22 February 2012

What's the count in the American empire debate?

As part of a project I'm working on this term, I ran these Google Scholar n-grams for a professor. (They aren't the first or even the most impressive ones I've run, but they are the ones that I'm most behind on sharing.)

The first corpus is a broad overview of the frequency of the use of the term "American empire" in Google's corpus of American English.

"American empire", corpus: American English, 1850-2008

What emerges is a somewhat surprising pattern: There's a long decline over the nineteenth century, a blip following the acquisition of an actual American empire in 1898, and then a long plateau following the United States' inadvertent victory in World War I.

The decline that set in after the height of Vietnam War guilt is reversed almost immediately by the events of September 11, which sent discussions of "American empire" to an all-time high (that's right, even eclipsing the period of time when we stole--there's no other nice word for it--Hawai'i, Puerto Rico, the Philippines, and Guam!).

"American empire", corpus: American English, 1968-2008
 This trend is confirmed by an examination of the British English corpus (more discussion below).

 (Note that although this chart and the subsequent one include the case-sensitive "American Empire" in addition to "American empire," the trend lines for all are essentially the same.) Here, we again see the same post-9/11 jump (and an odd post-World War II blues jump in the proper noun "American Empire"). (Note that the British were much less concerned about Vietnam as an exercise in American empire-building.)

"American Empire" and "American empire", corpus: British English, 1968-2008
Finally, we turn to the all-English corpus (which presumably includes the dominions of Canada, Australia, New Zealand, and Wee Britain). We see the same pattern regarding post-9/11 empire talk.

"American Empire" and "American empire", corpus: English, 1800-2008
P.S. Just for fun, the rise and fall of the "British Empire":

"Roman Empire", "British Empire", "Ottoman Empire", corpus: English, 1800-2008


  1. The divergence (and eventual convergence) of e and E is interesting. Maybe reflects a discursive shift from "empire" as a normal mode of great power foreign policy to "Empire" as an anachronistic and stigmatized political form.

    1. I think this actually may be correct. I've played around with more of these than I present here and the power of n-grams to analyze discourse should not be underestimated.

  2. Is there an easy way to generate data sets using google that you can use for more than producing graphs?

    1. Ah, "easy"--no .... you can download the data but that is just kind of a hassle.