Sunday, January 15, 2006
Moretti Responds (II)
Some more thoughts:
More thanks, and apologies for the slowness of my replies - I am travelling with my 1-year old, with VERY little time of my own. So:
First, a few details. I am less alone than I claim. Excellent. I didn’t even know of the existence of the nora project, which seems very interesting, and I have now several new titles on my reading list, which is great. As for evolutionary approaches, I would be more cautious; some of them - like memetics, or the recent Literary Animal collection - strike me as naïve in their disregard for formal analysis, and passion for fuzzy or crude units of content.
Maps, no, there I am really alone, unfortunately (the idea of a literary atlas began as a collective project, and I ended up working on my own only when the first project was denied funding, and dissolved). Even Bill Benzon’s very interesting post, as he says right away, is on a different type of map (his final question - how is it that the world makes its way into the mind there to be transformed into texts? - is however one of the truly Great Questions in front of us. I wish I knew what to answer).
One caveat for the more optimistic contributors. The field of, loosely speaking, science and literature is full of false starts. The best example of quantitative analysis ever done - Burrows’ multivariate analysis of Austen’s style - was published 20 years ago, and has had, if I’m not mistaken, hardly any effects. (No one has mentioned it in the course of this discussion either). We’re all working uphill, and I’m not sure it’s going to change soon.
Close reading and abstract models, or, interpretation and explanation. Bill Benzon is absolutely right in saying that even in the sciences research is not evenly spread, but clusters around specific issues - the fruit fly is a particularly neat example, because detective fiction is a sort of literary fruit fly (with few and clear variables, easy to manipulate). But this is not close reading, it’s actually much more similar to the “experiments” (on village stories etc.) that I try to do in the book. So, I still think that the strategies I outlined are antithetical to the mainstream of literary criticism. It may be tactically silly for me to say so now, given that the general consensus is that what I do could be interesting, as long as it doesn’t want to get rid of current procedures, but what can I do, this is not a matter of bragging, or of originality (originality, in a book that borrows all its models?!), or of democracy (a hundred flowers, yes, and more) - it’s a matter of logic. Between interpretation (that tends to make a close reading of a single text) and explanation (that works with abstract models on a large groups of texts) I see an antithesis. Not just difference, but an either/or choice.
This said, I understand Eric Hayot’s skepticism about how I formulate the antithesis at the end of the book: I am not happy with it either, and in the response to Prendergast’s article I will try a different take, stressing the causal aspects of explanation, which are usually absent from interpretation. In a formula, interpretation posits a relationship between a meaning and another meaning, and explanation a (causal) relationship between an external force (or constraint) and one or more meanings. And then again, it may well be that the study of literature will always require, or be enriched by, both close reading and abstraction, interpretations and explanations; but this will amount to saying that literature requires two conceptually opposite approaches. Which is odd, and will make for some interesting speculation on why it should be so.
Finally, Tim Burke’s “dumb luck” argument. (My favorite formulation comes from the extinction specialist David Raup: “Bad genes or bad luck?") It is certainly possible that dumb luck plays/played a much larger role than we imagine - probably larger and larger as we move back in history, and cultural products can disappear more easily. I will enlarge on this in my reply to Prendergast, but I may as well admit right away that every time I have studied competition, success, and failure, I have never found that luck played a major role (the only one exception I know: Austen’s novels, despite having soldiers, include no reference to war, unlike most of her contemporaries, and the so-called “Hundred Years’ Peace” that begun in 1815 rewarded her enormously for her choice: sheer external luck changing the world, and hence the expectations of generations of readers). (Needless to say, this is not the only ingredient in Austen’s long-term success). But I admire Gould too, and would have liked to find instances of dumb luck. If anyone comes up with convincing examples, I’d love to see them.
And I do need to add, I don’t ignore book history (it’s the starting point of “Graphs”, and about 50% of my last book), nor punctuated equilibria (which I first applied to literary history 20 years ago, and have used in the three books I have written since).
Thanks again to all of you for the ideas and the challenges.
“Explanation (that works with abstract models on a large groups of texts" strikes me as rather thin science, and only moderately relevant to the reasons why people are interested in literature. You could do a similiar study, of equal scientific interest, on the history of matchbooks or of shoes, and it would all be fine and would probably tell you a small amount about literature, but if you’re talking either/or, forget it. I was a tentative pluralist to begin with, but you don’t seem to want that, and the whole thing seems like a late repeat of the various past attempts to Put [name of study here] On A Really Scientific Basis For The First Time.
Austen’s novels contain plenty of references to war. From Mansfield Park:
Young as he was, William had already seen a great deal. He had been in the Mediterranean - in the West Indies - in the Mediterranean again - had been often taken on shore by the favour of his Captain, and int he course of seven years had known every variety of danger, which sea and war together could offer. With such means in his power he had a right to be listened to; and though Mrs. Norris could fidget about the room, and disturb every body in quest of two needlefulls of thread or a second hand shirt button in the midst of her nephew’s account of a shipwreck or an engagement, every body else was attentive; and even Lady Bertram could not hear of such horrors unmoved, or without sometimes lifting her eyes from her work to say, “Dear me! how disagreeable. -I wonder any body can ever go to sea.”
The Austenian method is to wedge in the “horrors” of war between domestic inanities or worse. The reader can focus on the sharply defined fidgeting in the foreground or consider the undercurrent. (Harder to ignore the allusions to war in Persuasion, since they’re essential to the plot, but then the shortsighted and self-interested people in that novel are correspondingly more offensive.) Perhaps the nineteenth century readers who established the view of Austen you allude to tended to concentrate on Mrs Norris & Lady Bertram and read Austen as provincial in scope and concerns, mentally adapting what she actually wrote to conform to their own tastes and preferences.
What a pleasure to come across this discussion. I am an evolutionary biologist, so quite a few things here have struck a chord.
I have not yet read through all the material here, but am looking forward to doing so. Literary folks might be interested to know that a few ideas have percolated from literary analysis into science, as well as the other way around. Three papers of my own have attempted to connect the techniques of historical representation across disciplines. They are available online (as HTML, with a pdf version containing all the illustrations linked from each page):
“Telling the tree: narrative representation and the study of evolutionary history”
“Trees of history in systematics and philology”
“Mapping the space of time: temporal representation in the historical sciences”
The interdisciplinary use of genealogical trees (sensu lato) has been one of my particular interests, and a lengthy bibliography (to 1993) is also available at:
Looking forward to reading more.
Alright everyone, I’ll see you guys in a couple of months. Our new friend Bob O’Hara has landed a crushing blow to my social life. How long does it take to read over 3,000 archived messages anyway? (I’m only half-unserious here, people.)
What you need, Scott, is to load those archives into the nora software so you can get a better handle on what’s in them.
I’ve been blitzing my way through Bob O’Hara’s “Trees of history in systematics and philology” and recommend it.
Professor Moretti, I certainly wasn’t suggesting that you ignore book history tout court, but rather that there are numerous book historians who, like you, produce graphs and count books in a way that reflects a sophisticated understanding of genre issues.
>What you need, Scott, is to load those archives into the nora software so you can get a better handle on what’s in them.
Classic data mining application. Here’s a fun example:
>I am less alone than I claim
A few URLs for Professor Moretti and anyone else who’s interested:
Let me say a bit more on dumb luck, which for me is related to thinking about Moretti’s trees in terms of emergent systems.
The software program NetLogo is a nice way to see emergence in action. To cite one of my favorite examples, the simulation “Termites” is built around a set of agents who have a small number of simple rules. Each time-step they move one space. In the environment there are also “sticks”, randomly placed. If a termite-agent ends its move proximate to a stick, it picks it up. If a termite-agent carrying a stick ends up next to another termite-agent carrying a stick, they drop their sticks and move a space. There’s another rule about jumping a large number of spaces if a termite is surrounded by sticks that’s no so important.
Anyway, the consequence is that in a simulation that begins with a random distribution of termites and sticks, they will in time invariably build a single large roughly circular pile and that once built this structure persists.
The analogy for me is with genre. I agree you can talk about the formation of genres as being anything but “dumb luck”, as being the product of an evolving process of divergence and convergence in relationship to some external social “fitness landscape”. The genre is the circular pile: at a certain moment, given certain conditions and a process of textual evolution, the genre is going to come into being.
