The Essay Club comes to an end, as the book club did, with a generous reply from the author. Franco Moretti is the author of Signs Taken for Wonders (1983), The Way of the World (1987), Modern Epic (1995), Atlas of the European Novel 1800-1900 (1998), and Graphs, Maps, Trees (2005) and the chief editor of The Novel (Princeton, 2006). He founded the Center for the Study of the Novel and, with Matt Jockers, the Stanford Literary Lab.
First of all, thanks to all of you for your generosity and engagement. Though I cannot address all your points, I will try to include all controversial and/or interesting issues. Answers 1-3 concern the pamphlet; 4-6 broader issues, that are open or require more thought. And 7, a couple of problems raised by Matt Huculak.
1] Why did I study a single text, rather than a large series?
It was a pragmatic decision: gathering the raw data for the networks proved to be quite time-consuming, and I found myself constrained to a single text. [Actually, I charted a dozen or so, out of curiosity; but they were a haphazard assortment, not a homogeneous series]. Now that Matt Jockers has created an algorithm that [with a minimum of human intervention] extracts a network from any text, the Lab is launching a study of about 250 texts from six different traditions, in which statistical measurements will be much more important. But don’t expect anything until next summer!
2] Why Hamlet?
I wanted a play, for the reasons explained in the pamphlet [dramatic networks are more reliable than novelistic ones]. And since I knew the theory would be new for most readers – more than evolution, or the simple charts I had used in the past – I decided to compensate with a well-known text. Plus, James is right, I thought that if network theory could say something new about Hamlet, this would prove its value for literary study.
3] Did I find anything that surprised me?
Three things: a] Horatio [more on him in answer #4]; b] the curve of characters’ centrality, which proves we must reconceptualize the notion of “character” avoiding, James puts it well, the fallacy of the ‘autonomous individual’; and, c] the findings about symmetry/asymmetry in Our Mutual Friend and The Story of the Stone. Incidentally, the silence [not just yours] on the two novels is bizarre. Ok, not many people know Cao Xueqin – I have read him only a few years ago myself – but guanxi is a very intriguing encounter between networks and plot.
4] Who is Horatio?
For me, he represents the State. For James, the reader. For Alex Woloch [in a discussion at Stanford], the friend. A pretty good case could be made for the confidante. And, I am sure, for even more interpretations.
The keyword, here, is “interpretation”. For better or worse, that page on Horatio and the State is different from the rest of the paper, because it veers away from description and explanation, and into interpretation.
To be clear, I take the distinction between interpretation and explanation from Ricoeur [and partly from Elster]: interpretation turns the words/actions in the text into a new text created by the interpreter [What Horatio really means in the play is…]; explanation establishes a relationship between the words in the text and an already existing external force or structure [for, respectively, causal or functional explanations]. Whence, usually, a greater sense of “solidity” conveyed by explanations over interpretations.
Clearly, my recent work is devoted mostly to explanations. But I think that literary study will always have to use both interpretations and explanations. The question is: how do we establish a relationship between these two activities? Can one validate – or invalidate – the other? How? And so on. For me, this interplay will be a very interesting aspect of future work – much more so than the diatribe between close and distant reading. To which however, having started it, I now dutifully turn.
5] Close/Distant Reading [CR/DR]
When I coined [half in jest, but never mind] the expression “distant reading”, I had to account – just as Lisa says, quite rightly, that we should always do – for my assumptions, methods, choices and so on. CR, on the other hand, is a practice without a theory – or an explicit theory, anyway. Shouldn’t any discussion begin with a little conceptual effort on the part of close readers? Or does their love of literature exempt them from theoretical work?
Maybe that was ‘strident’, as James says, so let me try again. I often hear people assert the need – or even the inevitability – of an interaction between CR and DR. I see the point, but the real problem is not the peaceful coexistence, or cooperation, of the two approaches; it’s the hierarchy between them. For James, CR “should remain for a long time the defining practice [why not theory, I wonder...] of literature departments”. I disagree, and see the basis of literary study in large-scale patterns and conceptual architectures – with individual textual analysis as a more idiosyncratic and peripheral endeavour. [How quickly and how well pedagogical practice may change is a serious but different issue – for another discussion].
That said, I am not asking for a mass conversion to my way of looking at things. DR doesn’t have a monopoly on “large-scale patterns and conceptual architectures” – far from it. Even in other attempts at theoretical investigation, however, CR seems to me to have become mere ballast. Take two excellent books that have nothing to do with DR: Woloch’s The One vs the Manny and Dames’s The Physiology of the Novel. Both open with a couple of theoretical chapters, which are followed by several chapters of close readings of individual novels. Now, how much do the readings actually add to the argument? In my opinion, almost nothing. Is that because Woloch and Dames are bad readers? No: it’s that CR has exhausted its capacity to contribute to knowledge. [I know of only one exception: the work of Roberto Schwarz. Since he is, in my opinion, the greatest critic of our age, it’s a very important exception, and I would like to understand it better.]
Finally. This doesn’t concern your blog, but it’s amazing how many of the vestals of CR, who are so shocked by my callous indifference for the wonders of reading, criticize my “distance reading”, thus revealing, first, that they have never bothered to read the very short piece in which I presented the idea, and, second, that they are illiterate.
6] Why use network theory?
What should we expect from a new theory, and from interdisciplinary work in general? Something new? Sure; but new how, vis-à-vis existing literary theories? A paradigm shift? That’s always exciting; but it presupposes a field trapped in Ptolemaic epicycles. Is that where we are? Some think so: the “Darwinian” collection The Literary Animal, for instance, quotes around seven hundred works, but nothing by Adorno, Auerbach, Bakhtin, Barthes, Benjamin, Benveniste, Brik, Curtius, Eco, Frye, Jakobson, Jameson, Lukacs, Shklovsky, Spitzer, Todorov, Tomashevsky, Tynianov, or Weinrich. They do quote Propp once in a footnote, how grand, but, clearly, for them a century of literary theory is a heap of garbage, and literary study begins with them. Lux facta est.
I have a different position. In general, I believe that what we should expect from an external model is not a Revelation, but a broader, more solid framework for those theories that already work well. Evolution is the greatest theory ever; but I became convinced it would work for literature only when I realized that it was consistent with all the key concepts of Russian Formalism, and a host of solid historical findings from other sources. Evolution was a good framework, because it agreed with what had been found independently of it.
A framework that agrees with previous findings … As Lisa puts it, if network theory hadn’t found Hamlet at the center of Hamlet, it would be in trouble. But this very “agreement” with already established [and/or theorized] data makes the introduction of new theories in an old field slightly paradoxical, as they should both confirm what we know, and yet change it [otherwise, what’s the point]. What form this may concretely take, is something I will return to in detail sometime in the future.
7] Magazine networks
We have no Hamlet character – i.e., no center – writes Matt; so, we have to set it up. No! You have to map the network: if it does have a center, Hamlet will appear; otherwise, you will have a different type of network. In our future study, we actually hope to find networks with two centers, or many, or none at all: if all networks turned out to have the same shape, they would be useless, because they wouldn’t explain the enormous variety of configurations we encounter in dramatic just as in social structures.
“Interaction”, ok; but what about “conflict”? Very good point. I don’t think that this difference – and others: conflict and love, love and cooperation, cooperation and disintegration etc. etc. etc. – can be captured at the abstract level of a network’s system of interactions. For this, vertices and edges are not enough: we will have to use the semantic content of the interactions. How to do so, is one of the things we are trying to figure out.