Buried treasure, castles and roads

Ironically, the Day of DH was accompanied by disappointing news from the NEH for many of us who had applied for digital humanities grants (sincere and non-ironic congratulations, however, to those projects that were selected). Posting for the former and processing the latter set some wheels in motion in my mind. Those wheels began to spin more quickly as I read tweets and posts from the Digital Classics Association conference, but it’s only in the last couple of days that they’ve gained enough traction for me to attempt to express them in writing. I hope, therefore, that the archivist of this Day of DH will excuse the post-day addition to the blog.

My thoughts began to take shape when I realized that my three unsuccessful grant proposals all had a similar characteristic: they involved the development of digital resources to allow semi-digitally-literate people like me to share data and make connections. One proposed the creation of a desktop toolkit to allow humanities researchers to automatically create metadata for their own complex digital collections, without having to delve deeply into the world of schemata and ontologies and without being dependent on a particular repository. The other two involved a project to create a gazetteer of published assertions about archaeological and historical periods that could serve as a hub for the linking and reconciliation of ambiguous references to periods in texts and datasets. We wanted to get away from a standardized vocabulary that would either be too general or too specific for anyone to want to use. Instead, we proposed to collect a potentially infinite number of authoritative claims about the temporal and geographic coordinates of period terms, which could in turn be used both to disambiguate less specific mentions of periods in texts. This gazetteer would also serve as a reference point through which datasets could be connected in a linked-data environment, just as Pleiades does for Pelagios.

Both projects were meant to produce tools that would allow humanists, in the broadest sense, to take control over – and responsibility for – the sharing or preservation of the digital data we are now all producing. They were meant to lower the bar for entry into the “digital humanities”, and make the majority of us who lack either resources or programming skills less dependent on the few who have one or the other. Both drew widely divergent reviews from the grant review panels. Some of the more critical reviews offered some input of real substance, for which my collaborators and I are grateful. But at least one set of comments for each grant denied that the issues we wanted to address were real problems. Digital humanists have digital repositories for archiving, they said, or they manage their own material with tools like Zotero; archaeologists work out the complex questions of periods on their own, and don’t need somewhere to look them up.

Apart from the sort of defensiveness one always feels about one’s own projects, I realized that I found these comments frustrating in a complementary way. The first set, about archiving and repositories, assumes that the best practice is standardization and centralization, and that people can be forced or cajoled into depositing their data in a few standard formats with a few major digital repositories. This may be true of data produced in the course of government-funded research, which now must be deposited in such repositories, or of data produced in environments like the UK, where cultural-heritage laws make the deposit of archaeological data mandatory. But it is patently not true of data collected by a vast sea of small projects and individual researchers, much of which will never even be described properly, let alone submitted to a central repository. The second set of comments, on the other hand, assumed that the job of digital tools was to promote the collection of new data and the generation of new interpretations, and since archaeologists can deal perfectly well with periods on their own in this process, a digital gazetteer will not lead to any new research. Here the emphasis was on what will produce new data or let us extract new information from existing data right now.

Although these were not the only voices in our reviews, I do think they represent the dominant view of the digital humanities. I also think that they reflect, to some degree, the reasoning behind decisions to fund many current projects. Grantors seem to respond to projects that will either facilitate brand-new research – especially in cool technological or graphic ways – or pull very very large sets of data together in one place, where they can be standardized for storage and reuse. It is harder to make a case for tools that decentralize, reject extensive standardization, and propose to make the best of scholars’ current bad habits and idiosyncrasies, rather than trying to reform them.

In attempting to make sense of my own thoughts on this subject, I found it easier to construct a metaphor in which I could work out exactly why this bothers me (apart from my projects not being funded). The more I thought about it, the more I felt that most self-described “digital humanities projects” fall into three categories that can be expressed in terms of a medieval landscape of buried treasure, castles, and roads (you could use gold mines, gold-rush banks, and the railroad too, except people saw railroads as a good investment). The buried treasure is new data recovered through new research. The castles are repositories where that treasure can be deposited and curated. And the roads are the connectors and links that allow people to bring treasure to castles, or to go from castle to castle to visit different treasures, or to look for buried treasure in unexpected or distant places.

Right now I think the digital humanities are investing overwhelmingly in projects that promise buried treasure or the construction of an even better castle. Investments in treasure hunts can be risky, but they pay off with lots of shiny gold, and the discovery of a new treasure dazzles the public (and the media, and politicians, etc.). Investments in castles – especially the expansion of existing castles – are more like blue-chip stocks, safer if not as flashy. Not only are the structures larger, more impressive and more secure than a scatter of individual houses, but once their construction has been funded, they seem as if they should be inexpensive to maintain. And once they’re built, people will naturally want to put valuable things in them to protect them, and they’ll be willing to sacrifice some control over those things, since what other choice will they have?

