SLäNDa: An Annotated Corpus of Narrative and Dialogue in Swedish Literary Fiction

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 826–834 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 826 SL¨aNDa: An Annotated Corpus of Narrative and Dialogue in Swedish Literary Fiction Sara Stymne and Carin ¨Ostman Linguistics and Philology, Scandinavian Languages Uppsala University, Sweden sara.stymne@lingfil.uu.se, carin.ostman@nordiska.uu.se Abstract We describe a new corpus, SL¨aNDa, the Swedish Literary corpus of Narrative and Dialogue. It contains Swedish literary fiction, which has been manually annotated for cited materials, with a focus on dialogue. The annotation covers excerpts from eight Swedish novels written between 1879–1940, a period of modernization of the Swedish language. SL¨aNDa contains annotations for all cited materials that are separate from the main narrative, like quotations and signs. The main focus is on dialogue, for which we annotate speech segments, speech tags, and speakers. In this paper we describe the annotation protocol and procedure and show that we can reach a high inter-annotator agreement. In total, SL¨aNDa contains annotations of 44 chapters with over 220K tokens. The annotation identified 4,733 instances of cited material and 1,143 named speaker–speech mappings. The corpus is useful for developing computational tools for different types of analysis of literary narrative and speech. We perform a small pilot study where we show how our annotation can help in analyzing language change in Swedish. We find that a number of common function words have their modern version appear earlier in speech than in narrative. Keywords: literary corpora, direct speech, narrative, dialogue, annotation 1. Introduction In literary studies, as in many other research fields, there has been a trend towards using computational methods. Traditionally, most research was based on close reading of books by researchers, whereas distant reading (Moretti, 2000), the computational treatment of potentially large amounts of literary text, has become an important comple- ment. While unsupervised methods, like topic modelling has been important in such studies, supervised models re- lying on annotated data are also often required. Annotated data is also useful for evaluation of proposed models, re- gardless of their type. In this paper, we describe version 1.0 of SL¨aNDa, the Swedish Literary corpus of Narrative and Dialogue (Stymne and ¨Ostman, 2020). It is made up of eight Swedish literary novels from the late 19th and early 20th centuries, annotated mainly for different aspects of dialogue. The full annotation also contains other cited materials, like thoughts, signs and letters. The main motivation for in- cluding these categories as well, is to be able to identify the main narrative. Such a corpus can be useful for train- ing and evaluating computational models, which in turn al- lows larger scale studies of cited materials, narrative, and dialogue in literature. Annotated literary corpora are quite rare, especially for Swedish, and as such we believe that SL¨aNDa could become a useful resource. We consider cited materials to be all parts that stands out from the main narrative of the novels. The most common type is direct speech from the different characters, forming …

Content truncated to 500 words