Workbook for Introduction to Digital Humanities: A-State

Edward's Text Analysis



In American author James Fenimore Cooper’s 1847 novel The Crater, the main character must not only survive on a deserted island, but even manufacture from volcanic rock the very soil in which to grow his crops. Instances such as this in the story seem analogous to the process of distant reading as a whole—deconstructing a monolith to its most minute pieces, then reassembling it in ways that better serve other purposes.
 
For this assignment, it helped that I had done a close reading of The Crater several years prior, and thus was able to see instances where Voyant’s data-crunching occasionally confirmed my suspicions, but more often opened entirely new dimensions of potential interpretation. I had known that Cooper and other authors of his era often wrote lengthy and descriptive passages, but I was surprised that the average sentence length of The Crater was 29 words—up to twice the recommended average length today of 15-20 words.
 
The most helpful feature, in my opinion, was the trends feature located in the upper-right of the dashboard. Tracking the prevalence of selected words over the course of a novel provides a way to gather and present data without dismissing with the linearity of the book’s plot altogether. Instead, for instance, it allowed me to see that the narrator of the book focuses less on the main character in the second half, while the usage of the word “colony” grows after that point. Such minute details would have been too time-consuming and even too trivial to determine in the days before text analysis, but their current ease of computation means that they can now contribute to more multifaceted and nuanced arguments.
 
The version of the book used for this analysis was the Project Gutenberg text, published in 2004. While at first I had challenges stripping the Gutenberg boilerplate additions from the beginning and end of the file—Notepad would automatically add the sections back after I ostensibly saved my edits— I finally succeeded in making sure that the only thing being analyzed was the text itself. Voyant’s upload process was well-designed and, for only one document, deceptively easy. The range of tools and their different visual impacts was quite impressive, and I appreciated that manipulating one of them caused real-time responses in the others. Still, some of the tools are not necessarily the most useful in every case—the correlations section only found words that occurred near each other just once in a chapter, for instance.
 
Overall, Voyant is an impressive tool that has the potential to make some features of distant reading as ubiquitous as a browser search. As many have stated, though, its greatest potential lies in complementing close reading, not downright replacing it.



Resources Used:
Sinclair, Stéfan and Geoffrey Rockwell, 2016. Voyant Tools. Web. http://voyant-tools.org/.
Cooper, James Fenimore. The Crater; or, Vulcan’s Peak: A Tale of the Pacific.  1847. Project Gutenberg, 2004. http://www.gutenberg.org/ebooks/11573


 

This page has paths: