Saturday, February 01, 2014

Graham Harman, Kenneth Goldsmith, & Franco Moretti Walk Into a Bar: A Future for Literary Studies

Let's throw some big words onto the table: speculative realism, object-oriented philosophy, the digital humanities, and conceptualist poetics.  They all belong there together, I think, because they all have something to do with one possible future for literary studies.  Graham Harman is the leading figure of the movement known in philosophy as speculative realism or object-oriented philosophy.  While his name seems to strike fear into the hearts of the more uptight members of the Anglo-American philosophical community, he remains relatively unknown among those concerned with literary study, even in those circles where philosophically-informed critical theory holds sway.  All that may change, though: in the past few years he seems to have become more focused on the interdisciplinary possibilities of his thinking, and he's begun to speculate about just what his brand of thinking could mean for the future of literary study.  It's a weird prospect he presents, but an intriguing one.  What is more, it's a future that could easily build on two growing developments in thinking about literature: digital humanistic study of the sort practiced by Franco Moretti at Stanford's Literary Lab; and conceptualist poetics as practiced most prominently by Kenneth Goldsmith.

Speculative Realism

Harman tells us that speculative realism is "weird realism" (he likes the phrase so much it appears as part of the title of his book Weird Realism: Lovecraft and Philosophy).  While we literary types might want to think of "realism" as a nineteenth century literary movement, Harman uses it in a sense specific to philosophy: as one term in the realist/idealist debate.  Realists, in this context, are those who believe in an objective, exterior world of things-in-themselves, existing independently of our minds; while idealists like Kant believe that the only world we can know is the world of things present to our mind, the world of mental phenomena.  In the long history of realists vs. idealists, realists have tended to be the commonsensical bunch, those who seem least weird and most down to earth.  One thinks of Boswell's anecdote about Samuel Johnson responding to Berkeley's idealist speculations, in which the rebuttal to idealism could not have been more concrete:
After we came out of the church, we stood talking for some time together of Bishop Berkeley's ingenious sophistry to prove the nonexistence of matter, and that every thing in the universe is merely ideal. I observed, that though we are satisfied his doctrine is not true, it is impossible to refute it. I never shall forget the alacrity with which Johnson answered, striking his foot with mighty force against a large stone, till he rebounded from it -- "I refute it thus."
Vulgar empiricism! It worked for Johnson, but for Harman it just won't do.  He comes by his realism from another source: Heidegger, specifically the tool theory sections of Being and Time.

Most of us in literary studies who cut our teeth on continental literary theory have absorbed, to one degree or another, the sense of reality as socially constructed, as the product of discourse or language, or as constituted in social systems of one sort or another.  Whether our maître à penser was Foucault, Derrida, Deleuze, Bourdieu, or, if we're very hip indeed, Bruno Latour, we've tended to believe in objects as formed by human consciousness or consciousness' products (such as language or systems of prestige).  Harman believes, instead, that objects necessarily exceed their relationship to humanity, and even their relationship to each other.  As he put it in an interview for the Cultural Technologies podcast, his theory centers on "the notion that the objects of the world are not exhausted by their interactions, that there is some nucleus in the object that is never fully deployed in its relations."  This grew out of Heidegger, especially those parts of Heidegger's thinking that rebel against the phenomenological philosophy of his mentor Husserl, who was interested in the way objects appear in human consciousness.  For Husserl, objects were important in terms of their appearance to us, but Heidegger saw that things exist, for the most part, without our consciousness.  We depend on them as tools or equipment, but even as we're doing this they're not something we think about, or are conscious of.  Things, for Heidegger, are mostly not in our consciousness, but are unconsciously relied on and taken for granted.  Things don't tend to appear in consciousness until something goes wrong with them: you don't think about your computer keyboard until the space bar sticks; you don't think about the hammer you're using until it breaks (the latter is Heidegger's famous example in Being and Time—you might recall something similar being said about shoes his "Origin of the Work of Art").

