Mapping CanLit and Austin Clarke

This post is part of an ongoing Digital Scholarship Ontario series which highlights the scholarship and teaching experiences of Ontario’s digital scholars.

by Paul Barrett

My current research is driven by the perplexing problem of trying to understand Austin Clarke’s marginalization in Canadian Literature. Clarke is one of our earliest, most prolific, and one of our most awarded authors. He has written 11 novels, many short stories, poetry collections, and his archives at McMaster (where I work) are extensive. Despite all this, however, there is very little scholarship on Clarke — even after winning major awards for The Origin of Waves and The Polished Hoe, Clarke’s writing still have not received much scholarly attention. My current research uses digital humanities research to explain this absence in Canadian literary scholarship and to offer new methods of interpreting Clarke’s work.


One way that I consider Clarke’s place in Canada is by using the method of topic modelling. Topic modelling is a digital method of analyzing a large group of texts. A topic modelling algorithm surveys the texts and identifies the underlying topics that define this collection of texts. Topic modelling purports to identify the hidden, underlying themes that define a collection of texts and thereby provides a new way of interpreting a collection of texts.  So my project involves both investigating the methods of topic modelling (Are they actually helpful in helping us understand a collection of works? Do they tell us something that we can’t learn through regular forms of reading?) and then using those methods to see what themes emerge from a topical analysis of Clarke’s writing.

To do this, I have digitized all of Clarke’s novels and short stories and used the topic modelling software, MALLET, to create a number of topic models of Clarke’s work. I have also digitized and modelled the entirety of two Canadian literary journals: Canadian Literature and Studies in Canadian Literature/Études en littérature canadienne. My thesis is that a comparison of the two models will reveal interesting and unexpected connections between Clarke’s writing and the broader field of Canadian literary studies. Right now, I’m in the process of comparing these topics. Visually, they look like this:

https://i0.wp.com/paulbarrett.ca/wp/wp-content/uploads/2016/05/CanLit-80-Keys-Azure-Blue.pngOne surprising dimension of this work has been the discovery of how topic modelling is not a completely automated process where a bunch of texts go in, and some meaningful topical information comes out. Rather, the user needs to make a number of critical decisions that affect what kinds of topic output will be produced: they must choose the number of topics, the list of excluded words, and a series of other choices, all of which affect the results. This is one obvious way in which digital humanities work is revealed to be not an instrumental, pseudo-scientific process, but an act that requires interpretation and human intervention; in this way digital humanities work is a lot closer to our traditional ideas of the humanities than we may have realized.

Topic modelling is just one dimension of my Clarke work. I’m also looking at his use of nation language (the black language of the Caribbean) to see how it changes over time. I’ve also been mapping movement in his texts to see where his characters go over time and comparing movement throughout his texts. Movement is very important in Clarke’s work and this mapping provides an alternate way of reading Clarke’s work and comparing the possibilities and restrictions on movement that his characters experience. A screenshot of the prototype of this mapping:

https://i2.wp.com/paulbarrett.ca/wp/wp-content/uploads/2016/05/Screen-Shot-2016-05-30-at-11.52.37-AM.png

This research will form the basis for a forthcoming website entitled Austin Clarke’s Aesthetics of Crossing. In a broader sense, I treat this research as an opportunity to experiment in the digital humanities. The digital humanities is an emerging field full of possibilities but I think both the conclusions that we draw and the tools we use need to be more effectively theorized. Therefore, I hope this work provokes both these kinds of theorizing of digital humanities tools as well as the beginning for thinking about what it might mean to do digital humanities work specifically in Canada.

Paul Barrett is a Banting Postdoctoral Fellow in the Department of English and Cultural Studies at McMaster University. He is the author of Blackening Canada: Race, Diaspora, Multiculturalism (2015). He has recently published articles in ARIEL and Studies in Canadian Literature. His writing is concerned with questions and representations of race and diaspora in Canada and global contexts. His current research assesses the role of digital forms of textual analysis in Canadian literature. The most recent manifestation of this work is a digital humanities project that employs algorithmic forms of ‘reading’ to survey Austin Clarke’s archives and corpus.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: