Coding editions. Computational approaches to the editing of pre-modern texts.

Introduction

The endlessness of the digital space allows scholarly editors to conceive edition projects on a grand scale. Indeed, the eternal mission of philology can be identified in the quest for the best way to represent and/or reconstruct primary sources, in order to save them from oblivion and grant the public the most informed access to them. Consequently, the lack of space boundaries characterising the digital environment is one of the first aspects making such an environment perfectly suitable for hosting ambitious edition projects. Considering then the possibilities disclosed by hypertextuality, multimodality and multimediality, one can hardly argue against the digital way of editing, despite the ongoing challenges it must face (cf. for instance Rosselli Del Turco 2016), of which sustainability is clearly in the forefront. However, there are reasons to believe that state-of-the-art digital editions have not yet overcome the limits of a still rather bookish paradigm, obeying a mostly representational logic (Van Zundert 2018, Cugliana and Van Zundert 2022).

In this contribution, I will argue that the key to the next level of digital scholarly editing is to be found in computation, that is, in the actual coding of the whole editorial workflow. I will present the theory, the advantages and the challenges of such a computational approach to the editing of pre-modern texts, backing up my claims with examples from the praxis and from my own scholarly work in the field of digital philology. Of course, the field of computational editing is still in its infancy, which means that fully computational editions have not been published yet. However, there are projects¹  which can already prove the validity of the arguments presented in this paper or that at least seem to move in the same direction.

Theoretical background

While the digital paradigm has been widely described and commented upon (among others by Stella 2007 and Sahle 2016), there is a crucial aspect that has not been sufficiently underlined, which however could represent a turning point in the history of (digital) philology. It is the case of the use of programming code throughout the different phases of the editorial process, aiming at the realisation of what Barabucci and Fischer (2017) defined as the formalisation of textual criticism. The authors, in their conclusions, state that a “shared formalization would lead to the semi-automatization of the editorial process”, where “the responsibility of the editors would be to describe their choices and decisions”, while that of the computers would be to “deal with applying these rules and decisions in the best way”.

The act of formalising the competence of the editor is to be seen as an achievement per se, for it can lead to more accountability and verifiability of the scholarly processes, which can be shared in its complexity with the community. As a matter of fact, such processes are of fundamental importance not only from a methodological perspective, but also for the proper understanding and contextualisation of the presented results. In this respect, Van Zundert (2018) points out how the current digital methods used in scholarly editing, instead, still aim at merely recording the results of the philological work in a representational fashion,²  without keeping track of the decisions, analyses and actions that led to those results. In a computational edition, on the other hand, the scholarly competence and actions are not only expressed in a formal language, but they are also operationalised. This, in turn, brings about new affordances, such as the replicability and reusability of the editorial work, as well as the introduction of a degree of indirection allowing for highly controlled modelling of the edition process.

If methods are there to realise theoretical principles in the best possible way, at the same time they can in turn influence the principles themselves, opening up new perspectives uncovering, for instance, implicit biases and illogicalities. Indeed, the use of computation for the modelling and operationalisation of editorial knowledge can contribute to perfecting the editorial workflow in the way envisioned by McCarty (2005), who referred to the “meaningful surprise” often arising from the process of modelling and computing.³ 

The praxis of the computational method

Some examples from the actual application of this approach will hopefully prove its potential for the field of Digital Scholarly Editing.

Conclusions

As can be evinced from this abstract, my contribution has a twofold purpose: on the one hand giving a short, but hopefully convincing introduction to the theoretical aspects justifying the scholarly value of a computational approach to editing and, on the other, presenting some meaningful results obtained in the praxis. In particular, the focus will mainly be on the success and challenges of the normalisation pipelines developed in collaboration with Gioele Barabucci, showing what it actually meant to translate into code the usually very analogue task of normalising medieval texts, what problems we encountered and how we solved them. Finally, as an outlook for the possible applications of the computational principle, I will briefly illustrate the work I am doing for my next project, the edition of the Prenostica Socratis Basilei, which was conceived from the start as a computational one. In my talk, I will show some prototypical examples concerning specific aspects of the computational workflow.

¹ An example of the application of an integrated computational approach for different phases of the editorial process is the digital critical edition of the Chronicle of Matthew of Edessa (Andrews, Safaryan and Atayan 2019).

² Although there seems to be a need for further discussions also as far as the “edition as interface” is concerned. In this respect, Andrews and Van Zundert (2018) argue that the interface is “not just an argument about the text, but also an argument about the ‘attitude’ of the editor, a window into his or her take on methodology and the digital edition itself. It is also a revelation of the technical skills available to the editor. The interface tells us something not only about the methodology but also about the import of the edition”. The authors observe, however, that many editors still underestimate the role of the interface, whose “language” is one of great complexity.

³ “Take, for example, knowledge one might have of a particular thematic concentration in a deeply familiar work of literature. In modelling one begins by privileging this knowledge, however wrong it might later turn out to be, then building a computational representation of it, e.g. by specifying a structured vocabulary of word-forms in a text-analysis tool. In the initial stages of use, this model would be almost certain to reveal trivial errors of omission and commission. Gradually, however, through perfective iteration trivial error is replaced by meaningful surprise” (McCarty 2005).

⁴ The pipelines are available online on the pre-release website of the DI edition: https://marcopolo.cceh.uni-koeln.de.

⁵ In February 2022 it received an “Initial Funding” from the Faculty of Arts and Humanities at the University of Cologne (https://phil-fak.uni-koeln.de/en/research/research-funding/initial-funding).

⁶ The roots of the sortes have been traced down to the Mediterranean area (Abraham 1968), but their use soon spread all over Europe. Since the Renaissance these texts have been used as entertaining games to play in a group (Urbini 2006), but this was not the case in antiquity and for the greater part of the Middle Ages. Then, they were taken seriously and could have a strong influence on people’s decisions.