“War,” what is it good for?

We historians like words. So we just love it when computers let us count words really quickly. In fact, I’ve even done it once or twice. So we should all be really interested when the AHA (courtesy of Robert Townsend) comes out with a study of lots of words. The words in 23,000 history dissertation titles since 1920, to be exact. Summarizing the results of the study, reportage from Inside Higher Ed included the following:
‘For the recent titles [1993-2012], some of the analysis may challenge conventional wisdom about the state of the disciplines. There has been much discussion in recent years from some historians who say that issues of race, class and gender have come to dominate history, at the expense of traditional studies of politics and war. But the new AHA study found that “war” appeared in 11 percent of dissertation titles and “politics” appeared in 7.6 percent of titles. By contrast “women” and “gender” appeared in 7.8 percent of the titles, and “race,” “ethnic” and “ethnicity” appeared in only 4.5 percent of the titles.’

I was struck by this interpretation. And I was reminded of the dangers of using single words to conclude too much from a work. I was particularly curious about that word “war. ” What is it good for? What does it really tell us about the extent to which a study adopts a ‘traditional’ approach to war? Does it tell us whether the author is doing ‘military’ history, and if so, what kind?

I don’t claim to know the answer. But while I don’t have those 23,000 dissertation titles at hand, I have played around with Google Ngram Viewer, and done some basic analysis of the paper titles for the upcoming SMH conference. And I’ve noticed how the word “war” really does need some kind of semantic markup (as the digital humanists like to say), because the word “war” could mean many different things. Often times it may refer to the study of war in a traditional sense, but just as often (not sure about the exact proportion) it is actually being used as a shorthand for a timeframe, rather than an approach, much less a subject of study. You can probably think of various titles that deal with, say the homefront during World War II, as examples. These examples clearly indicate a willingness to use the term “war,” and even study life during wartime. But that’s not exactly the same thing as military history, or at least not a traditional approach to military history as the above quote suggests. (I’ll ignore here the odd fact that ‘war history’ sounds more military than ‘military history.’)

If we really wanted to get a sense of the relative frequency of various historical subfields, we should be a bit more sophisticated. Better than “war” might be “wars” and “warfare,” since I’d guess it’s mostly military historians who care about multiple wars and the waging of war. More useful might be terms that traditional military historians actually impart substantive meaning to, that is to say, their disciplinary jargon: military, army/armies, navy/navies, battle, siege, strategy, tactics, campaign, etc. Check out my previous post if you need some additional examples. We might even rank such terms according to how ‘traditional’ they are. You could also search for specific battles, even the years of famous campaigns (1494, 1704, 1812, 1815…). Perhaps best of all would be some combination of “war” and these other terms used by traditional military historians. In short, I wonder how useful a frequency count of “war” really is.

Better software which lets us to see collocations and co-occurrences will allow us to analyze phrases and not just single words (“unigrams”) – Google Ngram Viewer even offers a little of this basic functionality now. As we improve our methodology, we will undoubtedly achieve a better understanding of our discipline and field.

5 thoughts on ““War,” what is it good for?

  1. Pingback: War by any other name | Skulking in Holes and Corners

    • Maybe for some things (like war*). Can regular expressions use synonyms, find collocates, do topic modeling…? I didn’t think so, but I could be wrong. What we’d really want would be to find the collocates/co-occurrences that go along with war, wars, warfare…

      • Collocates: sure. Synonyms: sure. Topic Modelling: mmm, maybe not.

        A RegEx can be as simple or as complex as you like. The art – and discipline – is figuring out what exactly the pattern is you’re trying to find, then figuring out how to construct the RegEx to search for that.


        (P.S.: I’m not great at constructing RegExs. I’ve used them a bit, but not for a while and the skillfade is pretty steep.)

        • Thanks for the info. Sounds like you’re volunteering to do a tutorial on the blog! I did find some discussion of using regex when searching for two given terms near each other, but beyond that, it looks a big hard-core. I think most historians will probably stick to the more user-friendly tools (Voyant, TAPOR…). In any case, we should be talking about what we can do with these a lot more.

Comments are closed.