Harman goes a step further than Heidegger: for him, it's not just that objects go deeper than our consciousness because of being used without being present in the mind: the object goes deeper even than its use.  The existence of the hammer exceeds both consciousness and use, and withdraws from us in ways upon which we can only speculate.  It can't be summed up by how we think about it or use it.  Neither can it be explained as a bundle of its components.  As Harman puts it,
Just as humans do not dissolve into their parents or children but rather have a certain autonomy from both, so too a rock is neither downwardly reducible to quarks and electrons nor upwardly reducible to its role in stoning the Interior Ministry.  The rock has rock properties not found in its tiny components, and also has rock properties not exhausted by its uses.  The rock is not affected when a few of its protons are destroyed by cosmic rays, and by the same token it is never exhaustively deployed in its current use or in all its possible uses.  The rock does not exist because it can be used, but can be used because it exists…. It is a real form outside our minds.  It is what medieval philosophers called a substantial form: the reality of an individual object over and above its matter, and under and beneath its apprehension by the mind.
One might say that Harman's philosophy is concerned with essences, in that it is looking for what makes a thing itself, in excess of its uses or relations or components.  "The term 'essence' gets a bad press these days," Harmon once told an interviewer for Design and Culture, "because it has come to be associated with all kinds of repressive and reactionary dogmas, but if we take 'essence' in a more minimalistic sense to mean 'what a thing is quite apart from its accidental circumstances,' then a certain essentialism is unavoidable."  Moreover, we're not going to get access to those essences, or even look for them.  We're going to "look instead at how individual entities… withdraw" from systems of definition or use or relation.

Essentialism and Weird Realism

Maybe the point about essentialism and its putative relation to reactionary views is worth elaborating on.  Harman certainly does elaborate on it—here's a passage from his article "The Well-Wrought Broken Hammer":
According to one familiar narrative… philosophers used to be naïve realists who believed in real things outside their social or linguistic contexts; these things were ascribed timeless essences that were not politically innocent, since they subjugated various groups by pigeonholing each of them as oriental, feminine, pre-Enlightenment or some other such tag.  According to this view, we have luckily come to realize that essences must be replaced with events and performances, that the notion of a reality that is not a reality for someone is dubious, that flux is prior to stasis, that things must be seen as differences rather than solid units…
The thing about this narrative, though, is that it takes one view of objects and their essences as the only possible view of them:
The problem with individual substances was never that they were autonomous or individual, but that they were wrongly conceived as eternal, unchanging, simple, or directly accessible by certain privileged observers.  By contrast, the objects of object-oriented philosophy are mortal, ever-changing, built from swarms of subcomponents, and accessible only through oblique allusion.  This is not the oft-lamented 'naïve realism' of oppressive and benighted patriarchs, but a weird realism in which real individual objects resist all forms of causal or cognitive mastery.
An emphasis on the question of what makes an object itself is not, then, a sure sign of reaction: this is not your father's essentialism.

Object-Oriented Literary Study

What could all of this have to do with literary study?  Well, for starters, its one more nail in the coffin of one of the old dogmas of the New Criticism—that every part of the poem exerts pressure on every other part, and that no part is extractable without utterly changing the poem's meaning and effect (this is really a more of a refutation of Cleanth Brooks' "Irony as a Principle of Structure" than of New Criticism as a whole, which was wider and more various than most of its current detractors and advocates seems to believe).  Here's Harman discussing the famous ending of Keats' "Ode on a Grecian Urn," with the assertion (the urn's, not Keats') that "beauty is truth, truth beauty."
If Keats' 'beauty is truth, truth beauty' can only adequately be read as the outcome of the earlier part of the poem, this is not true of the whole of the earlier portions, Cleanth Brooks notwithstanding.  We can add alternate spellings or even misspellings to scattered words earlier in the text, without changing the feeling of the climax.  We can change punctuations slightly, and even change the exact words of a certain number of lines before 'beauty is truth, truth beauty' begins to take on different overtones.  In short, we cannot identify the literary work with the exact current form it happens to have.
What is essential to the poem remaining itself is not immediately clear, but the insistence on total integrity doesn't hold up, as far as Harman is concerned, no more than the sense that a few protons of an individual rock being destroyed would destroy its identity as a rock.  What we can't do is determine the essence of the particular literary text: "the literary text," writes Harman, "runs deeper than any coherent meaning, and outruns the intentions of the author and reader alike."  What, then, can we do?  Here's where things get interesting—because we can try, through various methods of indirection, to allude to what makes Keats' ode itself.  We won't get to the core of the thing, but we can begin to understand the nature of what makes that poem that poem, and we can understand something about what would mark the limits of the poem, when we would find that it is no longer itself.  The critic of a text or set of texts would go about this sort of thing by
...attempting various modifications of these texts and seeing what happens.  Instead of just writing about Moby-Dick, why not try shortening it to various degrees in order to discover the point at which it ceases to sound like Moby-Dick?  Why not imagine it lengthened even further, or told by a third-person narrator rather than by Ishmael, or involving a cruise in the opposite direction around the globe?  Why not consider a scenario under which Pride and Prejudice were set in upscale Parisian neighborhoods rather than rural England—could such a text plausibly still be Pride and Prejudice?  Why not imagine that a letter by Shelley was actually written by Nietzsche, and consider the resulting consequences and lack of consequences? … all the preceding suggestions involve ways of decontextualizing works… showing that they are to some extent autonomous even of their own properties.  Moby-Dick differs from its own exact length and its own modifiable plot details, and is a certain je ne sais quoi or substance able to survive certain modifications and not others.
It's a different world than what we normally think of as the province of the literary critic, isn't it?  But it's also a world that has to some extent already come into being, independently of Harman's recommendation.

