“Pirating Texts” – an Interview with Grant Glass

Collage of Robinson Crusoe covers

Grant Glass is a PhD student at UNC Chapel Hill and a Graduate Fellow at Duke University, focusing on work in digital humanities and British literature. Grant led a research team during the summer of 2018 to create Pirating Texts, a project related to his dissertation work using algorithmic text analysis to find social, historical, and literary differences in various versions of Daniel Defoe’s Robinson Crusoe. Read his interview here:

How did you conceive of Pirating Texts?
I have always been interested in adaptation and found that the story of Robinson Crusoe was strangely retold over and over again (Cast Away, The Martian, Swiss Family Robinson, Lost in Space). My research on why this story was so compelling led me to the see variations in different editions of Robinson Crusoe. Soon enough, hundreds of editions became thousands and any hope that I could collate or read them all vanished before my eyes. I thought a machine could handle this large dataset and hopefully help me understand how many variations there were in the text.

How did you decide on which technologies to use in the project?
As a team, we looked at a variety of different methods and then looked at the results to tell us the differences. No one method is perfect, so we came up with a variety of different methods to use. Doc2vec became one of our main methods for parsing the texts.

What, if any, obstacles did your project face?
Access to texts is always one of the most difficult processes of any Digital Humanities project. Trying to access the plain text version of a work held by an institution like Google or Gale proves to be untenable for a smaller research project. The team spent most of the summer asking for permission from different libraries, web scraping and calling APIs to gain access to our texts.

How has the project changed and evolved over the process?
Many of the methods changed dramatically, like Text Tiling proved to be ineffective with this set of texts. Over the course of the project, we found that many editions changed very little or changed many of the plot points completely (like Crusoe never finds Friday).

If you had to describe the project in 5 words, what would they be?
Distant Reading meets Textual Studies

What’s your dream DH project (if you could work on anything with any skill set in any field, what would it be)?
I would love to be able to teach a machine Daniel Defoe’s style and try to have it find unknown works.


Image from https://researchblog.duke.edu/2018/09/06/robinson-crusoe/