Profiling of Intertextuality in Latin Literature Using Word Embeddings

Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 4900–4907 June 6–11, 2021. ©2021 Association for Computational Linguistics 4900 Profiling of Intertextuality in Latin Literature Using Word Embeddings Patrick J. Burns,1 James A. Brofos,2 Kyle Li,3 Pramit Chaudhuri,1 Joseph P. Dexter4† 1 Department of Classics, University of Texas at Austin 2 Department of Statistics and Data Science, Yale University 3 Department of Computer Science, Columbia University 4 Data Science Initiative and Department of Human Evolutionary Biology, Harvard University † Corresponding author: jdexter@fas.harvard.edu Abstract Identifying intertextual relationships between authors is of central importance to the study of literature. We report an empirical analysis of intertextuality in classical Latin literature us- ing word embedding models. To enable quan- titative evaluation of intertextual search meth- ods, we curate a new dataset of 945 known parallels drawn from traditional scholarship on Latin epic poetry. We train an optimized word2vec model on a large corpus of lemma- tized Latin, which achieves state-of-the-art per- formance for synonym detection and outper- forms a widely used lexical method for inter- textual search. We then demonstrate that train- ing embeddings on very small corpora can cap- ture salient aspects of literary style and ap- ply this approach to replicate a previous in- tertextual study of the Roman historian Livy, which relied on hand-crafted stylometric fea- tures. Our results advance the development of core computational resources for a major pre- modern language and highlight a productive avenue for cross-disciplinary collaboration be- tween the study of literature and NLP. 1 1 Introduction In “Lonesome Day Blues,” Bob Dylan sings, “I’m gonna spare the defeated...I am goin’ to teach peace to the conquered / I’m gonna tame the proud.” This lyric echoes a passage from Vergil’s ancient Latin epic, the Aeneid, as translated by Allen Mandel- baum: “to teach the ways of peace to those you conquer, / to spare defeated peoples, tame the proud” (Thomas, 2012). Such allusions or “inter- texts” transmit ideas across space and time, diverse media, and languages. Although researchers fo- cus on those intertextual connections felt to have special literary significance for the works at hand, in principle intertextuality refers to any verbal or 1Code and data are available at https://github.c om/QuantitativeCriticismLab/NAACL-HLT-20 21-Latin-Intertextuality. semantic resemblance within the literary system, ranging from direct quotation to topical similari- ties (Kristeva, 1980; Juvan, 2009). Given the im- portance of intertextual criticism to literary study, computational identification of text reuse in liter- ature is an active area of research (Bamman and Crane, 2008; Forstall and Scheirer, 2019). Classical Latin literature is a highly influential tradition characterized by an extraordinary density of allusions and other forms of text reuse (Hinds, 1998). The most widely used tools for the detection of Latin intertextuality, such as Tesserae and Dio- genes, rely on lexical matching of repeated words or phrases (Coffee et al., 2012, 2013; Heslin, 2019). In addition to these core methods, other research has explored the use of sequence alignment (Chaud- huri …

Content truncated to 500 words