Living Machines: A study of atypical animacy

Proceedings of the 28th International Conference on Computational Linguistics, pages 4534–4545 Barcelona, Spain (Online), December 8-13, 2020 4534 Living Machines: A study of atypical animacy Mariona Coll Ardanuy The Alan Turing Institute Queen Mary University of London mcollardanuy@turing.ac.uk Federico Nanni The Alan Turing Institute fnanni@turing.ac.uk Kaspar Beelen The Alan Turing Institute Queen Mary University of London kbeelen@turing.ac.uk Kasra Hosseini The Alan Turing Institute khosseini@turing.ac.uk Ruth Ahnert Queen Mary University of London r.r.ahnert@qmul.ac.uk Jon Lawrence University of Exeter j.lawrence3@exeter.ac.uk Katherine McDonough The Alan Turing Institute Queen Mary University of London kmcdonough@turing.ac.uk Giorgia Tolfo The British Library giorgia.tolfo@bl.uk Daniel CS Wilson The Alan Turing Institute Queen Mary University of London dwilson@turing.ac.uk Barbara McGillivray The Alan Turing Institute University of Cambridge bmcgillivray@turing.ac.uk Abstract This paper proposes a new approach to animacy detection, the task of determining whether an entity is represented as animate in a text. In particular, this work is focused on atypical animacy and examines the scenario in which typically inanimate objects, specifically machines, are given animate attributes. To address it, we have created the first dataset for atypical animacy detection, based on nineteenth-century sentences in English, with machines represented as either animate or inanimate. Our method builds on recent innovations in language modeling, specifically BERT contextualized word embeddings, to better capture fine-grained contextual properties of words. We present a fully unsupervised pipeline, which can be easily adapted to different contexts, and report its performance on an established animacy dataset and our newly introduced resource. We show that our method provides a substantially more accurate characterization of atypical animacy, especially when applied to highly complex forms of language use. 1 Introduction Animacy is the property of being alive. Although the perception of a given entity as animate (or not) tends to align correctly with its biological animacy, discrepancies are not uncommon. These may arise from the ways different entities are unconsciously perceived, or from the deliberately figurative use of animate expressions to describe inanimate entities (or vice versa). Machines sit at the fuzzy boundary between animacy and inanimacy (Turing, 1950; Yamamoto, 1999). In this paper, we provide a basis This work is licensed under a Creative Commons Attribution 4.0 International Licence. Licence details: http:// creativecommons.org/licenses/by/4.0/. 4535 for examining how machines were imagined during the nineteenth century as everything from lifeless mechanical objects to living beings, or even human-like agents that feel, think, and love. We focus on texts from nineteenth-century Britain, a society being transformed by industrialization, as a good candidate for studying the broader issue. This paper applies state-of-the-art contextualized word representations, trained using the BERT archi- tecture (Devlin et al., 2019), to animacy detection. In contrast to previous research, this paper provides an in-depth exploration of the ambiguities and figurative aspects that characterize animacy in natural language, and analyzes how context shapes animacy. Context is constitutive of meaning (Wittgenstein, 1921, 3.3), an observation acknowledged by generations of scholars, but which is still difficult to apply to its full extent in computational models of language. We show how the increased …

Content truncated to 500 words