QUANTITATIVE CODICOLOGY AND THE REPERTORIUM WORKSTATION

The Repertorium of Old Bulgarian Literature and Letters (henceforth “the Repertorium”) is a
series of projects begun in 1994 by Anisava Miltenova (Institute of Literature, Bulgarian Academy of Sciences), David J. Birnbaum (Department of Slavic Languages and Literatures, University of Pittsburgh), Andrej Bojadzhiev (University of Sofia), and Milena Dobreva (Institute of Mathematics and Informatics, Bulgarian Academy of Sciences) and designed to explore the opportunities for exploiting computational resources to advance the study of Slavic philology in original and innovative ways. From
its inception the Repertorium set out to transcend the impressive but ultimately limited achievements of
many other projects in humanities computing, such as structured document development, structured searching,
variable rendering for multiple uses, etc. While the
Repertorium has appreciated these technologies and has used them productively, its long-term goal has been not merely to enhance traditional philological activities (such as publishing and searching), but also to facilitate new
types of philological research that would have been not merely impractical, but essentially unthinkable without the use of computational tools.
Over the past ten years the Repertorium has overseen the encoding of more than 350 manuscript descriptions by members of the Institute of Literature, and it has also coordinated and directed compatible large-scale joint projects with British and Swedish researchers. The
Repertorium DTD, begun before both the TEI P5
manuscript description module and the European
MASTER initiative, is designed as a conformant
modification of the TEI DTDs, but it imposes tighter and more consistent structure than any comparable project of which we are aware, and this attention to structure (whereby a manuscript description is as much a database as a document) is designed specifically to facilitate not
merely rendering or structured searching, but also
data-mining for analytical purposes, an enterprise for which I have coined the term “quantitative codicology.”
In a keynote address at one of the recent Extreme Markup conferences, Tommie Usdin (Mulberry Technologies) observed that “XML has made true all of the lies that we told about SGML.” When the Repertorium team began encoding manuscript descriptions in SGML the mid-1990s, we took on faith that at some point it would be possible to perform the sort of analysis we needed for our data, but it is only with the introduction of such relatively recent ancillary XML technologies as XSLT, XQuery, and SVG that it has been possible to exploit
the Repertorium inventory for the innovative and
sophisticated analysis that is was originally intended to facilitate.
My presentation at the 2006 Digital Humanities
conference will concentrate on describing both the
technologies developed within the framework of the
Repertorium initiative for analyzing medieval Slavic manuscript materials and the philological principles that guided our research. In particular, this presentation will demonstrate the ability for a scholar seated behind a
standards-conformant web browser anywhere on the
Internet to combine on the fly and in varied ways
structured searching with the generation of such graphic representations of the relationships among manuscripts as dendrograms and plectograms (about which see my “Computer-Assisted Analysis and Study of the Structure of Mixed-Content Miscellanies,” Scripta & e-Scripta,. vol. 1 (2003), 15–54 [preprint available on line at http://clover.slavic.pitt.edu/~djb/2003_ljubljana/2003_ljubljana_paper.pdf]).
What is most new and significant since my initial
introduction of these technologies into manuscript
studies is my integration of them and of other tools into a “Repertorium Workstation,” an integrated and coherent platform where users can browse and search manuscript descriptions, generate plectograms on the fly (through behind-the-scenes XQuery and XSLT transformations to SVG), use the plectograms to launch new queries, and,
in general, explore the range of relationships among
manuscript witnesses without constraint in ways
that have never been exploited before in humanities computing, and that were essentially unthinkable due
to technological limitations until very recently. The
presentation will combine a demonstration of the
Repertorium Workstation with a discussion of its
philological motivation, its design principles, the types of original and innovative primary scholarship that it supports, and plans for further development.
All Repertorium materials, from the raw data files to
the Internet-based Workstation, are (or will be) freely accessible on the Repertorium web site (http://clover.
slavic.pitt.edu/~repertorium/). While many of the
technologies underlying the Repertorium Workstation
were developed to address specific needs of Slavic
medievalists (e.g., Unicode-based support for multiple and uncommon writing systems), the system has great general applicability to medieval studies in other cultural tradition, as well as to non-medieval studies of textual
traditions. In particular, its innovative use of SVG to model the structural similarities among manuscript
witnesses can serve as an example (easily adapted
to other projects) of how graphic representations can provide insights into textual information that would
otherwise remain virtually imperceptible.