Blue eyes and porcelain cheeks: Computational extraction of physical descriptions from Dutch chick lit and literary novels

Attentiveness to physical appearance is seen as an important motif in chick lit. However, this has never been researched in an actual comparison with literary novels. Our goal is to make a comparison between the two genres concerning the importance of physical appearance of characters. In this article, as a first step in researching this topic, we consider the task of identifying sentences that contain descriptions of a character’s physical appearance in a set of novels in both genres. We manually annotate sentences where the appearance of a character is described. We then apply a machine learning methodology and manually written lexical–syntactic queries to automate the extraction of these sentences. Contrary to our expectations, the queries outperform the classifier. When we use a combination of the two automated methods, the extraction of physical descriptions from chick lit is more successful. This suggests that such descriptions are more uniform in character.