Goldsmith, Moretti, and a Future for Literary Study

We've seen decontextualizations very much along the lines suggested by Harman— but not in the main works of literary critics working in any of our conventional modes (new historicist, formalist, Marxian, feminist, poststructuralist, etc.).  We've seen it much more consistently in the works of conceptualist poets like Kenneth Goldsmith.  Consider Goldsmith's transcription works: his typing, verbatim, of a year's worth of weather reports in his 2005 book The Weather, or his transcription, again verbatim, of the September 1, 2000 issue of The New York Times in the 2003 book Day.  These projects raise many questions (including questions about the meaning of manual transcription in the age of mechanical reproduction, about the meaning of authorship and the importance or unimportance of originality, and about whether a poem needs to be read or simply acknowledged as a concept).  But Goldsmith's conceptualist projects also raise exactly the kind of question Harman raises with " Why not imagine that a letter by Shelley was actually written by Nietzsche, and consider the resulting consequences and lack of consequences?"  Why not imagine that the September 1, 2000 issue of The New York Times was a book written by Kenneth Goldsmith?  What are the consequences or lack of consequences?  Goldsmith has undertaken other projects very much in accord with other elements of Harman's projected future for literary studies.  Consider, for example, the following items from his 2002 list poem "Head Citations":

1. This is the dawning of the age of malaria.
2. Another one fights the dust.
3. Eyeing little girls with padded pants.
4. Teenage spacemen we're all spacemen.
5. A gay pair of guys put up a parking lot.
5.1. It tastes very nice, food of the parking lot.
6. One thing I can tell you is you got to eat cheese.
7. She was a gay stripper.
8. Fly like a negro to the sea.
9. Hey you, get off of my cow.

It is, as you've probably noticed, a list of misheard musical lyrics.  The poem contains some 800 of them, taken from various sources.  The distortion of the original text is not so great that all of the lyrics are unrecognizable, though we've clearly moved beyond the words as written by the songwriters.  The question of how far is too far, of where the limits of the song are to be found, comes to the fore—and gets complicated by the fact that these mishearings are all actual instances of what people have experienced when they've heard the songs.  Once again, Goldsmith's practice anticipates Harman's call for action—it's object-oriented literary criticism avant la lettre.  

Goldsmith remains a controversial figure in poetry circles, and it would take a remarkably progressive English professor to consider what he does a kind of literary criticism.  But what about Franco Moretti?  His lit-crit bona fides are as respectable as one could wish: he's written extensively in the Marxian mode, holds the appropriate degrees, holds an endowed chair at Stanford, and has published a half-dozen books of criticism, well received in many quarters.  He's also at the forefront of the movement for the digital study of the humanities, running Stanford's Literature Lab.  The massive digitization projects done at Stanford under his direction have yielded all sorts of interesting results, but one of the directions Moretti has taken comes tantalizingly close to the project outlined by Harman.  The most fascinating part of Moretti's recent book The Bourgeois examines what many may think of as an unpromising topic: verb forms in Daniel Defoe's Robinson Crusoe.  His analysis of those verb forms leads him to notice that the most typical formation involves Crusoe narrating events in a sequence moving from the past participle to the simple past and on to an infinitive.  "Having stowed my boat," says Crusoe, "I went on to shore," and then concludes with "to look about me" (I am indebted to David Winters for this particular example).  This is important, says Moretti, because it embodies the bourgeois worldview of instrumental reason and deferred gratification: of doing something in order to prepare for a next step that will lead to another step in an ever-proceeding process of building and mastering.  It's a far cry from the verbal structures we find in Homeric epic, where the "Having done that I did that in order to do the new thing" pattern is in relative abeyance.  Gathering this kind of data is important—but monkeying with it could be even more important.  Moretti has the technical resources to make massive substitutions: he could change Crusoe's verb forms around electronically, and produce exactly the kind of modified text Harman describes, allowing us to investigate the boundary at which Robinson Crusoe ceases to be recognizably itself.  How important is that bourgeois verbal structure?  One modification would go unnoticed by even the most committed Defoe expert.  When would we meet the threshold?

