Nineteenth-century attempts to compress the complex relationship between the surviving versions of popular medieval romance traditions into the genealogical form of the stemma cannot but raise doubts among those medievalists whose interests tend toward the specificity of the reception context offered by each individual manuscript version. How can such complexity be captured in a tree, without over-privileging some features over others in order to arrive at a coherent picture? And, indeed, how can philologists establish the relationship between the variant versions of a text that have survived without ignoring the fact that the larger part of most traditions was probably contained in manuscripts that are now lost? The answer to these questions may lie in the field of digital humanities, where digital analysis offers us the opportunity of comparing textual relationships based on hundreds of textual features, rather than the small handful of key features often used in the past to distinguish different strands of a text. At the same time, digital visualisation techniques make it possible to view and understand the relationship between the manuscript versions that arises out of such analyses without resorting to tree and lists.
Using the romance Floire et Blanchefleur as a pilot study, I have, with the help of Antonia Murath of the Free University Berlin, created a database that identifies and weights 170 differentiating plot features from twenty versions of the Floire et and Blanchefleur narrative from across Europe. Christian Edlich-Muth and I have subjected this data to principal component analysis and digital clustering and visualisation techniques in order to identify which features of the different versions correlate with each other and how these correlations relate to what is known about when and where the different manuscripts they are contained in were produced. These visualisations make it possible to better differentiate between various subgroups of the Floire et Blanchefleur story and recognise how the tale came to be dominated by different generic features (devotional, romance, epic) in different regions. The results of this project are currently being prepared for publication.
Miriam Edlich-Muth 