If you’re in a history graduate program, odds are you like to read. A lot. You’ve probably already discovered which areas of history interest you the most and you’ve made it your mission to read as much about it as possible. But on your first trip into the archives in search of primary sources, you probably realized with a sense of dread that it’s physically impossible to read and give a proper analysis of everything that’s out there.
Computers have made this process significantly easier. Archives are digitizing and transcribing more and more historical documents. We have the ability to run searches of key words and phrases to locate sources. But as Matthew Jockers writes in Macroanalysis: Digital Methods & Literary History, “Close reading, digital searching, will continue to reveal nuggets, while the deeper veins lie buried beneath the mass of gravel layered above” (9).
So how can we access the deep recesses of the archives without engaging in a close reading of the sources we find? According to Jockers, the answer lies in computational analysis. In light of technological advancements, Jockers argues that historians and other practitioners of the humanities need to embrace new methodologies for utilizing the massive amount of information available to us through digitization. Macroanalysis, or distant reading, allows scholars to look at patterns and trends as they appear within an entire collection. Here are a few examples of how Jockers uses macroanalysis in the field of literary history.
Jockers examined 758 works of Irish-American literature over a period of 250 years by using metadata embedded in bibliographies. He was able identify larger trends in the literature based on the gender of the author, geographical region in which the work was produced, and the setting in which the story takes place (among other variables). With the full-texts available digitally, Jockers demonstrated how computing software could trace word use, book length, and ethnic markers. In subsequent chapters, Jockers examined how macroanalysis could provide insight into the style, nationality, theme, and influence of a large body of literature by looking at something as simple as the use of “the” or how many times the author used a particular pronoun. He describes how literary scholars could use correlation coefficients, topic modeling, and information cascades to draw conclusions about how texts relate to each other over time.
Some of you may be thinking that this sounds a bit too quantitative for you. (It sounds too much like math to me, and I hate math!) After all, the humanities is about contextualization and interpretation, which is absent from a distant reading of the literature. Jockers agrees with you. He points out that there are significant drawbacks to conducting macroanalytical research. These massive data sets produce outliers and exceptions. There is always a possibility that the metadata is incorrect (Jockers points out that in many instances the publication dates and author’s genders were incorrect). Computational analysis cannot provide context and interpretation. And in many cases, sources and books are not available digitally due to copyrights laws and the significant time it takes for archives to digitize their collections.
The main point Jockers is trying to make is that macroanalysis should not take the place of microanalysis. They “should and need to coexist” (9). A macroanalysis of literary or historical data can help scholars better understand the broad patterns and trends of their research interest. Those patterns and trends, taken from a massive data set, provide the context in which a close reading of a smaller set of texts can take place.
Jockers book is an excellent resources for those historians resistant to change. He is not lauding computational analysis as the new way of conducting research in the humanities. He is urging scholars not to ignore a tool that can provide new insights into their field of study.
As a diplomatic historian, this methodology is intriguing. If any of you have ever perused Foreign Relation of the United States (the last one I looked at was 4,752 pages long in epub form and only presented selected State Department documents on the Suez Crisis), you will recognize the benefits of having a tool that can identify trends in language and style in diplomatic correspondences or determine which themes are the most prevalent. As David Allen and Matthew Connelly suggest in “Diplomatic history after the big bang: using computational methods to explore the infinite archive,” distant reading can reveal what kinds of documents are more likely to be redacted or rank documents according to their relevance to your research interests (83). The results of this type of analysis are not sufficient by themselves, but when used to inform, conceptualize, and direct your microanalytical research, they are extremely beneficial.
Allen, David, and Matthew Connelly. “Diplomatic History after the Big Bang: Using Computational Methods to Explore the Infinite Archive.” In Explaining the History of American Foreign Relations, 3rd ed. Cambridge: Cambridge University Press, 2016.
Jockers, Matthew L. Macroanalysis: Digital Methods and Literary History. Urbana and Chicago: University of Illinois Press, 2013.
3 Replies to “Macroanalysis: Applying Computational Analysis to History”
Thanks for the post Erica, I am totally with you and Jockers. I don’t think that macro-analysis can replace micro studies, qualitative analysis, or close readings, but the doors macro-analysis opens are wide ranging and numerous. As historians we’re always looking to build on past research, reinterpret old sources to make new arguments, and identify important gaps in the literature our predecessors might have missed. I think macro-analysis has the potential to guide our work, help us find what others have missed, and provide a survey of the field as is. Of course it has faults and limitations, but what method of inquiry doesn’t?
I agree entirely! I appreciate, as you point out, Erica, that it seems these tools can alleviate some of the stress of recognizing it’s impossible to read or even just be familiar with all of the available documents and sources. Digital tools can better point us towards what we might otherwise miss in our own research. I’m also intrigued by Alex’s point about identifying gaps in literature that our predecessors might have missed. I’ve always wondered how it’s possible to feel knowledgeable enough even in your own subfield, to claim being an expert. Macroanalysis seems, to me, like it could be a way to make sure we’re engaging other scholars’ sources even more deeply. I think reading this book and set of articles helped me recognize the ways digital tools strengthen historical research and made it seem more attainable with practical examples of digital tools being put to work. Several of the readings, but Blevins in particular points out the importance of collaboration because it’s impossible to be familiar with all of the available digital tools. I found this to be very encouraging because (like you) math terrifies me, so while some of this book was a bit quantitative for me, I take solace in knowing that I don’t have to understand it but can collaborate with someone who does!
Maron, I think Blevins’ article does an excellent job applying Jockers’ argument to the historical field. I found his conclusion to be particularly relevant. Historians are needed more than ever because of “the kind of skepticism” they apply to all new methodologies. Historians need to assess the results of macroanalysis for biases and silences while taking into consideration the circumstances under which the sources were produced and the complexities hidden beneath the deluge of statistical data (147).
His comments on the importance of collaboration eased my mind as well! And as this methodology becomes more established, I hope that more universities will incorporate technical training in the computer sciences to facilitate its use in the field.
See Blevins, Cameron. “Space, Nation, and the Triumph of Region: A View of the World from Houston.” The Journal of American History, June 2014.