Oxford Digital Ethnography Group Seminar Series

For every species, a language: AI, interspecies communication, and the democratization of language to nonhuman worlds

About

One of the effects of the rapid transformation in artificial intelligence capabilities in recent years has been an explosion of scholarly and popular interest in the communication systems of different animal species. While there have been longstanding research projects investigating cetacean, bird, and elephant communications, a wide range of ethologists, animal rights activists, political philosophers, app builders, and others now argue that every species has a “language” of its own that can be understood with machine learning. Within this framework, the natural world is one in which many kinds of living things will be known and will be saved by listening to their languages. AI-enabled analysis of vast quantities of behavioral data holds out the hope of defining universes of semanticized “concepts” through which humans and nonhumans will be able to create expanded democracies and stave off ecological crisis. But rather than one new vision of language emerging out of this kind of work, I want to emphasize here the ways that language is becoming polymorphous in its definitions and limits even as people insist upon “language” remaining something that animals have. What is at stake for these utopian projects of democratic inclusion in insisting animals have language, even as the definition of language itself is under major revision?

Courtney Handman is an Associate Professor of Anthropology at the University of Texas at Austin. Having written about Christian missionary projects to teach Papua New Guineans how to communicate with God (Critical Christianity: Translation and Denominational Conflict in Papua New Guinea, 2015) and about the communicative infrastructures of colonialism (Circulations: Modernist Imaginaries of Colonialism and Decolonization, 2025), she now is researching the ways that people are learning how to communicate with and through large language models.