Collation Visualization: Helping Users to Explore Collated Manuscripts

Elisa Nury (King's College London)

Digital Classicist London seminar 2017

Friday July 7th at 16:30, in room 234, Senate House, Malet Street, London WC1E 7HU

Livecast at Digital Classicist London YouTube channel.

The comparison of manuscripts and recording of variant readings is an essential, but challenging task towards the preparation of a scholarly edition. A large amount of detailed data is collected during collation, and needs to be analysed and visualized. In the recent years, the use of digital format has been increasingly incorporated within the collation workflow: from writing down variants in Word or Excel documents to adopting complex automated collation tools such as CollateX. It has thus become necessary to display collation results on the computer in a meaningful way, which will help scholars to explore the collated manuscripts.

Collation is admittedly more than just a record of variant readings. It frequently incorporates additional notes and comments, “paratextual” elements such as changes of pages or folia, gaps, lacunae, and so on. This combination of variants and paratextual material produce a large amount of complex collation data, which are difficult to read and to interpret. Therefore, the scholar needs to visualize and analyse the collation results as a whole, and not only variant by variant. A good visualization should offer a way to check collation against the actual witnesses, whether they are manuscripts or printed editions. In addition, the user should be able to interact with the collation to analyse readings and variants. Collation could be filtered, so as to find patterns of agreements or disagreements between those witnesses, which can indicate how they are related to each other. Visualization and manipulation of collation results are thus essential in order to use collation for further research, such as studying the manuscript tradition and creating a stemma codicum.

To tackle these issues, I would like to present a method of visualization of collation results. The method of visualization consists of two aspects: first, a description of the collation table displayed in HTML; second, a Jupyter notebook where the user can interact with the collation through a Python script, to select agreements between a group of witnesses against another group, or to make small corrections in the alignment. The case study to which the visualization was applied is the Declamations of Calpurnius Flaccus. It is a classical literary text in Latin from the second century. The witnesses, four manuscripts and two critical editions, have been encoded in XML TEI and then pre-tokenized for collation with CollateX into a JSON format that allows to record a reading together with more detailed information. The Jupyter notebook will then serve to analyze the collation results: how are the witnesses of Calpurnius Flaccus related to each other?