But in the termite simulation, what is more or less random is where exactly on the grid the pile appears. That is perhaps a difference that doesn’t make such a difference if all you’re interested in is the structure as a whole, or the genre in its generality. But the specific texts that give genre its peculiar shape, the tropes and plot elements and template, strike me as analogous to the location of the termite mound: as a difference which has a strong element of the random to it.
I’ll give a science-fiction example. I think you could say that a subgenre of SF concerned with the representation of online computer use, hacking, and so on, was more or less an inevitability. But what was not inevitable was some of the precise imaginative landscape of that subgenre: “cyberspace”, “jacking in”, and so on, much of which came out of Gibson’s Neuromancer. One kind of cultural history or literary analysis would insist that Gibson’s impact on the genre was a result of individual authorial skill or vision; another more historicist argument would say that it had something to do with a larger sociohistorical predicate that made metaphors of space and frontier the most powerful way for Americans to map novel experiences of technology. I’d suggest that Moretti’s approach, and similar approaches to genre, open up another possibility: that the subgenre was the product of an evolutionary process, of cycles of divergence and convergence. But I’d go another step, and say that the first text to “hit paydirt” and provide a template for a genre that is about to come into being may be chosen with an element of randomness about it. Partly I’d say this because in any given claim about a genre’s beginnings, you can almost always find a text that preceded the origin but was unnoticed or unread and texts that followed that proposed radically different metaphors or tropes and were ignored.
“All criticism is doomed to analyze only its own perceptions. What the critic is finally saying is, this is my fantasy when faced with the work. What is essential for the interpreter is an ethics of modesty: that he not consider his own perception is the only one.”
<div style="text-align:right;">- “Edgar Allen Poe” [Benjamin Friedlander] quoting Stephen Rodefer quoting Julia Kristeva</div>
Less tentatively than John Emerson, I’m a pluralist. (As a consumer, anyway; as a producer, I’m merely an amateur belletrist.) I enjoy much of the present-day interdisciplinary work which might be classified as “materialist” or “quantitative”. And I understand that everyone who works in the humanities considers themselves a member of a beleagured minority, no matter how successful they might seem to the outside world. But as a pluralist, like John, I feel some need to push back against an either/or tone.
The illusion of rising above one’s own time is thrilling, whether the buoyant gas be eternal verities, the universal language of music, or scientific analysis. But history sinks us all eventually, and, despite the earnest (and often well-rewarded) effort put into social and psychiatric sciences, they have a short shelf life. If not many people read Poe’s or Twain’s or James’s criticism nowadays, even fewer build on studies of racial traits and phrenology and hysteria. To paraphrase Prof. Arthur Lee, the social science of today will be the symptom of tomorrow.
“I may as well admit right away that every time I have studied competition, success, and failure, I have never found that luck played a major role...."
That’s hardly surprising, since we’re not working in controlled laboratory conditions. Explanations can be invented for any event. The problem is proving them. (Disproving proposed explanations, on the hand, is often both doable and worthwhile.) Histories which preclude dumb luck and the strictly unpredictable influence of individuals tend towards the Whiggish or Spenglerian. The excluded arbitrary returns in the form of the historian’s decisions.
“And then again, it may well be that the study of literature will always require, or be enriched by, both close reading and abstraction, interpretations and explanations; but this will amount to saying that literature requires two conceptually opposite approaches. Which is odd, and will make for some interesting speculation on why it should be so."
As Laura Carroll indicates, “interpretation” is hard to shake free of. For example, Adam Roberts’s pre-1900 “science fiction” wouldn’t have been analyzable as such until “mainstream realism” became well-enough established to make those isolated works seem like part of a rival tradition. Even choosing what elements to trace through which representatives of a “genre” is an interpretive act, which is why I called the Strand spectrum analysis more convincing than the “Conan Doyle’s rivals” tree.
Insofar as literature can be said to exist at all, yes, I believe it calls forth both “interpretation” and “explanation”—that is, a desire to narrate the work’s intended meaning and a desire to narrate the work’s origins. (Already that seems not so binary a opposition.) Or it calls forth a desire to describe the work itself, or to describe our reaction to the work, or, more generally, to respond. Trying to pursue all of these approaches at once would, I agree, result in an unholy mess and a shattered career. But that doesn’t mean that they’re unambiguously opposed.
Let’s return to the celebrated clue of the drugged coffee. I haven’t read the source story, and so I don’t know what its effect in context would be for me, much less for a reader of the time. But I’ve seen the Continental Op and Philip Marlowe walk more or less knowingly into traps to “shake things up.” The detective mystery, broadly defined, shows someone gathering knowledge about criminal circumstances and bringing about an expository conclusion. Presumably drawing from the nearly autistic Holmes, the polite puzzle mystery depicts detached knowledge (represented by the handling of clues, maps, and timetables) and emotional disengagement from the (emotionally arbitrary) conclusion. Hard-boiled mysteries and psychological thrillers, on the other hand, show knowledge as physical experience and insist that observation is inseparable from influence. One might, Moretti-like, try tracing the split by tracking the amount of physical and emotional damage sustained by the detective—beatings in one column, druggings in another—but there would eventually be convergences: Hammett’s “drawing room” mystery (set among the detective’s dysfunctional friends, with the detective shielded by almost pathological smugness and with a viciously undercut solution); the “Ellery Queen” team’s late decision to make Queen a fallible bringer of injustice; Sayers’s turn from Wimsey toward Harriet Vane....
There’s not necessarily a conflict between intellectual-emotional engagement and the deployment of evidence. Much of the quantitative, historical, comparative, and materialist research that’s been mentioned results in richer or refreshed readings. In that list, I’d include what I’ve read of Moretti’s previous work; on early acquaintance, he sounds like a Marxist Hugh Kenner: a congenial combination. For whatever conscious or unconscious reasons, he seems to have evolved in the direction of Philo Vance, insisting on the need to keep the reading experience (whether current-day or historical) at arm’s length. Me, I’m a Black Mask fan. Literature, like murder, is no place for clean hands.
Timothy Burke: “But I’d go another step, and say that the first text to “hit paydirt” and provide a template for a genre that is about to come into being may be chosen with an element of randomness about it.”
I’m not sure that I agree. Imagine grouping the SF authors who are commonly thought to have started subgenres or common SF tropes and looking for similarities among them. (Let’s say: Verne, Wells, E.R. Burroughs, H.P. Lovecraft, E.E. “Doc” Smith, Wlliam Gibson.) I don’t have much time at the moment, so I won’t try to categorize the similarities that I see, but I’d say that these authors have an unusual ability to create or transmit an archetype, with the impression of it having its full energy but without much in the way of overly skilled writing, complex characterization, or even the SF author’s putative preoccupation with science, etc. to distract from it. (They may think that they’re actuated by science, and often you can find a short story or minor novel or something that uses science as something more than a mythopoetic trigger, but that’s not what was influential about their fiction.)
So even though other, more or less forgotten, authors may have supposedly started their various subgenres, an author with their particular type of skills was required in order to take something that may have been in the background and bring it out. In that sense I disagree that the choice is random.
I’d also like to thank Bob O’Hara for his opportune arrival. Besides fitting nicely into the discussion of Moretti’s “Trees”, his “Telling the Tree" is helping me with a post about Brooks and Henry Adams’ attempts to “scientize” history.
I’m sympathetic to Moretti’s take on explanation and interpretation. I’ve been arguing a similar point in various recent posts here, here, and here at The Valve. Rather than continue with an argument, let me toss out an analogy for consideration: medical diagnosis.
Diagnosis obviously involves quite a bit of interpretation. The patient exhibits signs with the physical must interpret. On the basis of that interpretation therapeutic intervention will be planned and executed. If all goes well, the interventions will have the predicted effects. There are, of course, situations where no regime of interventions yields a favorable prognosis.
Diagnosis can involve a wide variety of things. The patient is asked about her condition. Some observations are made with the “naked eye.” Others involve relatively simple instruments – thermometer, stethoscope. Others involve very complicated instruments – CAT scans, EKGs – and still others involve laboratory tests. All of these observations require interpretation. And the whole process is suffused with a large and complex body of explanatory theory about the workings of the body and of disease processes.
In what sense is a literary text like a patient? What is there to be examined? Do we have any bodies of explanatory knowledge comparable to those available to contemporary medicine?
I rather suspect the analogy is loaded in some way. And that I am strongly biased in these matters. But I’ve not explored the analogy enough to have a sense of these matters.
