2022年度採択の日本学術振興会科学研究費助成事業 学術変革領域研究(B) 「言語相互行為における身振りと手話を対象とした身体記号学(略称：身体記号学)」のクロージングカンファレンスを実施します。

日程： 2025年2月10日(月)，11日(火・祝)
会場： 国立情報学研究所　12F会議室(1208/1210)　(＋Zoomを用いたハイブリッド形式)
言語： 日本語(JPN), 日本手話(JSL), 国際手話(IS)（各セッションの主たる言語は[ ]内）
参加申込（無料）：対面・オンラインともに定員に達しましたので参加申し込みを締め切りました。

2月10日

プレイベント：日英クロスサインプロジェクト研究成果報告[IS] (通訳：現地 JSL，JPN/オンライン JSL，JPN)
10:00-10:10 概要説明(大杉豊)
10:10-10:40 プロジェクト説明：Robert Adam (Heriot-Watt University)
10:40-11:10 研究成果報告①：岡田智裕
11:10-11:40 研究成果報告②：Sujit Sahasrabudhe (Heriot-Watt University)
11:40-12:00 コメント(Maartje De Meulder, HU–University of Applied Sciences Utrecht/Heriot-Watt University)

身体記号学クロージングカンファレンス第一部:AIを味方にするためのコミュニケーション研究[JPN/JSL] (通訳：現地 JSL, IS/オンライン JSL)
13:30-13:50 概要説明(菊地浩平)
13:50-15:30 話題提供 (4件，各20分)
発表①「日本手話話し言葉コーパス (JSLコーパス) データセットの概要と事例紹介」(岡田智裕)(JSL)
発表②「科学未来館SCコーパス データセットの概要と事例紹介」(坂井田瑠衣)
発表③「AIを利用した指さし動作認識の概要」 (Junwen Mo, Minh-Duc Vo, 中山英樹)
発表④「指標とAI認識: 記号論の観点から」(榎本剛士)
15:30-15:50 休憩
15:50-17:20 パネルディスカッション（指定討論者：喜多壮太郎, University of Warwick）
18:00 懇親会

2月11日

身体記号学クロージングカンファレンス第二部:研究成果報告[JPN] (通訳：現地 JSL, IS /オンライン JSL)
10:00-10:15 領域全体成果報告(坊農真弓)
10:15-10:35 A01成果報告「記号論に基づいたアノテーションデザイン」(井上昂治，高梨克也)
10:35-10:55 A02成果報告「日本手話日常会話コーパス」 (坊農真弓，岡田智裕)
10:55-11:05 休憩
11:05-11:25 A03成果報告 「LLMのリソースとしてのYouTube」（中山英樹，Minh-Duc Vo, Junwen Mo）
11:25-11:45 B01成果報告 「Tracrin」(牧野遼作)
11:45-12:00 外部アドバイザーによるコメント(伝康晴)
12:00-13:00 昼食休憩 (60分)

身体記号学クロージングカンファレンス第三部：招待講演[JPN/IS] (通訳：現地 JSL/オンライン JSL)
13:00-14:00 招待講演①「ろう者の生活・アクセス・主体性：AI時代のコミュニケーションを再考する」(Dr. Maartje De Meulder, HU–University of Applied Sciences Utrecht/Heriot-Watt University) (IS)
14:00-15:00 招待講演②「AI技術と手話： 慣習化された手話を越えて」(Prof. Kearsy Cormier, DCAL–Deafness Cognition and Language Research Centre, University College London)(IS)
15:00-15:15 休憩(15分)
15:15-16:15 招待講演③「ジェスチャー・メタファー・空間言語」(Prof. Sotaro Kita, University of Warwick)(JPN)
16:15-16:20 クロージング(JPN)(坊農真弓)

Title :Deaf lives, access, and agency: rethinking communication in the age of AI
ろう者の生活・アクセス・主体性：AI時代のコミュニケーションを再考する

Abstract : The rapid advancement of AI-driven language technologies, including generative AI tools, is transforming how deaf people navigate communication. Tools like real-time captions and speech recognition facilitate access, while emerging communicative agents such as conversational AI systems and experimental signing avatars act as communicative agents across multiple modalities and degrees of embodiment, from text-based interfaces to human-like robots. Meanwhile, human sign language interpreting, a cornerstone of linguistic access, faces systemic pressures across Western societies due to supply-demand imbalances and heavily reliance on interpreters. This presentation explores the implications of these shifts, including the creation of new access hierarchies that privilege some deaf users while marginalizing others. It examines how technoableism and technochauvinism in technology development can overestimate AI capabilities, risking harm through compromised access and the erosion of hard-won accessibility rights. These techologies also reshape language ideologies, and redefine the role of sign language interpreting as a profession. Additionally, the presentation addresses systemic biases in the field of sign language AI and the need for sustainable, equitable research practices, centering deaf leadership so that impactful design decisions are made by those with a larger stake.

