Metrical Tagging in the Wild: Building and Annotating Poetry Corpora with Rhythmic Features

Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics, pages 3715–3725 April 19 - 23, 2021. ©2021 Association for Computational Linguistics 3715 Metrical Tagging in the Wild: Building and Annotating Poetry Corpora with Rhythmic Features Thomas Haider Department of Language and Literature, Max Planck Institute for Empirical Aesthetics, Frankfurt am Main Institute for Natural Language Processing (IMS), University of Stuttgart thomas.haider@ae.mpg.de Abstract A prerequisite for the computational study of literature is the availability of properly digi- tized texts, ideally with reliable meta-data and ground-truth annotation. Poetry corpora do ex- ist for a number of languages, but larger collec- tions lack consistency and are encoded in vari- ous standards, while annotated corpora are typ- ically constrained to a particular genre and/or were designed for the analysis of certain lin- guistic features (like rhyme). In this work, we provide large poetry corpora for English and German, and annotate prosodic features in smaller corpora to train corpus driven neural models that enable robust large scale analysis. We show that BiLSTM-CRF models with syl- lable embeddings outperform a CRF baseline and different BERT-based approaches. In a multi-task setup, particular beneficial task re- lations illustrate the inter-dependence of po- etic features. A model learns foot boundaries better when jointly predicting syllable stress, aesthetic emotions and verse measures benefit from each other, and we find that caesuras are quite dependent on syntax and also integral to shaping the overall measure of the line. 1 Introduction Metrical verse, lyric as well as epic, was already common in preliterate cultures (Beissinger, 2012), and to this day the majority of poetry across the world is drafted in verse (Fabb and Halle, 2008). In order to reconstruct such oral traditions, literary scholars mainly study textual resources (rather than audio). The rhythmical analysis of poetic verse is still widely carried out by example- and theory-driven manual annotation of experts, through so-called close reading (Carper and At- tridge, 2020; Kiparsky, 2020; Attridge, 2014; Men- ninghaus et al., 2017). Fortunately, well-defined constraints and the regularity of metrically bound language aid the prosodic interpretation of poetry. However, for projects that work with larger text corpora, close reading and extensive man- ual annotation are neither practical nor afford- able. While the speech processing community explores end-to-end methods to detect and con- trol the overall personal and emotional aspects of speech, including fine-grained features like pitch, tone, speech rate, cadence, and accent (Valle et al., 2020), applied linguists and digital humanists still rely on rule-based tools (Plech´aˇc, 2020; Anttila and Heuser, 2016; Kraxenberger and Menning- haus, 2016), some with limited generality (Navarro- Colorado, 2018; Navarro et al., 2016), or without proper evaluation (Bobenhausen, 2011). Other ap- proaches to computational prosody make use of lexical resources with stress annotation, such as the CMU dictionary (Hopkins and Kiela, 2017; Ghazvininejad et al., 2016), are based on words in prose rather than syllables in poetry (Talman et al., 2019; Nenkova et al., 2007), are in need of an aligned audio signal (Rosenberg, 2010; …

Content truncated to 500 words