@inproceedings{rossini-van-der-plas-2026-124,
    title = "From 124 Million Tokens to 1,021 Neologisms: A Large-Scale Pipeline for Automatic Neologism Detection",
    author = "Rossini, Diego  and
      van der Plas, Lonneke",
    editor = "Oleskeviciene, Giedre Valunaite  and
      Giouli, Voula  and
      Armaselu, Florentina  and
      Liebeskind, Chaya  and
      McGillivray, Barbara",
    booktitle = "Proceedings of the Workshop Neology and Large Language Models",
    month = may,
    year = "2026",
    address = "Palma, Mallorca (Spain)",
    publisher = "ELRA Language Resources Association (ELRA)",
    url = "https://preview.aclanthology.org/test-year-match/2026.neollm-1.1/",
    doi = "10.63317/4o6ks86o293r",
    pages = "1--15",
    abstract = "We present a scalable, modular pipeline for automatic neologism detection that combines rule-based filtering with LLM classification. The pipeline is grounded in two complementary word-formation frameworks, grammatical and extra-grammatical morphology, which jointly define the scope of what counts as a neologism and inform a four-class classification scheme (NEOLOGISM, ENTITY, FOREIGN, NONE). While designed to be modular and transferable at the architectural level, the pipeline is instantiated on 527 million English-language Reddit posts spanning 2005{--}2024. From this corpus, we extract 124.6 million unique tokens and reduce them by over 99.99{\%} to yield 1,021 neologism candidates, a set small enough for manual expert verification. Multiple LLMs independently classify each candidate via majority vote, with a final verification step, revealing substantial cross-model disagreement and highlighting the challenge of operationalizing neologism detection at scale. Manual annotation of all 1,021 candidates confirms that 599 (58.7{\%}) are genuine lexical innovations."
}