The Index Thomisticus Treebank as Linked Data in the LiLa Knowledge Base

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 4022–4029 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 4022 The Index Thomisticus Treebank as Linked Data in the LiLa Knowledge Base Francesco Mambrini, Marco Passarotti, Giovanni Moretti, Matteo Pellegrini Universit`a Cattolica del Sacro Cuore Milan, Italy {francesco.mambrini,marco.passarotti,giovanni.moretti,matteo.pellegrini}@unicatt.it Abstract Although the Universal Dependencies initiative today allows for cross-linguistically consistent annotation of morphology and syntax in treebanks for several languages, syntactically annotated corpora are not yet interoperable with many lexical resources that describe properties of the words that occur therein. In order to cope with such limitation, we propose to adopt the principles of the Linguistic Linked Open Data community, to describe and publish dependency treebanks as LLOD. In particular, this paper illustrates the approach pursued in the LiLa Knowledge Base, which enables interoperability between corpora and lexical resources for Latin, to publish as Linguistic Linked Open Data the annotation layers of two versions of a Medieval Latin treebank (the Index Thomisticus Treebank). Keywords: Treebank, Linguistic Linked Open Data, Latin 1. Introduction and Motivation Linguistic annotation on textual corpora is an in- valuable support for studying historical languages like Latin. Language learning and corpus-based re- search are the two most obvious applications. Since Latin has a very rich morphology, lemmatization and morphological-feature annotation are particularly im- portant for tasks like word search or for vocabulary acquisition. The distinctively “free(er)” word order of Latin, as compared to many modern languages like Ital- ian or English, also greatly complicates the syntactic analysis of texts for modern readers (or even for parsers fine-tuned for modern languages). While services and libraries that support lemmatiza- tion and/or morphological analysis for Latin have been available for decades,1 a series of Latin treebanks, with word-by-word account of the syntax and morphology of Latin texts, have been published only in recent years. Latin’s long and rich history is well reflected also in the spectrum of existing treebanks, which include texts of different genres and periods. The treebank devel- oped within the Perseus project (Bamman, D. et al., 2017), consists of a small (about 53,000 tokens) selec- tion of texts of the Classical period, from 1st century BC to 1st century AD (Bamman and Crane, 2011); the Latin portion of the PROIEL treebanks (Haug, D. et al., 2018) contains the Vulgata by Jerome (4th century), plus some Classical and Late Latin texts, for a total of about 200,000 tokens (Haug and Jøhndal, 2008). The Late Latin Charter Treebank (LLCT) (Korkiakangas, T., 2020) is the only one featuring non-literary texts, as it is entirely composed of charters written in Tuscany 1See for instance: LEMLAT (Passarotti, M. et al., 2020), and the modules for Latin in UDPipe (Straka, M., and Strakov´a, J., 2021) and CLTK (Johnson, K.P. et al., 2021). See Passarotti et al. (2020) for a more detailed list. in the 8th and 9th century, for a total of 242,000 to- kens (Cecchini et al., 2020b). Moreover, two treebanks display texts from a …

Content truncated to 500 words