I have been writing a new piece about using computational modeling to study the legacy of Augustinian conversion in the modern novel. The project began when I tried to devise ways of detecting greater or lesser degrees of narrative binaryness across a broad range of texts, including novels and autobiographies. I was interested in studying where a sense of revolutionary change as a key marker of the individual life went in the modern period. The idea was that conversion would be strongly marked by a difference in language use before and after the experience, a feature that was noticeably on display in Augustine’s Confessions, one of the Ur-texts of narrative conversion in the Western tradition.
The challenge that I have been finding — apart from getting statistical analyses to work, not break, and be repeatable — is how to talk about these quantitative measures in ways that are intellectually as well as aesthetically satisfying. I’ve been finding that the language one uses to talk about numbers is very different than the language one uses to talk about language. This may be obvious to anyone who has ever read a computer science article, but I initially thought that this was just a problem of disciplinary training. With the right literary sensibility, surely numbers could be made to sing or at least be enjoyable to read.
Not so. The essay I’m writing right now is so clearly divided between the interpretive and theoretical parts and the quantitative parts that I’m not sure it is a surmountable difference. I have no interest in burying either half, however. I didn’t go into this to become a social or computer scientist, but I also didn’t start experimenting with computational reading just to shove it under the rug (“why not put it in the footnotes?” as one respondent once remarked). My goal was, and is, to see how reading quantitatively and computationally changes my/our relationship to literature.
This is all to say that writing about books has historically been based on a certain way of talking about language, one that is in flux now that there is a different technology mediating our relationship to language. It seems to me a long road to integrate this type of analysis into the tradition of essay writing and criticism, whether academic or popular. Among the many challenges this transition has posed for me, this was one I hadn’t foreseen, but it is as of now a very strong one.
How might we develop ways of talking about quantity that are as cognitively and stylistically satisfying as talking about the woof and warp of words?