BIO : Dr. Maartje De Meulder is a senior researcher at the University of Applied Sciences Utrecht and a leading scholar at the intersection of Deaf Studies, Disability Studies, Sign Language Interpreting Studies and Human-Computer Interaction. Her interdisciplinary work addresses key societal challenges facing deaf communities, with a current focus on the ethics of AI-driven language technologies and their impact on deaf communication and access practices. As one of the few deaf researchers working on sign language technologies and AI, she has pioneered ethical and responsible approaches to their development, critically examining their socio-political implications. Her research bridges traditionally siloed fields, focusing on the intersections of language, technology and disability, while advocating for deaf leadership and sustainable, equitable research practices. Dr. De Meulder's work is widely published, and she actively supports the capacity building of deaf schoalrs through the Dr Deaf workshops.

Title: AI technologies with sign languages: Moving beyond conventionalised signs
AI技術と手話：慣習化された手話を越えて

Abstract: Sign languages are rich visual languages, comprising many different types of signs. These include lexical signs similar to words in spoken languages, the kinds of signs one might find in dictionaries, and also fingerspelling via manual representation of manual letters or characters – both of which are highly conventionalised in form and meaning. But a high proportion of any sign language production involves partly lexical or non-lexical signs – including pointing signs, depicting/classifier signs, and manual enactments (constructed action). Work on AI with sign languages has focused in large part on the conventionalised aspects of sign formation, particularly lexical signs and fingerspelling – in large part because these are the easiest to translate to/from a spoken/written language e.g. via dictionary lookups, with minimal context. Partly/non-lexicalised signs pose a much bigger challenge with AI because they vary so much in form and require considerable context to be interpreted. This presentation will explain these challenges – both for manual signs and nonmanual expressions - and consider what is needed to address them.

BIO:Kearsy Cormier is Professor of Sign Linguistics at in the Department of Linguistics within the Division of Psychology and Language Sciences at University College London. She is also Director of the UCL Deafness Cognition and Language Research Centre, the largest research centre in Europe focusing on deaf people's communication and cognition. Her work in sign language linguistics is broad and covers all areas of major subfields of linguistics. She is director of the national British Sign Language Corpus and BSL Signbank, and using these and other resources is working with computer scientists towards automatic translation between signed and spoken languages.

Title: Gesture, metaphor, and spatial language
ジェスチャー・メタファー・空間言語

Abstract: I will discuss how co-speech (i.e., speech-accompanying) gestures relate to language and conceptualisation underlying language. I will focus on “representational gestures”, which can depict motion, action, and shape or can indicate locations (i.e., “iconic” and “deictic” gestures in McNeill's 1992 classification). I will provide evidence for the following two points. Various aspects of language shape co-speech gestures. Conversely, the way we produce co-speech gestures can shape language. I will discuss these issues in relation to manner and path in motion event descriptions, clause-linkage types in complex event descriptions, and metaphor. I will conclude that gesture and language are parts of a "conceptualisation engine”, which takes advantage of unique strengths of spatio-motoric representation and linguistic representation.

BIO: Sotaro Kita is Professor of Psychology of Language at the University of Warwick, in the UK. After a bachelor and a master's degree in mathematical engineering and information engineering from University of Tokyo. He obtained a PhD in linguistics and psychology from the University of Chicago, in 1993. He established and led the "Gesture Project" at Max Planck Institute for Psycholinguistics, in the Netherlands, from 1993 to 2003. Then, he held a faculty position at University of Bristol and University of Birmingham, before joining University of Warwick. He served as the President of the International Society of Gesture Studies and also as the Editor of the journal, GESTURE. He has investigated how gesture relates to various aspects of language: motion events (Kita & Özyürek, 2003, Journal of Memory and Language) and metaphor (Argyriou, Mohr & Kita, 2017, JEP: Learning, Memory and Cognition). He has also investigated how gesture varies cross-culturally (Kita, 2009, Language and Cognitive Processes), how gesture production shapes gesturer's conceptualization (Kita, Alibali, & Chu, 2017, Psychological Review) and how deaf Nicaraguan children turned gesture into an emerging sign language (Senghas, Kita & Özyürek, 2004, Science). His current research topics include how children's word learning can be facilitated by sound symbolism and gesture, how communicative contexts influence infants' and adults' gesture production, and how “silent gestures” show language-like features.

主催： 日本学術振興会科学研究費助成事業 学術変革領域研究(B) 「言語相互行為における身振りと手話を対象とした身体記号学(略称：身体記号学)」(領域代表者：坊農真弓)
https://research.nii.ac.jp/EmSemi/

共催： 日本学術振興会国際共同事業　英国との国際共同研究プログラム（JRP-LEAD with UKRI）「コロナ禍／コロナ後におけるオンライン会議状況でのクロスサイニング現象の理解」(研究代表者：坊農真弓)
https://www.cross-sign.nii.ac.jp