Monday, September 28, 2020

Beginning with a dataset of 300 African American texts

By Kenton Rambsy and Howard Rambsy II

In a limited amount of time, how do we make students in our classes aware of the vast body of texts that comprise African American literature? That’s one of the main questions and challenges that we face as professors. There are thousands of texts and a relatively small number of weeks to cover the material. Still, given our ongoing Digital Humanities projects, we considered how we might most effectively organize, analyze, and share large bodies of information.

[Link to list of 300 African American Literary Texts]

As one possibility, we created a visualization that focused on mentions of 300 African American novels, essays, poems, autobiographies, slave narratives, and short stories. The bubbles below represent the range of scores. The bigger the bubble size, the higher the weighted score (and number of citations/references).

Metacanon served as a useful model for how we conceived of this project. The website tabulates and ranks novels based on an accumulation of citations in thousands of articles.  We wanted to apply a similar approach but with a focus on black literary texts, and we sought to chart the mentions of works by Black writers across publications and search engines. 

We chose the 300 texts in the dataset based on our own interests and knowledge of major, recurring topics in our field as well as works frequently cited by our friends and colleagues. Thus, while many of the texts that appear in our dataset are deemed canonical, some were selected for personal and arbitrary reasons.

We collected information about the texts to get a general sense of what focal subjects have persisted and changed over the decades. We were not seeking to be comprehensive, as we certainly understand that thousands of texts comprise African American literature. We merely wanted to choose a large enough number of items that would be suggestive of the range and complexity of the field.

We focused on five databases: JSTOR, Project Muse, ProQuest, Google Scholar, and the online archives for The New York Times archives. We enlisted Peace Ossom-Williamson to help with computational methods. We used the returns from the five databases to calculate and assign each text a score. We used the same formula as Metacanon, and took the square root of the counts; however, we did not weight any particular database as higher than another.  Then we took the sum of all the points together.  

Works by some authors like William Wells Brown and Pauline Hopkins appear in scholarly articles but not in The New York Times. Conversely, recent books by writers like Colson Whitehead and Ta-Nehisi Coates appear frequently in newspapers, but not in scholarly contexts. Our findings do not prove canonicity. Instead, this chart gives users a sense of the frequency of mentions in tens of thousands of articles.

After calculating the numbers of references, we ranked them from the most to least.  We also organized the ranking into three separate tiers. The first tier consists of works that have a score of 20 and above. These works are frequently cited. The second tier consists of works that have a score of 19 – 10, and the third tier has scores ranging from 9 to 0.

Since 1980, select novels were among the most frequently cited works. Fifteen novels appear among the top 34 publications, or the top tier. Toni Morrison’s Beloved (1987), Song of Solomon (1977), and The Bluest Eye (1970), are among the most frequently mentioned. Nonetheless, the text that ranked the highest on our list was W. E. B. Du Bois’s book The Souls of Black Folk (1903).

Two essays, Alaine Locke’s “The New Negro” and Alice Walker’s “In Search of Our Mothers' Gardens,” and only one poem, , “I, Too” by Langston Hughes and one short story “Sweat” by Zora Neale Hurston appear in the first tier.

But ranking the various authors is just one approach. The dataset also allowed us to consider differences among a few different texts by a single author. Morrison’s novels Beloved and Sula (1973), were referenced quite frequently, while her novel God Help the Child (2015) and her short story “Recitatif” receive far less citations. Similarly, Richard Wright’s Native Son (1940), and Black Boy (1945), are cited by scholars and journalists, but his book The Long Dream (1958), is mentioned infrequently.

It’s worth noting that only 11% of the works on are list appeared in the top tier, which is to say, most books do not receive long-term frequent mentions. It is somewhat rare for a book to receive many passing mentions over the course of more than two decades.

Even though the field of African American literature is wide and varied, certain texts are privileged over others, at least in terms of references or mentions. Those privileged texts constitute an African American literary canon, and what we’ve discovered is that some so-called canonical texts are perhaps more canonical than others.



Matthew Wilkens said...

Really cool work! Do you have a list of included texts that you can share? This would be a great starting point for both research and teaching.

H. Rambsy said...

Yes...the link is now included up above in the post. Here it is:

Matthew Wilkens said...

Thanks, that's super helpful. Much appreciated!

aearhart said...

Excited to see your work published here. Thanks for sharing.