he purpose of this paper is to demonstrate how two non-native English speakers residing in Canada jointly evaluate and judge appropriate language behavior in Canada when they create a certain norm through narrative. The data used are lingua franca spoken in English by residents of Canada. We asked them to talk about their current occupations or future careers in a semi-structured interview. The recordings were conducted online (Zoom) for a total of six people in the format of two-person pairs (both met for the first time). The languages that the participants recognized as their mother tongue are as follows (Participant A: Swedish; Participant B: Yoruba, Bekwara, Mbe; Participant C: Nepalese; Participant D: Persian; Participant E: Arabic, French, Participant F: Thai). This paper shows that the participants evaluate the narrative together, make adjustments, and come to an agreement based on a certain kind of norm that white Canadian native English speakers with a background in Western culture generally have. The paper also shows that the co-constructed discourse becomes a socially shareable, understandable and desirable narrative because it is based on norms and beliefs of white Canadian native English speakers.

We are conducting a practical research project with the goal of promoting dialogue on equal footing between native speakers (NS) and non-native speakers (NNS) of Japanese living in the same community, on the topic of community development. In the theater workshop analyzed in the present study, participants experienced dis-communication situations through a series of short dramatic sketches and then were divided into five groups, including equal numbers of NS and NNS, to discuss solutions to the situations presented in the sketches. As a communication aid, all participants were given multilingual feeling cards, and interpreters were provided for groups that included NNS at the beginner level of Japanese. We analyzed the group dialogues, finding that the NNS began as side-participants, and that the use of feeling cards by NNS had effects including expanding the engagement area which had been formed only by the NS to include NNS and demonstrating understanding of NS's speech by NNS. Thus, the feeling cards tool allowed NNS at the beginner level to shift from peripheral to central participation without the need for interpretation, which in turn influenced the later development of the dialogues.

In daily conversation, listeners respond to speakers in various ways. Such reactions are called listener responses, and especially, aizuchi and laughter are typical listener responses. Listeners sometimes use single-modal response, and sometimes multi-modal responses simultaneously. However, there have not been many studies analyzing listener responses from a multimodal perspective. In this study, in order to clarify one aspect of the phenomenon of listeners speaking with laughter, we focused on repetitive utterances and laughter, and analyzed how listeners use repetitive utterances with laughter at the same time. As a result, the analysis revealed that repetitive utterances with laughter enable listeners to laugh at utterances that are not laughable or at positions that are not immediately after the preceding utterance.

The purpose of this study is to examine the relationship between the theme of the discourse based on the Coimagintion Method, a conversation support method aimed at activating cognitive functions of older adults, and the rhetorical functions of the utterances in that discourse. The taxonomy of Rhetorical Function Analysis was applied to the analysis. Rhetorical Function Analysis is a modification of the taxonomy of Rhetorical Unit Analysis, a discourse analysis method within the framework of Systemic Functional Linguistics, and within the framework of Japanese grammar. In Rhetorical Function Analysis, rhetorical functions and decontextualization indices are identified based on the combination of speech functions, predicate tenses, and subject and theme categories in the unit of analysis called the message (clause). In the Coimagintion Method, participants are required to bring pictures related to an assigned theme, give a short speech on the topic, answer questions from other participants, and write a short essay after the session is over. In this paper, we analyzed short speeches and essays on nine themes given by 12 participants. First, we confirmed that the relationships between themes and rhetorical functions observed in the pilot study of three themes by four participants were generally found in the present data of 12 participants. We then examined the nine themes, finding that there are rhetorical functions that tend to be used in each case: when talking about one's preferences and habits, when talking about one's future actions, when looking back on past events, and when talking about content not related to the individual. These results suggest that the thematic setting of the Coimagination Method discourse could adjust the rhetorical functions that are likely to be used in discourse.

