Abstract
It has been suggested that computational tools can enable us to identify narratives in large data sets, and that we can use those narratives to facilitate social science research at scale in the same way as distant reading techniques facilitate literary analysis at scale. In this paper I attempt to determine whether a Bag of Words approach to natural language processing can identify such narratives. I argue that a Bag of Words approach can capture some features of narratives but fails on others, especially the extent to which the meaning of the parts of narratives are dependent on the ending.
Get full access to this article
View all access options for this article.
