Thoughts on Franco Moretti’s Distant Reading
John C. Hunter
Too much polyphony, and too much monotony: it’s the Scylla and Charybdis of digital humanities.
Franco Moretti, “Style Inc.”
As we have observed throughout our considerations of Moretti’s work, what he is interested in first and foremost are new methods of approaching literature. With this is mind, I think that it is important to begin by locating his methodology in the context of other branches of scholarship. His self-description as a plucky naïf trying new things without really knowing what will happen is charming and may work in a purely literary context (and I stress “may” – there were people involved in what used to be called “humanities computing” doing similar things as far back as the early 1990s). But his “distant reading” can also be described as a version of what scientists call a “meta-analysis,” and some attention to this context—about which Moretti should be much more explicit—can help us to understand the strengths and weaknesses of his methods.
In the two article abstracts that I circulated last night, you can see how this works in the sciences. In the Li, Mao, and Wei article, a group of scientists in Beijing studied literally hundreds of scientific studies written over a thirty-year period to look for patterns of connection between addiction and certain human genes. Experimental scientists are trained to be as reductive as possible (in the non pejorative sense of that term): try to isolate single genes, single causes, eliminate as many variables as possible. The Beijing team looked across all of these local, isolated studies to try to find patterns: are there genes that are consistently associated with addiction? Which ones and in what ways? No single team of experimental scientists could hope to determine this by themselves, but aggregating the work of hundreds of labs could and this is what they did. The result is a scientific research article in which the authors have done very little of the research themselves: they are just collating, interpreting, and analyzing the research of others. And this is a perfectly respectable scientific approach these days, one that is made possible by online access to research from many countries, increased data processing power, and the tyrannical leveling effect of the English language (the world-wide language of research science and a subject to which I will return). It is just like Moretti’s refusal to close read and his move to see what can be learned from far away, rather than up close – he is performing a meta-analysis of the literary corpus.
The Evans and Foster piece on “Metaknowledge” can be read as a theoretical reflection on this process – reading the sentence in their abstract about how meta-analysis “uncovers regularities in scientific claims and infers the beliefs, preferences, research tools, and strategies behind those regularities” and we can almost hear Moretti studying the history of the novel in local contexts all over the world and finding his pattern of the modern novel in the periphery: Western forms in a complex interaction with local material and a local narrative voice. Regularities, beliefs, preferences, and strategies . . . exactly what Moretti wants to find. Evans and Foster are sociologists of science rather than bench scientists themselves, and they represent another broad current of academic inquiry that bears on Moretti’s method (one that goes back to Thomas Kuhn’s path-breaking Structure of Scientific Revolutions ). Kuhn outraged many scientists by asserting that there were patterns that scientific activity follows that are, in effect, unconscious, and Moretti is saying the same thing: think of his fascinating observations on the length of titles in “Style Inc.” – how they got shorter, why, and what patterns can be observed in the process. As with the scientists, the novelists didn’t consciously collude to do this to their titles (one can’t imagine Jane Austen thinking to herself: “What shall I call my new novel? The Education of Miss Elizabeth Bennett, perhaps? Nah – too long . . .”), but Moretti has shown (to my satisfaction at least) that the process DID happen.
The point here is not to denigrate Moretti or to imply that he is dishonestly piggy-backing on work done in other fields (though I do think it would behoove him to write about how his work fits in these scientific and social-scientific trends); I enjoyed reading this book and think there is a lot of fruitful material here. My point is to assert that his new methods are very much of a piece with much larger trends in the global academy, trends connected with the exponential growth of our data processing power and its applicability across the academy. He may present as a maverick in the literary context, but he looks much less so if one takes a step back and looks at larger academic trends. This can lead to a number of important questions that do not bear on our debates this week (the place of the humanities in the academy, for example), so I will leave this with just one more observation: all of this work may have been published in the New Left Review and in a Verso book, but its heritage suggests that there is nothing necessarily oppositional or subversive about it.
Translation appears here only as the means of diffusion “and hence reformulation from one language into another.” Plot endures and style is changed to suit local conditions. This is clearly not an important issue for Moretti, and this strikes me as a much more serious challenge to business as usual in comparative literature than distant reading will ever be. It is also the build up to his largest historical claim: that there are two distinct world literatures, a pre 18th c. one with a mosaic of local cultures and lots of diversity and a post 18th c. one that shows a “stunning amount of sameness” and a change mechanism based on convergence. This puts a LOT of weight on the bloom of full modernity (in the shape of mercantile capitalism and colonial expansion). It makes sense in the Marxist context because (as we have already observed) it is a fine example of base and superstructure, but a lot of nuance is lost in the process. On p. 127, Moretti asserts that powerful core literatures interfere with peripheral ones but that this is seldom true in reverse. But is it this simple? The single most fertile scene of formal innovation in the novel over the past fifty years has unquestionably been Latin America: could any English-language novelist claim to be as formally innovative or influential as Garcia Marquez, Fuentes, or Roberto Bolano? Could anyone deny their huge influence on writers in the English-speaking world? All this to say that diffusion is always a two-way street.
I think that this complicates his assertion about modern world literature a lot, and I would enjoy discussing this further with the group. The danger of treating literary texts and traditions as data that do not need to be read but aggregated is that it tempts us to think that we are merely compiling data and not always already interpreting it. This is what our scientific colleagues do when they say “the data doesn’t support that” as if data speaks for itself. To cite the title of a very useful book that I teach in my foundation seminars, “raw data is an oxy-moron.” Even in distant reading there is interpretation . . .