OGA v0.2.0 (2024-11-24) and GLAUx v1.0 (2024-04-09) Compared

Overview This deposit includes three csv files comparing version 0.2.0 of the Opera Graeca Adnotata (hereafter, OGA) corpus with version 1.0 of the GLAUx corpus. The files are intended as finding aids for scholars and the general public to identify where to look for the latest Greek digital editions with rich morpho-syntactical annotation. All fields in the csv files are wrapped in double-quotation marks and are pre-sorted by year-date in ISO 8061 format. The three files represent: OGA and GLAUx matches OGA unique texts GLAUx unique texts Method The matching and deduplication process used the CTS_URN (Canonical Text Service Uniform Resource Name) in the metadata of each corpus as a key. While generally reliable, occasionally some duplications or overlaps persist because the respective corpora use different base texts, collections, or partitions. For example, GLAUx annotates Aesop's Fables with two CTS_URN values, "tlg0096.tlg002a" and "tlg0096.tlg002b", corresponding to two different collections of the fables. OGA has only a single entry for Aesop's Fables, "tlg0096.tlg002", corresponding to a single collection. Searching all three files simultaneously is advisable to determine if a text is unique to one corpus or if it overlaps to some degree between corpora. Author and title values were extracted from the metadata of both corpora, but they default to OGA values for matched texts. Composition date ranges were also extracted from the metadata of both corpora, with adjustments made to all dates to conform them to the ISO 8061 date standard. Sources and Bibliography Version 0.2.0 of the Opera Graeca Adnotata, curated by Giuseppe Celano at the University of Leipzig, was archived on November 24, 2024 on Zenodo at https://doi.org/10.5281/zenodo.14206061, with 40M+ tokens and 1999 base texts/files, with annotation of various features (e.g., part of speech tag, XPOS tag, etc.). These feature values are mostly consolidated within CoNNLu files, yet separated out by distinct layers corresponding to several document-specific XML files. ISO 8061 dates and author designations (sometimes provisionally made) for all texts in the corpus make it particularly useful for stylochronometry, i.e., diachronic linguistic studies or the study of language evolution. The OGA corpus was described in this earlier article that corresponds to v0.1.0. Giuseppe G. A. Celano, "Opera Graeca Adnotata: Building a 34M+ Token Multilayer Corpus for Ancient Greek", arXiv cs.CL, 2024-03-41, https://doi.org/10.48550/arXiv.2404.00739. Version 1.0 of the GLAUx annotated corpus of classical and postclassical Greek, curated by Alek Keersmaekers at the Catholic University of Leuven, was archived on April 9, 2024, at https://doi.org/10.5281/zenodo.10948375. It contained 20M+ tokens and some 1400 texts. The corpus is regularly updated on Github at https://github.com/alekkeersmaekers/glaux. The dataset is described in the following scholarly publication: Keersmaekers, Alek (2021): The GLAUx corpus: methodological issues in designing a long-term, diverse, multi-layered corpus of Ancient Greek. Proceedings of the 2nd International Workshop on Computational Approaches to Historical Language Change 2021, 39–50. Online: Association for Computational Linguistics. https://doi.org/10.18653/v1/2021.lchange-1.6. Rights Statement These csv files are (c) 2024 by Mark G. Bilby and distributed under a CC BY-SA 4.0 license, honoring the license of the entire OGA corpus and of many of the texts in the GLAUx corpus.