Distinguishing Fictional Voices: a Study of Authorship Verification Models for Quotation Attribution

Recent approaches to automatically detect the speaker of an utterance of direct speech often disregard general information about characters in favor of local information found in the context, such as surrounding mentions of entities. In this work, we explore stylistic representations of characters built by encoding their quotes with off-the-shelf pretrained Authorship Verification models in a large corpus of English Figure 1: Example of quotation attribution on an excerpt novels (the Project Dialogism Novel Corpus). of Pride and Prejudice by Jane Austen (1813). Results suggest that the combination of stylistic Underlined text are identified mentions, and arrows link and topical information captured in some quotes to their relevant entity mention (solid arrows are of these models accurately distinguish characters explicit references and dashed arrows are anaphoric references).