Roads, by contrast, are boring, labor-intensive, and always demanding some kind of maintenance or additional investment (this is why the US currently scores a D for roads on its infrastructure report card). They are not sexy, unless they themselves are the object of research or represented in cool new ways. But without roads it’s harder to find buried treasure, and harder still to show it to others or compare it with treasures that have been found before. As my brother pointed out when I described this metaphor to him, there is one group that has historically recognized the utility of roads, from the Bronze Age to the present: the military. This comment had unintended resonance, since while our period gazetteer grants were being reviewed and rejected, a colleague forwarded to us a call for post-doctoral research put up by the US Intelligence Community which included a suggested program to develop… period gazetteers.

The IC call notes that information gathering is made easier by semantic-web features like machine-actionable gazetteers for place names, which allow natural language processing and disambiguation. It therefore solicits efforts to create flexible ontologies and authoritative or crowd-sourced vocabularies for divisions of time not based on clock or calendar, since “no existing gazetteer offers a solution for communicating other meaningful designations of time, such as periods and events.” This call is backed up on the Classical side with the appeal for a period gazetteer made by Tom Elliott at the Digital Classics Association conference and repeated in the closing address by Geoff Rockwell, who sees the need for gazetteers for the “primitives” of the discipline more generally.

The treasure-castles-roads metaphor also fits traditional scholarship, with new results from primary research as the treasure, libraries or archives as the castles, and print publications like books or articles as the roads. Digital tools and approaches make it easier and faster to find treasure, and make the display of that treasure in castles even more shimmery and alluring — but the treasure is still basically the same, as are the underlying architectural principles of the castles. In this respect, “digital humanities” just means “doing the same things humanities scholars have always done, except using computers to do it more quickly, more efficiently, and more attractively ” — in which case, everyone who uses a word-processing program or a spreadsheet or email is a digital humanist.

On the other hand, the digital versions differ in one important respect from their analogue counterparts: there is a much greater risk that we will awake to find that the shimmering treasures — only fairy gold after all — have disappeared, while the castles turn out to be nothing more than heaps of rocks in a field. Libraries and archives, the castles of traditional scholarship, hold their treasures as physical objects, and the roads between them are held open by the presence of publications in large numbers of copies. The latest platforms and interfaces used to discover and present new digital treasures, however, are almost certain to disappear. Both they and the contents of centralized repositories often represent the only copies of a particular dataset. If one digital castle falls, its contents may well fall with it. The old roads provided by publication are not sufficient for this new landscape: they cannot guarantee the survival of the primary data, and they cannot connect isolated castles with each other.

Roads in the digital landscape are significantly different from the roads in the analogue landscape. In traditional humanities scholarship, we have accepted a split between published interpretation and primary data, on the assumption that the latter can at least be consulted in physical archives in the future (especially in fields like archaeology, where the primary data cannot be replicated). We have also counted on persistent access to our treasures independent of the survival of any particular castle, since information tends to be reduplicated. When both data and publication exist primarily in digital form, however, and when individual repositories (or publishers, for that matter) can serve as the sole portal through which a particular dataset can be accessed, neither of those assumptions still apply. It is therefore all the more important in this environment to invest in new roads that provide links between castles and allow treasures — the data themselves, not only their interpretation — to be reduplicated and described more thoroughly to ensure their long-term survival. More and finer treasures and more and taller castles only exacerbate the problem.

All this leads me to a definition of the digital humanities that is somewhat different from the one I provided on the Day of DH. Digital humanities is not using computers to do better humanities research, or to answer the same questions with larger datasets, or to present our results in fancier, shinier interfaces. Digital humanities must be about the creation of digital tools and methods to facilitate unstructured exploration and discovery within the humanities — and not simply for a few lucky scholars, but for everyone, as Greg Crane has recently emphasized. Treasures and castles will always be important, but — at least in the eyes of an archaeologist — they will also always be ephemeral if they do not form part of a larger landscape of connectivity.

The projects that will have the greatest long-term impact in this landscape, I have come to feel, are neither the most polished nor the most centralized, but those that carry out the hard, slow, and distinctly unglamorous work of road-building. By the middle ages, the finest ancient buildings had become stables, rubbish heaps, at best the work of mythical giants, and proud new castles had been built by proud new kings. Now those castles in turn are inhabited only by wind and weeds — but in not a few parts of the Mediterranean world, you can still drive the path of a Roman road. “Digital humanities” will only represent a real transformation in the way we approach the humanities if digital humanists focus their collective efforts on new foundations for equally lasting paths of communication.