This was then turned into the seventeen-volume collection in 1941, Slave Narratives: A Folk History of Slavery in the United States from Interviews with Former Slaves. Overall, this work features more than 2,300 accounts of life in slavery and is open to the public through the Library of Congress. These are all visible online through scanned transcripts that have also been transcribed. For my print project, I will analyze the tones and word choices utilized in these collections to form a glimpse of the experience of slavery by state and gender.
While not many primary accounts exist by African Americans on their experiences for various reasons, the WPA interviews offer historians a way to examine slavery through the memories of those who survived until the next century to recount their childhoods. This is where the interviews are often criticized. Since these accounts are based off of memories, it’s often argued that these are questionable in accuracy. However, this does not dismiss that they are excellent tools to examine history as long as you keep that fact in mind. I first came across the heavy use of these interviews in Deborah Gray White’s seminal piece, Ar’n’t I a Woman? As a historian on women’s perspectives in slavery, White admits that there are not many sources to find that include first hand accounts. In fact, her book was one of the first to be written on the topic in the 1980s. Yet through these interviews, White was able to draw a narrative to discuss what life was like for enslaved women.
To add further analysis to this topic, it would be interesting to examine all of the WPA interviews side by side to see the similarities and differences by state and gender. As stated above, all of these interviews are digitally accessible thanks to the Library of Congress. I would then upload these to the database that Jockers used in Macroanalysis to determine if there are any patterns in speech between former slaves of the same state and gender. I would also look at the overall tone of these pieces to see similarities. Since some of the interviewers were the grandchildren of their former masters, the interviewees are often believed to be reserved in their recounts due to this fact. It would be interesting to track a correlation between interviewers who knew the participants and those who had no connection. The tone and word choice would change, if I were to guess. I would further utilize Voyant Tools to track if similar adjectives were being used to describe former masters. This would also help to establish the tone of the pieces and the memories of the formerly enslaved.
This would be a huge task to undertake in such a short span of time, so in that sense this proposal is mostly theoretical. That being said, the tools are all there to someday turn this into a feasible project that could offer new insights to African American slavery studies in the United States.
Feedback is always appreciated 😉
The Slave Narratives are an invaluable resource. It would indeed be interesting to explore trends within the corpus.
You’re observations about the limitations of approaching retrospective accounts of the past are well taken. With that noted, when these texts are approached for what they are, recollections on the past at the moment they were created, then they persist as invaluable sources of memory. That is, they are authentically of their moment in time and valuable as sources for exploring that moment.
Getting the text of these prepared and ready for computational analysis is going to be a good bit of work. Along with that, as you note, it’s at a really massive scale.
So if you did end up wanting to work on a project like this for the course, you would likely want to zero in on some subset of the texts to pilot the work with them. I would also add that the specificity of what analytic methods you would use are going to have a lot to do with how you go about organizing and prepping the texts, so you would want to do some piloting with a few narratives for starters before getting to far into the work.