“We believe that the humanities are unlikely to remain relevant, unless significant changes are made in how professional humanists are trained. Our review of relevant literature and data, both from Stanford and from outside, has given us a good sense of what these changes are. We believe that Stanford — with its educational prominence, culture of innovation and its great human and material resources — should be a leader in driving those changes.”

Why do I quote this passage from a white paper about “The Future of the Humanities Ph.D. at Stanford,” mentioned in a very interesting article on today’s NYTimes, “The Repurposed Ph.D”? Not just because Franco Moretti teaches at Stanford but because I believe that Moretti’s book provides some useful indications – on a theoretical level – about a “repurposing of the humanities,” or at least of their “literary studies” branch.

Indeed, Graphs, Maps and Trees not only points to a different (or divergent) type of textual data mining and modeling, but also to a different way of training doctoral students in literary studies. The kind of brilliant analyses that Moretti performs in his book in order to demonstrate his theoretical point (“a more rational literary history,” in reaction to close reading as “secularized theology” emanating from New Haven) imply a different (and broader) set of interdisciplinary competences, drawing from quantitative history, geography and evolutionary theory. They also potentially envision broader professional applications beyond the promised (but increasingly elusive) destination of tenure-track professorships. Perhaps, as the NYTimes article also suggest, such a repurposing of the (digital) humanities can help form a new generation of culture analysts with a new set of skills which can better prepare them for “jobs within universities but outside the professoriate, like administrator or librarian, as well as nonacademic roles like government-employed historian and museum curator,” and can even help them move across the job market, from higher education to industry, governmental institutions, foundations, etc.

I will comment now on a few points in Moretti’s book that seem particularly relevant to this week’s topic, the intersection of texts and images, but also refer to things we have discussed in previous weeks. Questioning the narrowness of the canon (about 200 novels for 19th-century Britain, even fewer in the Italian), Moretti provides some methodological and theoretical guidelines that are useful for working with digital tools and large scale textual data sets for literary studies:

1. Quantitative research provides a type of data which is ideally independent of interpretations…its limit is that it provides data, not interpretation…(italics mine)
2. Quantitative data demand an interpretation that transcends the quantitative realm.
3. Quantitative data can falsify existing theoretical explanations (for example, interpretive assumptions about the history of the novel).

Moretti distinguishes between graphs, that are not really models, and maps and (especially) evolutionary trees that, instead, are such models, that is “simplified, intuitive versions of a theoretical structure.” (p. 8). We can discuss in class the difference between these forms of visualization and their particular usefulness for our specific research goals. Graphs allow Moretti to formulate and/or falsify hypotheses about the system of novelistic genres as a whole (the life cycles of genres and subgenres). Maps are a good way to prepare an individual text (from the “village stories” series, for example) for analysis: by placing a story in space, the map offers a model of the narrative universe that, compared to other maps in the series, can bring some hidden textual patterns to the surface (“the road from birth to death of a specific chronotope”). The map becomes a diagram. Diagrams look like maps, yet they represent relations not their distribution in space, “a matrix of relations, not a cluster of individual locations.” (GMT, p. 54) Diagramming the use of the free indirect discourse or the use of “clues” in detective fiction allows Moretti to make macro-morphological divergences appear: “this system of differences at the microscopic level [the sentence] adds up to something that is much larger than any individual text, and which in our case is of course the genre – or the tree – of detective fiction” (GMT, p. 76).

As Matt Kirschenbaum writes in his contribution to Reading Graphs, Maps, Trees, a volume which collects critical responses to Moretti’s book and Moretti’s own replies to his critics, “the goal of data-mining (including text-mining) is to produce new knowledge by exposing unanticipated similarities or differences, clustering or dispersal, co-occurrence and trends.” The key word, here, is “unanticipated.” Visualization tools allow us to see/make this new knowledge emerge (graphs, maps and trees “place the literary field literally in front of our eyes – and show how little we still know about it…” – as Moretti effectively puts it).

I think the ambiguity of “seeing/making” touches upon one of the crucial questions we have debated, in relation to most of the tools we have considered: visualizations are (computational) artifacts that may give us the illusion of “discovering” what we are actually “configuring” through our tools. Viceversa, they can help us find unanticipated patterns, and formulate interpretive hypotheses in controlled exploratory experiments with our data models. This, by the way, is what Moretti does best.

In conclusion, distant reading seems several steps removed from close reading: it deals with data models (and forms of visualization, or interfaces) that extract and represent abstract patterns from textual data, “translating the traditional way of formulating critical problems in the humanities into reasoning that can be tested, algorithms that can be run” (Matt Kirschenbaum). Yet, Moretti seems to suggest that a successful and meaningful application of his theory should allow the critic to make discoveries valid also at the microtextual level. Ultimately, if “theories are nets” (as Moretti states, quoting Novalis), we must “evaluate them not as ends in themselves, but for how they concretely change the way we work,” and perhaps also for how they help us humanists find new purposes for our research.