The humanities make for awkward data

I’m in the earliest stages of a project that hopes to combine data visualization and network analysis to cast light on English medieval scribes, medieval manuscripts, medieval texts, and the unstable connections between them. While generating a small sample dataset in order to help conceptualize what usefully-structured data might look like for the project, it has become abundantly clear how awkward the fit can sometimes be between the digital and the raw, messy “stuff” of the humanities.   There is an enormous amount of uncertainty in the things I work on: books dated to within a quarter century or so, texts dialectally localized not to lat./long. but to a county or a large chunk of the country, manuscripts about which we know very little until they entered libraries and collections in the 18th century.  It’s not simply that we lack metadata for medieval books and texts, but that the nature of the metadata, and expectations for the information, are simply different when looking at 600-year-old books.  There are ways to accommodate these uncertainties in nice, tidy schemas, of course, but one thing I hope to spend some time thinking about this weekend is what messier schemas might look like and how they might work: what digital structures more closely modeled on the analog sources of the humanities might resemble, what ignorance might look like when acknowledged and included rather than elided.