@inproceedings{fundal-bizzoni-2026-directional,
    title = "Directional Alignment and Narrative Agency in Human{--}{LLM} Co-Writing",
    author = "Fundal, Halfdan Nordahl  and
      Bizzoni, Yuri",
    editor = {Hamilton, Sil  and
      {\"O}hman, Emily  and
      Hicke, Rebecca M. M.  and
      Bizzoni, Yuri  and
      Bax, Axel  and
      Matthews, Jacob A.  and
      H{\"a}m{\"a}l{\"a}inen, Mika},
    booktitle = "Proceedings of the 6th International Conference on Natural Language Processing for the Digital Humanities",
    month = jul,
    year = "2026",
    address = "San Diego, USA",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl-workshops/2026.nlp4dh-1.18/",
    pages = "190--201",
    ISBN = "979-8-89176-427-9",
    abstract = "We investigate narrative agency in hu-man{--}LLM creative co-writing, asking whodrives story development in turn-based collabo-ration. Using a new corpus of human{--}LLM co-written stories, we apply sentiment and seman-tic modeling to quantify affective alignmentand semantic novelty in turn-taking, and direc-tional measures to assess which agent shapesnarrative progression. Our results show asym-metric influence: human turns introduce greatersemantic novelty and are more likely to shapesubsequent developments, whereas LLM con-tributions predominantly elaborate on human-introduced elements. At the sentiment level,alignment is also asymmetric, but more bidirec-tional: LLMs exhibit stronger turn-level emo-tional adaptation than humans, but both agentstrack each other{'}s emotional valence and LLMsshow an independent tendency to more pos-itive emotional baselines. These findings in-dicate a complementary division of labor inhuman{--}LLM co-writing, where humans drivenarrative innovation and direction, while LLMsact as adaptive amplifiers that sustain coherenceand elaborate emerging narratives."
}