Two weeks ago, I attended the 6th edition of the workshop for Computational Models of Narrative in Atlanta. Together with my colleague Marten van der Meulen, I presented a paper on detection animate entities in stories. Animacy is often conceived as a categorical binary distinction, i.e. a chair is inanimate and a monkey, for example, is animate. This view has been challenged, however, by researchers from different fields. To give an example from linguistic typology, it is well-known that not all languages award animacy to the same entities in different grammatical categories. In Persian, for example, a tree is grammatically marked as animate whereas a flower is inanimate. In the paper we argue that animacy should be treated as epistemological stances rather than fixed states in the world: not ineffable qualia but behavioral capacity defines our stance towards objects.
The paper presents a linguistically uninformed computational model for animacy classification. The model makes use of word $n$-grams in combination with lower dimensional word embedding representations that are learned from a web-scale corpus. We compare the model to a number of linguistically informed models that use features such as dependency tags and show competitive results. We apply our animacy classifier to a large collection of Dutch folktales to obtain a list of all characters in the stories. We then draw a semantic map of all automatically extracted characters which provides a unique entrance point to the collection.
To our surprise, the paper won the best paper award of the workshop!
Karsdorp, Folgert & Van der Meulen, Marten & Meder, Theo & Van den Bosch (2015). ‘Animacy detection in Stories’. In Mark A. Finlayson, Ben Miller, Antonio Lieto, and Remi Ronfard (ed.). Proceedings of the Workshop on Computational Models of Narrative (CMN’15), May 26-28, 2015, Atlanta, USA, pp. 82–97. (Full text)