The project is partially finished, as corrections will be made in the following months mainly in the dissertation after deliberation with the defense jury. A paper publication concerning this MA work is expected in 2025 in the series Studia Artistarum of Brepols, as the publication of the proceedings of the journes d'tude in the memory of Louis Holtz that were held in Paris in June 2023 https://chartes.hypotheses.org/9407

Eutyches Grammaticus, but glossed

Navigating the repository

Structure

| Folder Name | Description | | ----------- | ---- | | .github/workflows | Contains the implementation of chocomufin and htrvx yml files for our dataset| | BambergMsc30 | Contains all the files data-GT-notebooks-graphs pertaining to the Bamberg Msc 30 witness | | Lat14087 | Contains all the files data-GT-notebooks-graphs pertaining to the Lat14087 witness | | Lat7499 | Contains all the files data-GT-notebooks-graphs pertaining to the Lat7499 witness | | VLO41 | Contains all the files data-GT-notebooks-graphs pertaining to the VLO41 witness | | comparativemssanalysis | contains the graphs and notebook for the comparative analysis of the witnesses| | kraken-YALTAi | contains the results and the test set for the YALTAi and kraken models. The models themselves can be dowloaded here: https://drive.google.com/drive/folders/16K4tPTaZvcc-rYqEFlUAoHX9vtaYLfnF?usp=sharelink | | pipelinediagrams | General diagrams concerning the pipeline and methodology | | rendu | The pdf and .tex file of the Master2 dissertation |

The rest of the repository contains the table.csv for the chocomufin character control.

Introduction

Eutychs was a Latin Grammarian, active in the mid-6th century AD Constantinople, a disciple of Priscianus Caesariensis. His contribution to the field of grammar consists mainly of a treatise called De uerbo, addressed to his diligentissimum discipulorum Craterus.The ars elaborates on the criteria for the classification of the conjugation of verbs, in short, if the genitive of dico is dicas or dicis. Itself a rather arid development of principles that do not escape the general rule, nevertheless thanks to De uerbo 141 citations of classical authors have been passed down to us -some passages being otherwise lost-, mostly early Augustan poets. Adding to this, exhaustive lists of examples of verbs and derived nouns make an integral part of the treatise, that have been recycled from posterior Grammatici.

De uerbo was last edited at the end of the 19th century by Heinrich Keil in his fifth (out of seven) tome of the Grammatici Latini, alongside an enormous corpus of artes dating from the 4th to the 7th century. This monumental edition lacks methodological rigor and exhaustive paleographical research, is, for the most part, outdated, leaving a substantial amount of grammatical works in need of critical editions, including Eutychs' own contribution.

The manuscript tradition of the work, as reported by the excellent work of Colette Jeudy, is rather modest, as it is comprised of 32 manuscripts transmitting partially on entirely the text, and span from the end of the 8th to the 11th century AD. De uerbo enjoyed a fulgurant posterity, that inspired two medieval scholars, namely Sedulius Scottus and Remigius of Auxerre to make a commentary rendition out of it, attesting to its importance in the educational milieu.

The typology of the manuscripts comes with two interesting features that pose at the same time the biggest difficulty for perspective editors (and, as we suppose, for Keil). Firstly, as many other works on grammar, actively used in the "medieval classroom" for the teaching of Latin, lists of examples are incorporated in the narrative between bits of theory, forming columns or/and tabular boards, a mise-en-page apt for the memorization of these exceptions. At first, this feature may not seem as big of a deal, but a thorough examination of these tables and of the posterior works that use them indicate that it yields important information for the understanding of the manuscript tradition (see especially Conduch 2019 "La mise en page d'Eutychs"). Secondly, several witnesses are richly annotated with multiple layers of interlinear and marginal notes, the progressive accumulation of which finds its climax in Remigius of Auxerre's commentary. Gloses in grammatical manuscripts have only recently (better late than ever!) started to interest editors (see Monella/Rosselini for Priscian and Evina Steinov for Isidore of Seville's Etymologies)most of them making use of digital tools. And that is because a typology as complex as that of grammatical glossed manuscripts cannot be easily handled and, better, be utilized for research purposes without the flexibility, absence of spatial constraints and ability of to handle of big and multilayered data that digital tools offer.

This project adheres to this renewal of tools and methods for the edition of grammatical texts in order to make full use of the precious information present in the manuscript tradition. Our case study for the first year of the Master's program in Digital Humanities in the marvelous codex Vossianus Latinus O. 41, a real-life specimen of a grammar manual used in class. Written by 4 hands and annotated by another 4 (and some more for corrections), and containing ordered -or not- lists of examples that find their way into doctrinal discussion it constitutes an exemplar case study for a diachronic survey of the different "annotation campaigns" as well as an attempt to model its heterogenous layout structure. A telling example from folio 5r:

Pipeline Overview

The following graph concisely describes the pipeline that was developed and implemented for the M1 dissertation. Limitations and improvements (in French) are discussed in the final chapter of the mmoire.

Quality control

We decided, following HTR-United workflows, to proceed with the implementation of chocomufin for the control of the special characters and HTRVX for the proper implementation of SegmOnto. Chocomufin is a software workflow developed by Clrice Thubault and Pinche Ariane that creates a table of the characters (table.csv) that exist in the ground truth for both documents and checks, with every push and pull request that the .xml documents found in the folders are conforming to this particular table. HTRVX, developed -surprise!- also by Thibault Clrice, is an excellent tool that reports back on SegmOnto non-compliant ALTO files, containing blank/non-attributed lines/unbound lines in zones/non-detectable extra empty zones created on eScriptorium and particularly helped with some gazillion lines that we had to manually annotate :)