experiment 1, participants listened to pairs of utterances and had to indicate whether they heard the same or different speakers. Even though all utterances were spoken by the same speaker, our results show that when two segments of conversation are spoken by the same speaker but make sense for different speakers to say, listeners report hearing different speakers. In experiment 2 we removed pragmatic information from the same stimuli by scrambling word order while leaving acoustic information intact. In contrast to experiment 1, results from the second experiment indicate no difference between our experimental conditions. We interpret these results as a top-down effect of pragmatic expectations: knowledge of conversational structure at least partially determines a listener’s perception of speaker changes in conversation.

The development of Hebrew zero and pronominal subject realization in the context of first and second person.

The study investigates the acquisition of Hebrew zero and pronominal subjects in the context of first and second person. We provide distributional evidence relative to verb tense, number, person, and conversational utterance type, in a peer-talk corpus (2;0-8;0 years). Findings show that acquisition starts early on, that verb inflectional morphology is crucial for the development of pronominal subjects, and that communicative contexts affect subject realization. Zero and pronominal subjects are not evenly distributed relative to the study variables, and cannot be treated as an alternation. A cluster analysis shows that each realization is linked to a distinguishable usage pattern, corresponding with particular discursive and communicative functions. These are defined as three Discourse Profile Constructions: (A) "calling for action" by 1st.Pl.Fut zero subjects (3;0 year olds); (B) "commenting on the interlocutor's actions" by 2nd.Sg.Past zero subjects (ages 4;0-6;0); and (C) "planning one's own actions" by 1st.Sg.Fut pronominal subjects (7;0-8;0 year olds).

User-Chatbot Conversations During the COVID-19 Pandemic: Study Based on Topic Modeling and Sentiment Analysis.

Chatbots have become a promising tool to support public health initiatives. Despite their potential, little research has examined how individuals interacted with chatbots during the COVID-19 pandemic. Understanding user-chatbot interactions is crucial for developing services that can respond to people's needs during a global health emergency. This study examined the COVID-19 pandemic-related topics online users discussed with a commercially available social chatbot and compared the sentiment expressed by users from 5 culturally different countries. We analyzed 19,782 conversation utterances related to COVID-19 covering 5 countries (the United States, the United Kingdom, Canada, Malaysia, and the Philippines) between 2020 and 2021, from SimSimi, one of the world's largest open-domain social chatbots. We identified chat topics using natural language processing methods and analyzed their emotional sentiments. Additionally, we compared the topic and sentiment variations in the COVID-19-related chats across countries. Our analysis identified 18 emerging topics, which could be categorized into the following 5 overarching themes: "Questions on COVID-19 asked to the chatbot" (30.6%), "Preventive behaviors" (25.3%), "Outbreak of COVID-19" (16.4%), "Physical and psychological impact of COVID-19" (16.0%), and "People and life in the pandemic" (11.7%). Our data indicated that people considered chatbots as a source of information about the pandemic, for example, by asking health-related questions. Users turned to SimSimi for conversation and emotional messages when offline social interactions became limited during the lockdown period. Users were more likely to express negative sentiments when conversing about topics related to masks, lockdowns, case counts, and their worries about the pandemic. In contrast, small talk with the chatbot was largely accompanied by positive sentiment. We also found cultural differences, with negative words being used more often by users in the United States than by those in Asia when talking about COVID-19. Based on the analysis of user-chatbot interactions on a live platform, this work provides insights into people's informational and emotional needs during a global health crisis. Users sought health-related information and shared emotional messages with the chatbot, indicating the potential use of chatbots to provide accurate health information and emotional support. Future research can look into different support strategies that align with the direction of public health policy.

State Graph Reasoning for Multimodal Conversational Recommendation

Conversational recommendation system (CRS) attracts increasing attention in various application domains such as retail and travel. It offers an effective way to capture users' dynamic preferences with multi-turn conversations. However, most current studies center on the recommendation aspect while over-simplifying the conversation process. The negligence of complexity in data structure and conversation flow hinders their practicality and utility. In reality, there exist various relationships among slots and values, while users' requirements may dynamically adjust or change. Moreover, the conversation often involves visual modality to facilitate the conversation. These actually call for a more advanced internal state representation of the dialogue and a proper reasoning scheme to guide the decision making process. In this paper, we explore multiple facets of multimodal conversational recommendation and try to address the above mentioned challenges. In particular, we represent the structured back-end database as a multimodal knowledge graph which captures the various relations and evidence in different modalities. The user preferences expressed via conversation utterances will then be gradually updated to the state graph with clear polarity. Based on these, we train an end-to-end State Graph-based Reasoning model SGR to perform reasoning over the whole state graph. The prediction of our proposed model benefits from the structure of the graph. It not only allows for zero-shot reasoning for items unseen in training conversations, but also provides a natural way to explain the policies. Extensive experiments show that our model achieves better performance compared with existing methods.

Analysis of Individual Conversational Volatility in Tandem Telecollaboration for Second Language Learning

Second language (L2) learning can be enabled by tandem collaboration where students are grouped in video conference calls while learning the native language of other student(s) on the calls. This places students in an online environment where the more outgoing can actively contribute and engage in dialogue while those more shy and unsure of their second language language skills can sit back and coast through the calls. We have built and deployed the L2L system which records timings of conversational utterances from all participants in a call. We generate visualisations including participation rates and timelines for each student in each call and present these on a dashboard. Students can self-reflect and perhaps target improving their levels of engagement for subsequent calls. We have recently developed a measure called personal conversational volatility for how dynamic has been each student’s contribution to the dialogue in each call. This measures whether a student’s contribution was interactive with a mixture of interjections perhaps interrupting and agreeing with others combined with longer contributions, or whether it consisted of regular duration contributions with not much mixing. We present an analysis of conversational volatility measures of a sample of 19 individual English-speaking students from our University at lower intermediate-intermediate level (B1/B2) in their target language which was French, in each of 86 tandem telecollaboration calls over one teaching semester. Our analysis shows that students varied considerably in how their individual levels of engagement changed as their telecollaboration meetings progressed. Some students got more involved in the dialogue from one meeting to the next while others did not change their interaction levels at all. The reasons for this are not clear from the data we have and point to a need for further investigation into the nature of online tandem telecollaboration meetings. In particular there is a need to look into the nature of the interactions and see if the choices of discussion topics were too difficult for some lower intermediate students and that may have influenced their engagement in some way.