Data Visualization of Drafts
In the back of The Poems of Emily Dickinson (Reading Edition, Belknap Press, 1998), R. W. Franklin includes some data: the number of extant poems Emily wrote from 1850-1886. It is basic, but as Dickinson scholars know, much discussion emerged and continues to emerge regarding the Civil War years. From 1861 to 1866, Dickinson wrote, respectively, 88, 227, 295, 98, 229, then 10 poems. What happened between 1865 and 66? What can account for this steep decline? This is not the place to provide a gloss on the scholarly discussion surrounding this data. Rather, I simply highlight the fact that scholars and students find the sudden drop perplexing and bewildering. As such, the simple data set begins to tell a story. It raises questions. It builds suspense. Data sets—even simple ones—often lead to breakthroughs in thinking.
Ever since I first saw one of Cummings’ drafts, I have been thinking of the ways they impact how we understand and read his published poems. Cummings himself would cringe at putting his organic process of poiesis into an excel sheet! And I am the first to admit that every approach we have to reading poetry eclipses even as it reveals. Nonetheless, I find the data visualization below to begin to tell a story about Cummings’ process of making a poem.
At this point, I include data visualization on just one poem: “silence” (CP 712).
A few preliminary observations.
- The final poem is 1-3-1-3-1 lines per stanza. However, in the drafts, he uses this pattern only twice.
- He doesn’t add or subtract a single word of “silence” across roughly 32 drafts. This suggests that, for at least this poem, Cummings grappled not with what to say but how to say it.
- He does add and then withdraw a hyphen. Also, in the drafts, a closed parenthesis is used much more than the open one, which is the parenthesis that found its way in the final poem.
Here are the drafts, beginning, as the Houghton arranged them, from the last drafted sheet to the first drafted sheet.
To collect accurate data, I typed out each draft in a Word Document (click here for pdf). I made the penciled characters a different color than the typed characters. Upon finishing, I searched the document for each word and punctuation mark. Using the Word Count feature, I quickly could see how many total words were in the drafts as well as the number of characters typed or penciled . . . and so forth . . . put the data into Excel, and then into Tableau. The result? Basic data.
However, this data tells a story. The simple fact that the final poem is made out of 14 words (I count “turn” as well as “turning”), but the drafts contain 635 words tells us that Cummings worked hard at his serious play. Indeed, he put this short poem through 30+ drafts. New readers of Cummings often assume that he had a reckless, haphazard approach; however, the data points toward the sheer attentiveness to every mark and blank space of his craft. Additionally, Cummings is known to be the poet of the typewriter; however, the data shows us that he used the pencil in the drafting process much more than one might realize. Other stories, too, wait to be told from what our qualitative and quantitative data might reveal.
In the near future, I plan on adding a few more examples of one way of approaching the drafts of Cummings; the following is just a start.
Aaron M. Moe, Ph.D.
Saint Mary’s College, Notre Dame