Elamur Research⌕
☰
home / papers / 2026 / Construction and Analysis of Japanese Parent-Child Dialogic Reading Corpus for Conversational Agents

Construction and Analysis of Japanese Parent-Child Dialogic Reading Corpus for Conversational Agents

Summary

Dialogic reading, where a parent and child talk about a picture book while reading it, supports early language growth. Existing reading-support agents, however, follow scripted question sequences and cannot handle this kind of flexible exchange. To inform better agent design, the authors built a multimodal Japanese corpus of real parent-child reading sessions.

Thirty-six dyads, with children aged 3–7, were recorded in separate soundproof booths connected by video chat. This produced clean per-speaker audio and video. Parents were asked to ask quizzes and questions during or after reading. Transcripts were created with whisper-large-v3 and then manually corrected. Utterances were labeled with a 13-tag dialogue-act scheme derived from INCA-A, assigned automatically by gemma-3-27b-it and checked against two expert annotators. Parent prompts were further classified as quizzes or open questions and categorized by type.

The authors split the dialogues by how much the child spoke. Parents' interaction strategies and question content differed between these groups. This suggests that reading agents should adapt their behavior to each child's speaking style rather than following a single fixed script.

Key points

  • The corpus contains 70 dialogues from 36 native Japanese dyads, totaling 10.7 hours (about 9.17 min per dialogue). It includes 2,699 child and 8,495 parent utterances.
  • Audio was captured on eight channels (48 kHz, 32-bit float WAV), with video in MP4. In the first session every dyad read the same bestseller. In the second, the child chose from five comparable books.
  • The corpus has 11,194 dialogue-act tags. Agreement between the two experts was Cohen's κ = 0.826, while automatic tags agreed with the humans at an average κ = 0.591.
  • Parents posed 323 quizzes and 196 questions. Each was classified as WH- or Yes/No, and quizzes were also mapped to the seven FairytaleQA categories (e.g., Character, Causal Relationship, Prediction).
  • Stated limitations are the small amount of data, especially from male parents (only 3 of 36), and an analysis restricted to linguistic features. The authors plan to study prosody and nonverbal cues in future work.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Linguistics & languages

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.63317/3khbbpw33gyp

@inproceedings{nakagi2026construction,
  title = {{Construction and Analysis of Japanese Parent-Child Dialogic Reading Corpus for Conversational Agents}},
  author = {Nakagi, Yuko and Chiba, Yuya and Fujita, Sanae and Araki, Shoko},
  booktitle = {Proceedings of the Fifteenth Language Resources and Evaluation Conference},
  year = {2026},
  doi = {10.63317/3khbbpw33gyp},
  url = {https://doi.org/10.63317/3khbbpw33gyp}
}

