Preserving Digital Archives

Most attendees at the Beinecke Library’s recent conference on digital archiving–“Beyond the Text: Literary Archives in the 21st Century“–arrived equipped with the idea that there is no preservation without loss.

What may have given some attendees pause, particularly those who work primarily on the first two centuries following the Reformation, is how much 21st-century digital stuff is being preserved–and how idiosyncratic the process of selection can be.

Faced with the data deluge of a contemporary literary figure’s electronic correspondence, for example, how do archivists determine what gets archived and what gets tossed?  Now that archiving can begin during a writer’s or publisher’s lifetime, without a family member’s interference (think Cassandra Austen), who shapes the archive?  And if digital archivists shape the archive, what principles of retention do they use?  Where do their loyalties lie? With the author?  Or with the data-hungry and feverishly scandal-mongering scholars of posterity?

The two-day conference raised unresolved and provocative questions, many of which focused on the problem of selection.  Fran Baker, the Assistant Archivist for John Rylands Library at the University of Manchester, discussed the complexity of archiving the Carcanet editorial papers, including email.  Hearing about the decision-making process determining what stays and what gets tossed may not seem new to librarians familiar with the problem of sorting and discarding, but in the context of shaping an archive, that decision-making process and its likelihood of error takes on urgency.

There were stories of forensic success, the most notable of which is Matthew Kirschenbaum’s narrative of the extensive and collective effort tracking down William Gibson’s electronic poem, “Agrippa,” which was designed to encrypt itself after a single reading.  That a text programmed to go away can be recovered suggests both the value of collaborating on large digital projects like The Agrippa Files and the perils of assuming that an author has control over her or his electronic archives.  Similarly, Beth Luey’s account of the rich storehouse of data contained in publishers’ records–sales data, copies printed, copies sold, print runs, design decisions, contracts, marketing files, legal disputes, reviews, book jacket design, subsidiary rights, and so forth–both encouraged work on publishers’ records and raised ethical and legal issues.  In the discussion that followed, for example, it became clear that though some publishers did not retain rejected manuscripts, others did, including pertinent correspondence and readers’ reports.

The Keynote talk by David Sutton noted that literary manuscripts are like no other manuscripts in that they offer insights into the act of creation.  He showcased ongoing projects that promote an awareness of digital literary archives:

• GLAM Group for Literary Archives & Manuscripts
• Diasporic Literary Archives Network

Hazel Carby’s eloquent, harrowing, and culturally resonant account of tracing her family genealogy back to a slave owner’s carefully archived records, reminded everyone that archives preserve both the beautiful and the monstrous.

Diane Ducharme drew on her experience at the Beinecke to warn that however much we may desire an unmediated past and a pristine archival order free from editing and explicating, all archives arrive shaped and selected.  Her discussion underscored the importance of searching for the traces of a previous archivist’s work.

Micki McGee described her experience with the Yaddo Archive Project, which aims at providing visualizations of the social network of writers who worked at Yaddo.  She described the process of seeking a relational database with social network mapping and a visualization widget.  Though the project, Yaddo Circles, requires authentication and is not yet available for public view, this vimeo provides an overview.  Clicking here reveals the kind of relational visualization this project might produce.

McGee also recommended looking at the following projects:

• Crowded Page
• Phybo
• Orlando Project
• SNAC Project
• The RoSE (Research Oriented Social Environment)
• Linked Jazz

These projects have potential for helping us recover the intensely sociable and highly competitive literary worlds of the long eighteenth century.   Like the many other provocative and interesting papers and introductions to sessions, they point a way forward even as they raise methodological, logistical, and even ethical questions.

This conference made clear the value of a longer conference, with sessions focusing on specific problems posed by digital archives of material both old and new.  I welcome contributions by others who attended the conference to help complete this cursory overview.