Community
June 9, 2012
DH Tools and Proletarian Texts
(This post originally appeared on the Proletarian Literature and Arts blog.)
I’m at the Digital Humanities Summer Institute at the University of Victoria this week. I’m taking a course on the Pre-Digital Book, which is already generating lots of interesting ideas about how we think and work with material texts, and how that is changing as we move into screen-based lives. There are, of course, many implications for how these differing textual modes relate to how we study and teach proletarian material, and more importantly, how class bears on these relationships. I hope to share some of these ideas as they have developed for me over the week. The course has taken up these questions in relation to medieval manuscripts and early modern incunablua and print, but the issues at stake are relevant for modern material as well. The instructors and librarians were kind enough to bring in a 1929 “novel in woodcuts” by Lynd Ward for me to look at – more on that will follow.
But, for fun, I also wanted to post about a little analysis experiment I did with some textual analysis tools.
I used the Voyant analysis tool to examine a set of Canadian manifesto writing. I transcribed six texts either from previous print publications or from archival scans for use as the corpus. These included: (1) “Manifesto of the Communist Parties of the British Empire”; (2) Tim Buck, “Indictment of Capitalism”; (3) CCF, “Regina Manifesto”; (4) Florence Custance, “Women and the New Age”; (5) “Our Credentials” from the first issue of Masses; and (6) Relief Camp Workers Strike Committee, “Official Statement”. [The RCWSC document remains my favorite text of all time.] Once applied, the tools let me read the texts in new ways, pulling out information or confirming ideas that I had about them in meaningful ways. You can find the summary of my corpus here.
The simplest visualization is the Cirrus word cloud, which at a glance shows that these texts are absolutely dominated by the language of class and politics (unsurprising, as they are aimed at remaking the existing class order). Michael Denning’s statement in The Cultural Front that the language of the 1930s became “labored” in both the public and metaphoric spheres is clearly reflected in this image.
Cirrus visualization of Canadian manifesto texts
Looking at the differences among the materials, an analysis of distinctive words is a simple way to get at the position of a given text in relation to the others. We might think of these Canadian manifestos as occupying the same ground of debate (though they are not responding to one another directly), but not necessarily sharing the same tent. For example, the “Manifesto of the Communist Parties of the British Empire” shows a much higher concentration of the term “war”, which helps situate it to later in the 1930s. The “Regina Manifesto” is overwhelmingly concerned with the “public” as it plans for a collective society. Florence Custance’s feminist statement shows itself to be more unique in its own time, as it uses “women” and female pronouns far beyond the other texts. And the Masses text betrays its literary periodical background with its heavier use of “art”.
The density of vocabulary in the texts can tell us something about intended readerships, and purpose of the text. Masses plays with the linguistic conventions of the manifesto to develop a text that is both assertive and creative; accordingly, it uses the largest variety of words to do so. However, the RCWSC is not far behind in its forthright call to action, which tells me something interesting about the role of the imaginative mode in connecting revolution with creative acts. Buck’s “Indictment” is the least dense text. It’s also the longest, which makes for a highly repetitive text. The “Indictment” has a strong oral quality to it, commenting on Buck’s trial and defense and with response and Marxist analysis. It is also highly indebted to that style, parsing its terms minutely and using them for step-by-step explanations. It is in many ways the most didactic of the texts, as the word density suggests, though such analysis misses the purposeful element of the limited word choices. I find Buck’s repetition to have an incantatory quality connecting it more closely to spoken debate than the other texts, an impression that comes out of working with the text closely, while typing and re-typing, and reading it aloud for myself. Word density is not for me an assignation of value; rather, it is one of many ways of framing some thoughts on how these texts – and manifestos more broadly – employ particular rhetorical modes and how we can follow them through.
Here is the link to the Voyant analysis of my manifestos. I invite you to take a look, play around, and consider throwing up some text from other working-class and proletarian sources. It seems to me that a lot of textual analysis begins by reaching for “important” texts – those that are canonical, or historical. The tools make no distinction – I would like to see more examples of writing from below feeding into the ways we think about texts in the DH realm.
Hi Andrea – just wanted to say that I think this is a really nice example of the way that word clouds can be useful in helping to give a general sense of what a text is about, and also when comparing texts from similar genres. Next time I use Voyeur with students I will point them to your post as an example!