Docuinformatics and Historical Modeling

When asked about a discipline that would benefit from a more cross-disciplinary research approach, Nick Matzke proposed a data-driven approach to the study of history. While Matzke stressed that there is no replacement for "old-fashioned, document-based, interpretive history" he does sketch out a fascinating notion of quantitative history and wonders out loud if it might be possible to determine the degree to which various philosophical and aesthetic influences shaped the thinking of Adolf Hitler, and in turn the direction of the 20th century. Spurred by the ongoing digitization of everything, Matzke asks: now that we've amassed this huge body of documents, how can "the corpus" inform historical research?

Like Matzke, I'm hesitant to imagine that history will be completely reinvented, but it is safe to assume that the "modeling" and analysis he has proposed will undoubtedly start to augment traditional scholarship. A quick scan of some developments in related fields might provide some clues as to what the near future might hold.

  • Pictured above is Jeremy Douglass of the UCSD Software Studies Initiative demonstrating the application of cultural analytics to organize abstract painter Mark Rothko's entire body of work as assets in a database. The traditional archive would limit cataloging art with relatively pragmatic fields such as year created, dimensions of the piece, materials, etc. This system actually takes the image content into consideration which allows compositional or color palette analysis. Aided by these extra parameters, a researcher might be able to isolate, quantify and critically frame specific pockets of work or tendencies that may not have been noticed otherwise. In Douglass' words, cultural analytics allows for the consideration of paintings as "not only aesthetic objects, but data."
  • An earlier venture that occupies similar territory is Hans Rosling's Gapminder, a foundation with a self-described mandate to "use statistics to promote a fact based world view." Gapminder is well known for developing Trendalyzer, an animated statistics tool for visualizing and exploring global demographic information.
  • The most recent edition of The Programming Historian (written by William J. Turkel, Adam Crymble and Alan MacEachern) asks the following of historians: "Do you need to learn how to program?" The authors' answer: "If you don't program, your research process will always be at the mercy of those who do." That is not to say that Matzke's "docuinformatics" demand that all historians learn to code but if historians don't develop the appropriate tools, then who will?

This is admittedly quite a nebulous discussion and we haven't even touched on how semantic analysis might figure into processing large volumes of historical documents. The important thing to remember here is that documents are (in David M. Levy's words) "talking things" and we can assume that as a discipline, history is invested in mining as many associations and as much context as possible from available archives.

More like this

Below, Nick Matzke answers the second of our three questions. History--old-fashioned, document-based, interpretive history--is a field that is ripe for an infusion of new methods and technology. Much as the biosphere and the organisms in it have evolved, individuals, documents, schools of thought…
image: history flow edit log of the Wikipedia article on evolution Nick Matzke is ambitious when he exercises his imagination. In answering our final question, Matzke sketches out a methodology for tracking how public policies or scientific hypotheses were "copied, repeated, modified and…
Okay, if you're anything like me, you don't have time to read the blogs you already follow. But I do recommend that everyone head over to SEED's Revolutionary Minds Think Tank, where Greg Smith is guiding a conversation on visualizing science. That's where I found the video above, demonstrating…
Margaret Turbull's comments about knowing when to build bridges between disciplines (and when not to) is an appropriate final installment of the RevMinds interview series. Over the last several months we've heard our multidisciplinary experts sound off on the following timely questions: The…

That Cultural Analytics video is incredible! I wonder how effective programs like that would be for teaching art history to young students.

Very interesting! Wonder if techniques like network analysis and text mining can help making history a quantitative discipline?