I’m not sure what I think of the “dumb luck” factor. I’ve read a bit in the complexity literature and, as theory and model, I rather like it. But it’s not clear to me what it means for literary history or, more generally, cultural history. Let me throw out two examples.
Sometime early in the 20th century manga emerged in Japan, mostly as a form of graphic novels for children. It really took off after WWII to the point where manga now constitutes roughly 40% of the print output in Japan and exists in many forms - including non-fiction - and is read by Japanese of all ages. The most important single figure in this post-War florescence is Osamu Tezuka, best known for Astro Boy (Tetsuwan Atom), a series which began in 1951 and then continued for 17 years (and was developed into an anime series first in the 1960s, then later in the 1980s). But Astro Boy was not his first success. Before that Tezuka had published including New Treasure Island (1947, over 400,000 copies), Lost World (1948) and Metropolis (1949). Metropolis appears to be a predecessor to Astro Boy, both are about robots looking for parents. Here’s what Tezuka says about the origins of Metropolis:
This manmade person was based on the image of the female robot in the famous pre-war German film Metropolis. That said, I hadn’t seen the movie at the time and I didn’t even know what it was about. During the war, in Kinema Junpou, or some other such magazine, there was a single still from the movie of the female robot’s birth scene. I remembered it and it just gave me a hint. I also really liked the sound of the word metropolis so I used the title.
That hint, it seems to me, is our “dumb luck” factor in this case. Tezuka saw an image and read a word, worked them over in his mind, and several years later produced a hit manga.
What’s important about this particular hit is that it involved themes that were to become quite important in manga culture. The central character, Michi, is a robot and, not only that, but a robot who is looking for a place in the social order - which is important in the subsequent development of manga. Moreover, Michi is both male and female, thus exhibiting a gender “problematic” that will become very important in manga.
Just how that initial hint became developed into a story, that we can fairly attribute to Tezuka himself. But why that particular story, and others, became a hit, that requires a different kind of explanation. That explanation, it seems to me, will involve the imaginative needs of the Japanese population, starting with children, during the post-War era.
To a jazz musician, the term “rhythm changes” designates a certain harmonic structure, namely that of Gershwin’s “I Got Rhythm.” There are a large number of tunes in the jazz repertoire that are based on that harmonic structure. Other than the blues (“blues changes”) there’s nothing else like it in the jazz repertoire. Jazz musicians routinely composed (performed and even published) their own melodies based on the harmonic patterns of popular songs; thus “Ornithology” is based on “How High the Moon,” “Donna Lee” is based on “Back Home in Indiana,” “Bright Mississippi” is based on “Sweet Georgia Brown,” and so forth. But there are dozens, if not hundreds, of tunes based on rhythm changes. And Gershwin’s tune is the only one that’s given rise to a phrase (“rhythm changes”) designating it’s characteristic functional structure. Why?
One approach to an answer would look to the intrinsic properties of the tune. The structure is build of a few simple elements yet it “covers” a wide range of “tonal space.” That characteristic - and others - makes it an attractive vehicle for improvisation. One might also look to the historical context. The tune was in a Gershwin musical that appeared on Broadway in 1930, Girl Crazy. Gershwin was a well-established figure by that time; jazz was well-established in the popular psyche and the Swing Era was about to explode; and music publishers where looking for properties to market. All this worked to Gershwin’s advantage. (See Richard Crawford, George Gershwin’s “I Got Rhythm” (1930), The American Musical Landscape, U of Cal Press, 1993, 213-136).
Beyond this we might look to the competitive ethos of jazz, especially during the 1940s and after. If musicians are going to compete in virtuosity, there needs to be a vehicle for that competition. “Rhythm Changes” certainly fills that bill. But some other suitable tune might have done just as well had it come along at the opportune moment. Along these lines it is worth noting that Charlie Parker loved rhythm changes and composed many tune over them. During the mid-40s into the early 50s he was the iconic jazz soloist; though, by that time, jazz had largely fallen out of public favor.
What, if anything, do these examples have to do with recent work on emergence? I don’t really know. But it seems to me that, in both cases, an amplifying context is important. Perhaps what these various computer simulations are about is setting up an amplifying context, a dynamic system where certain patterns can emerge and grow to the point where they dominate the system.