An attempt to storify the Twitter feed from the DH Topic Modeling Workshop (at least the second half of it)
Back in August, Miriam Posner’s post “What are some challenges to doing DH in the library?” initiated a wide-ranging conversation in the blogosphere examining the relationship between DH and libraries. As the dh+lib blog gets a’rolling, it seems useful both to revisit Miriam’s post, but also remind ourselves of the potential DH holds to enable new modes of discovery, knowledge, and interpretation, both for those in the academy and those in the broader field of cultural heritage.
Miriam cited a host of challenges to doing DH in libraries, including insufficient training, lack of authority, organizational stasis, overcautiousness, and lack of professional incentive. Along with a wealth of comments, the post also elicited responses by Michael Furlough, The Library Loon, and by Trevor Muñoz. As the Loon noted, many of the challenges described in these posts and the comments are issues applicable to any creative or forward-looking initiative within a library. Many readers here can no doubt attest to that truth.
The post that resonated most with me, however, was Trevor Muñoz’s assertion that DH is not just something librarians support through instruction or as liaisons or project managers, but is something that librarians themselves should be actively undertaking. As Trevor acknowledges, he has a unique occupational role straddling both a traditional academic library and a DH center; and as DH becomes more prevalent within academia, I imagine we will see more of his type of dual-appointment roles.
At the same time, I think Trevor’s point can be extrapolated even further to encompass how DH tools, methods, and technologies have the potential to help enhance and evolve a wealth of professional practices beyond academia and across all of cultural heritage. It is this ability to reinvigorate the work of non-academics, such as librarians, archivists, and collection managers, that has many of us in cultural heritage excited about DH as an emerging idiom within memory institutions. As the work of DH centers like CHNM and MITH gain more exposure outside the confines of academia, the broader cultural heritage community will better understand how DH can, as the recent book Digital_Humanities asserts, “open up important new spaces for exploring humanity’s cultural heritage and for imagining future possibilities using the transmeta methods and genres of the digital present.”
But that’s the gauzy version — some examples please! The attendance of a number of librarians, archivists, and other cultural heritage workers at the recent Topic Modeling workshop at MITH (#dhtopic for the twitterati; GDoc of notes) was a good example of this community’s interest in the promise of topic modeling as a tool to enhance discovery and rethink many fundamental practices around collection management and accessibility. As one of the non-academics in attendance, I was there to better understand how the enviable work of the talented DHers in the crowd can be adapted or transplanted (or, hell, brick-through-window’ed) into everyday practices in libraries, archives, and museums.
What galvanizes many of us working in cultural heritage is how DH tools and practices will enable us to move beyond the traditional methodologies of description of, and access to, archival or cultural collections. These traditional practices, holdovers from a world of physical materials and all the attendant requirements of arrangement, bulk, and storage, have also been fundamentally subjective. Catalogs, finding aids, LCSH — all are products of interpretive biases. That inherent subjectivity engendered a minor, if ongoing, crisis of conscience once contemporary criticism called into question the façade of objectivity in the management of cultural and historical materials (see, for instance, in archival studies, the work of Terry Cook, Heather MacNeil, and David Bearman). But tools like topic modeling, text mining, data visualization, and other methods of distant reading have the power to obviate (or at least largely reduce) the interpretive imposition of the cultural heritage professional at the point of access. They will allow collection stewards to refocus their efforts on providing the tools necessary for users to interpret and understand materials instead of focusing on the descriptions and classifications that group or arrange them.
Because of this, when discussing topic modeling and its promise for cultural heritage at the workshop, I was less fretful of the fallibility of the algorithmic presumptions of Latent Dirichlet allocation and more interested in what I’ll call DH’s confrontational potential — something similar to Mark Sample’s declaration of “an insurgent humanities.” Trevor Owens, in his cogent post about the tweets from #dhtopic, captured something of this seeming divide, noting the assumed disjunction between the exploratory and the evidentiary and how DH dialogues often run on two parallel tracks, one focusing on the freewheeling use of DH tools for discovery and the other on the use of DH tools to validate a specific argument. Here, the “generative discovery” possible with computational tools (akin to Stephen Ramsay’s “The Hermeneutics of Screwing Around“) is merely prologue to the overall process of building an evidence-based defense “against alternative explanations.”
But for librarians, archivists, and collection managers, there is no need to take that next step — enabling alternative explanations is entirely the goal of supporting accessibility and building or providing discovery tools. Argumentation or justification, while essential to scholarly knowledge production, is counter to the goals of the collection manager in describing and making available records. Our traditional methods of making materials available (taxonomic, ontological — the finding aid, the OPAC) are complicated by all sorts of logical or illogical interpretations and subjectivity. Tools like topic modeling offer the ability to bypass that interpolation, that annotative interruption, and hand users and researchers the tools to construct their own topics, queries, pathways, and meanings.
While subjectivity may be inescapable when it comes to archival appraisal and collection acquisition policies, librarians and archivists strive for anonymity and objectivity when creating systems of discovery. We strive for something beyond the influence of the idiographic. But even more than that it is the ability to enable simultaneous and, especially, contradictory means of discovery and interpretation. The more context we can provide the better; the more contradiction and interpretation we can enable, the stronger are our discovery tools. This is the opposite of argumentation. DH tools and methods offer, then, if not an antidote to the immutability of traditional descriptive and discovery methods, then at least a confrontational alternative and a substantive corrective.
That one collection of resources can give different users different outcomes and support contradictory arguments may sound too post-structuralist or relativist for some. But I think it signals a healthy reassessment for cultural heritage institutions who for too long have placed the collection manager — the librarian, the archivist, the collection manager — at the gate of discovery and access. The digital humanities have the potential to change the nature of that equation, to upset established methods of description and access, and to reaffirm the role of the cultural heritage professional as essential to preservation, accessibility, and usability of information and the cultural record. Here’s to making this blog and the dh+lib discussion group a place to explore how that affirmation is happening.
