@inproceedings{xu-etal-2026-cross,
    title = "Cross-Modal Taxonomic Generalization in (Vision-) Language Models",
    author = "Xu, Tianyang  and
      Sandoval-Casta{\textbackslash}{\textasciitilde}neda, Marcelo  and
      Livescu, Karen  and
      Shakhnarovich, Greg  and
      Misra, Kanishka",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.acl-long.742/",
    pages = "16319--16337",
    ISBN = "979-8-89176-390-6",
    abstract = "What is the interplay between semantic representations learned by language models (LM) from surface form alone to those learned from more grounded evidence? We study this question for a scenario where part of the input comes from a different modality{---}in our case, in a vision-language model (VLM), where a pretrained LM is aligned with a pretrained image encoder. As a case study we focus on the task of predicting hypernyms of objects represented in images. We do so in a VLM setup where the image encoder and LM are kept frozen, and only the intermediate mappings are learned. We progressively deprive the VLM of explicit evidence for hypernyms, and test whether the knowledge is recoverable from the LM. We find that the LMs we study can recover this knowledge and generalize even in the most extreme version of this experiment (when the model receives no evidence of a hypernym during training). Additional experiments suggest that this cross-modal taxonomic generalization persists under counterfactual image{--}label mappings only when the counterfactual data have high visual similarity within each category. Taken together, these findings suggest that cross-modal generalization in LMs arises from an interaction between linguistic structure and the information present in the input."
}