References

24 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Adriana G. Bus, Marinus H. van IJzendoorn, and Anthony D. Pellegrini. 1995. Joint book reading makes for success in learning to read: A meta-analysis on intergenerational transmission of literacy. Review of Educational Research, 65(1):1– 21. find
  2. Huili Chen, Yue Zhang, Felix Weninger, Rosalind Picard, Cynthia Breazeal, and Hae Won Park. 2020. Dyadic speech-based affect recognition using DAMI-P2C parent-child multimodal interaction dataset. In Proceedings of the 2020 International Conference on Multimodal Interaction, page 97–106. find
  3. Mirela Conica, Linda Kelly, Elizabeth Nixon, and Jean Quigley. 2023. Father and toddler language during shared book reading with text-based and wordless picture books. Read. Res. Q., 58(4):655–667. find
  4. Warwick B Elley. 1989. Vocabulary acquisition from listening to stories. Read. Res. Q., 24(2):174. find
  5. Brad M Farrant and Stephen R Zubrick. 2012. Early vocabulary development: The importance of joint attention and parent-child book reading. First Lang., 32(3):343–364. find
  6. Zoe M Flack, Andy P Field, and Jessica S Horst. 2018. The effects of shared storybook reading on word learning: A meta-analysis. Dev. Psychol., 54(7):1334–1346. find
  7. Sanae Fujita, Tessei Kobayashi, Yasuhiro Minami, and Hiroaki Sugiyama. 2015. Target age estimation of texts for children. Cognitive Studies: Bulletin of the Japanese Cognitive Science Society, 22(4):604–620. find
  8. Laura M Justice and Helen K Ezell. 2000. Enhancing children’s print and word awareness through home-based parent intervention. Am. J. Speech. Lang. Pathol., 9(3):257–269. find
  9. Aishwarya Kamath, Johan Ferret, Shreya Pathak, Nino Vieillard, Ramona Merhej, Sarah Perrin, Tatiana Matejovicova, Alexandre Ramé, Morgane Rivière, and Louis Rouillard et al. 2025. Gemma 3 technical report. arXiv preprint arXiv:2503.19786. find
  10. Hüseyin Kotaman. 2008. Impacts of dialogical storybook reading on young children’s reading attitudes and vocabulary development. Reading Improvement, 43:134–145. find
  11. Kikuo Maekawa. 2003. Corpus of spontaneous Japanese: Its design and evaluation. In Proceedings of the ISCA & IEEE Workshop on Spontaneous Speech Processing and Recognition, pages 7–12. find
  12. Toyomi Meguro, Ryuichiro Higashinaka, Kohji Dohsaka, Yasuhiro Minami, and Hideki Isozaki. 2009. Analysis of listening-oriented dialogue for building listening agents. In Proceedings of the SIGDIAL 2009 Conference, pages 124–127. find
  13. Joseph E Michaelis and Bilge Mutlu. 2017. Someone to read with: Design of and experiences with an in-home learning companion robot for reading. In Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems, pages 301–312. find
  14. Paul L. Morgan and Catherine R. Meier. 2008. Dialogic reading’s potential to improve children’s emergent literacy skills and behavior. Preventing School Failure: Alternative Education for Children and Youth, 52(4):11–16. find
  15. A Ninio, C E Snow, B A Pan, and P R Rollins. 1994. Classifying communicative acts in children’s interactions. J. Commun. Disord., 27(2):157–187. find
  16. Claire Noble, Giovanni Sala, Michelle Peter, Jamie Lingwood, Caroline Rowland, Fernand Gobet, and Julian Pine. 2019. The impact of shared book reading on children’s language skills: A meta-analysis. Educational Research Review, 28:100290. find
  17. Anastasia Stoops, Mengqian Wu, In Ho Ted Jung, and Jessica L Montag. 2024. A novel corpus of naturalistic picture book reading with 2-to-3 year old children. Language Development Research, 4(1). find
  18. Medha Tare and Susan A Gelman. 2011. Bilingual parents’ modeling of pragmatic language use in multiparty interactions. Appl. Psycholinguist., 32(4):761–780. find
  19. G J Whitehurst, F L Falco, C J Lonigan, J E Fischel, B D DeBaryshe, M C Valdez-Menchaca, and M Caulfield. 1988. Accelerating language development through picture book reading. Dev. Psychol., 24(4):552–559. find
  20. Ying Xu, Joseph Aubele, Valery Vigil, Andres S. Bustamante, Young-Suk Kim, and Mark Warschauer. 2022a. Dialogue with a conversational agent promotes children’s story comprehension via enhancing engagement. Child Development, 93(2):e149–e167. find
  21. Ying Xu, Dakuo Wang, Penelope Collins, Hyelim Lee, and Mark Warschauer. 2021. Same benefits, different communication patterns: Comparing children’s reading with a conversational agent vs. a human partner. Computers Education, 161:104059. find
  22. Ying Xu, Dakuo Wang, Mo Yu, Daniel Ritchie, Bingsheng Yao, Tongshuang Wu, Zheng Zhang, Toby Jia-Jun Li, Nora Bradford, Branda Sun, Tran Bao Hoang, Yisi Sang, Yufang Hou, Xiaojuan Ma, Diyi Yang, Nanyun Peng, Zhou Yu, and Mark Warschauer. 2022b. Fantastic questions and where to find them: FairytaleQA… find our card
  23. Bingsheng Yao, Dakuo Wang, Tongshuang Wu, Zheng Zhang, Toby Jia-Jun Li, Mo Yu, and Ying Xu. 2022. It is AI’s turn to ask humans a question: Question-answer pair generation for children’s story books. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics, volume 1… find our card
  24. Zheng Zhang, Ying Xu, Yanhao Wang, Bingsheng Yao, Daniel Ritchie, Tongshuang Wu, Mo Yu, Dakuo Wang, and Toby Jia-Jun Li. 2022. StoryBuddy: A human-AI collaborative chatbot for parent-child interactive storytelling with flexible parental involvement. In Proceedings of the 2022 CHI Conference on Human… find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/papers/nakagi-2026-construction-analysis-japanese-parent-child/