“She was not a woman, but a mere machine for reading and writing”(!)

(Many of us, appropriately, read Gissing’s New Grub Street for our Victorian Fiction seminar this week)

 

In her blog post, Lucina helpfully outlines the work computers, through machine learning, are able to perform in Ted Underwood’s chapter. Of course, Underwood’s algorithm is also intensely connected with the human—he explains the process of selecting the texts to be analyzed, the time period selected, the periodicals considered, etc. Equally obvious is the human work of interpretation Underwood performs to make his results legible to readers, particularly in connecting his finding of gradual “diversification” of a reading audience as a corrective to Huyssen’s “great divide” between “high art and mass culture” (103-4). However, I was also really struck by the way Underwood’s methodology required not just that he interpret the data once it has been run through the algorithm, but that he “read over [the model]’s shoulder” (82). It seems, as the model works to read trends in literary prestige, one at the same time must read the model’s readings to have any kind of detailed understanding of what is being measured. 

 

I was also fascinated by the moments in Underwood’s study where he calls on the specialized knowledge of other literary scholars not only in the design of his study (he creates the initial list of periodicals by “quizzing friends who are who are scholars of this period” (this is also how I figure out many things but have never this kind of informal canvassing so explicitly acknowledged in a piece of scholarship)) but also in its implementation. Underwood explains that he presented pages from both reviewed texts and random samples to graduate students and professors who study 19th and 20th century literature, which proved that the model could differentiate where humans could not (79). I was left wondering, how this kind of method of sociological survey of scholars could be used to ask other questions and possibly reevaluate other issues across literary fields.

Leave a Reply