Meditation on Metadata

making boundaries between evidence and interpretation, and crossing them.

This fall I am working with librarians and technologists at Northwestern University’s library to digitize one full year of the Berkeley Folk Music Festival in anticipation of my spring 2012 history research seminar, Digitizing Folk Music History.

One fascinating aspect of preparing to move from the Special Collections archive to a digital repository is the question of metadata. As archivists, historians, and technologists increasingly work together, cooperatively, we find ourselves considering what should accompany an archival object when it takes virtual form. Where does the object end and the interpretation begin?

My revelation as a historian who has done his time (pleasurable labor I might add) in actual archives is that what librarians call metadata might extend to all the work historians have traditionally done when archivists hand over documents to them. In one sense, all historical work—monographs, articles, syntheses, teachings, conversations, student essays—are metadata. They are elaborations, interpretations, analyses built upon the empirical evidence in archives (from written to multimedia to oral archives).

But then again, all these elaborations, interpretations, analyses in turn become the very data of future historical study! When my class last spring decided we wanted to contribute student digital projects back to the Berkeley Folk Music Festival archive, we were, in a way, turning the metadata back into data.

To quote the stoned hippie-beatnik inside me, “Dude!”

So one of the exciting challenges of digital history is to think about this space between data and metadata that the virtual archive enlivens, that the digital makes tantalizingly visible in all its complexity.

This is the very site of historical meaning-making, the place where the rubber of empirical evidence hits the road of interpretative discovery (is that a tortured metaphor or what?).

The question becomes: where do we close the sides of the database field and where do we open up the architecture into and out of which data and metadata flow?

To get anywhere answering this question, we’ll need a humanistic conceptualizing of data and metadata, not a rigid engineering one. We will need a supple sense of how meaning works, how it slides around, how it isn’t just a machine or an algorithm or a simple code, but a complex, living system filled with imperfections, contradictions, malleable shapes and forms, and profound possibilities as well as difficult challenges.

That’s as much a cultural theory question as a computer science one—only by bringing the two approaches together can we think about the past in new, futuristic ways. For this one, we’ll need to go meta to gather the data.