This paper aims to analyze the interactive function of chant employed during the exchange of heavy objects, as exemplified in Abe's work [2], within the context of a local community. To achieve this objective, a specific case study is drawn upon: the "boya" delivery scene within the preparatory stages of a festival, as outlined by Abe [2]. Through an examination of the interactional dynamics inherent in the vocal calls during this scene, the study identifies potential interactional functions. These functions encompass the following possibilities: firstly, signaling the transition of the "boya" carrier to a second-person role; secondly, indicating the bearer's comprehension of the delivery method; thirdly, foreshadowing the imminent initiation of delivery to a third party; and fourthly, establishing a rhythmic tempo for the activity. The research proposes an ordering of these functions. Lastly, the discourse delves into the significance of vocalizations within local communities. This inquiry is approached from two key perspectives: the contributions of vocalizations to the components of a festival, and their role as a means for external observers to discern and acknowledge the essence of the festival.

Although machine learning-based speech enhancement has been reported to have some degree of effectiveness in handling non-stationary noise and to outperform statistical methods such as the spectral subtraction based on the Minimum Mean Square Error (MMSE), there is a phenomenon where the performance is limited by the resolution of the Discrete Fourier Transform (DFT), a commonly used method for analyzing input signals. This limitation is particularly prominent in noisy environments with strong non-stationarity. To face with this problem, we propose to use two DFTs with difference size to analize input signal for machine learning based speach enhancement. As the result, we achieved up to 2.6 dB improvement in average Segmental Signal-to-Noise Ratio (Seg.SNR) across ten different noise environments when the input signal SNR was 0 dB.

In recent years, remote meeting has become widely used, and solutions for analyzing spoken dialogue are becoming more widespread.For example, in the context of business negotiations, a key focus lies in developing solutions to assess conversational skills through the extraction and summarization of linguistic features. Additionally, paralinguistic information holds significant value in analyzing spoken dialogue as it offers insights into a speaker's impressions and emotions,manifesting in various conversational cues such as laughter, filler words, and gasps. In this work, for the purpose of detecting laughter, we propose a detection system based on a pre-trained speech recognition model and a semi-supervised learning method using weakly labeled data in which the laughter interval is unclear. We conduct experiments on a corpus of Japanese conversation and show that the proposed method outperforms conventional methods. Moreover, we discuss the diversity of speakers, cultures, and sound collection environments as factors affecting the laughter detection performance.

In dialogue systems, it is especially important to ask questions that can be answered with "yes," "no," or "I don't know" intentions (YES-NO questions) in order to confirm the user's intentions and status, and to accurately interpret the intention of the user's response. In this study, we aimed to estimate the response intention to these YES-NO questions with high accuracy. Specifically, we created a dialogue corpus (Japanese Yes-No question-answer pairs), designed several intention estimators using a large-scale language model, and compared and evaluated their accuracy. As a result, the GPT model with the addition of an all-combining layer showed the highest estimation accuracy, achieving an accuracy of 91\%. Comparison among the estimators also confirmed a tendency to mis-estimate "I don't know" when using prompt programming with the GPT model. This study contributes to the ability of dialogue systems to improve their ability to estimate the intent of interrogative sentences and provides new metrics and insights into the performance evaluation of machine learning models.

Rapport is a harmonious relationship with others. We aim to automatically estimate a speaker's subjective rapport using their nonverbal behavior during a conversation. Rapport estimation is generally formulated as a regression. However, it is difficult to learn a mapping between nonverbal behaviors and rapport ratings during a conversation because of individual differences in rapport ratings. To alleviate this problem, we formulate rapport estimation as learning to rank. Learning to rank avoids the problem of individual differences in rapport ratings using preference learning, which learns the ordinal relationship between two conversations based on rapport reported by the same user. To evaluate the proposed method, we used a dataset consisting of first-meeting conversations and friend conversations that includes subjective rapport ratings. We compared the proposed model with the regression model using metrics for ranking. The result indicates that the proposed model is more suitable than the regression model for rapport estimations.