One often hears that there's little happening in literary studies lately, that after the theory wars of the last decades of the twentieth century we have become complacent, or given up on innovation in literary criticism.  The model propose by Harman is not as far-fetched as it may seem—in fact, it has already taken root among creative writers and those most committed to bringing technology to the study of literature.  We won't know how valuable it could be until we try it in earnest.  So I say we work to get Harman, Goldsmith, and Moretti installed on three adjacent bar stools.  When they walk out of the joint, they might just be able to point us to where we could be going.


  1. The passage from Harman about "Moby Dick" recalls Borges's "Pierre Menard." The two phrases from Borges that have become touchstones with me are "erroneous attribution" and "deliberate anachronism." The former would be Harman's Shelley/Nietzsche example. The latter, Borges writes, would be reading Homer as a version of Joyce, rather than the other way around. Or in the example that's on my mind right now, reading Coetzee's "Lady Chandos letter" as the source of Hofmannsthal's "Lord Chandos letter" rather than the other way around.

    Does Harman mention Borges at all?

  2. Not that I've seen, but I haven't come close to reading all of Harman's work. Let's invite Borges to the bar with the others, though!

  3. Excellent blog post and one that will be very helpful for me as I finish up a book on "weird reading" influenced by Harman's work; you might be interested to read this issue of Speculations, also:

  4. Thanks, Eileen! I'll look it up today (while everyone else is watching the Superbowl). Would also be interested to see the book when you're ready to share the manuscript. I'm thinking someone needs to put together a conference panel on "weird reading," since the speculative realism ball is just starting to roll in the literary field. If you're interested, let me know at archambeau [at] lakeforest [dot] edu.

  5. The thing about Harman's counter factual criticism, as I initially pointed out in this post and then in this one, in a different mode, is that it pretty much collapses onto the actual process of literary culture. We already have abridged versions of Moby Dick and of many other texts of varying literary quality and importance. Reader's Digest made a big business out of cranking our abridgments of texts of all kinds. I don't know whether his Pride and Prejudice variation has been tried, but Francis ForD Coppola took Conrad's Heart of Darkness and set it in Vietnam and Cambodia in Apocalypse Now (and Apocalypse Now Redux) and there's a Japanese manga version of Heart of Darkness that owes debts to both Conrad and Coppola.

    It's not clear to me that Harman's suggestion is anything more than an empty thought experiment. One can play with it in a Borgesian way, but I don't see that it has the capacity to tell us anything new or interesting about how literature works.

    1. I see what you mean -- I suppose that, like any method, it can be applied in ways that give us utterly banal results. I actually think that versions of texts (graphic novel or film adaptations or abridgments or what have you) make for interesting test cases of the identity of a narrative—at what point is it not itself? But I'm interested to give your posts a read: thanks for the links!

    2. By the way -- are you the Bill Benzon who wrote "Beethoven's Anvil"? Read that some years ago in a copy I checked out from a library. Just added it to my Amazon wish list, since there are things in it I'd like to read again.

    3. Yes, I'm THAT Bill Benzon.

  6. At what point is it not itself is pretty soon after it gets monkeyed with - it becomes the acknowledged abridgement, or the lightly edited plagiarism, or the evidence in the linguistic experiment with connotations of syntax, or further down the line the film script or the musical or the fan fiction. The thought experiment of Nietzsche writing Shelley doesn't make him write Shelley of course, though the thought experiment might lead to some conclusions about how they differ as authors. But they are authors and they wrote what they wrote unless proved otherwise.

    1. What do we make, though, of things like Mary Shelley's Frankenstein, where there are versions with or without the suggestions made my P.B. Shelley (mostly diction changes) -- Mary Shelley approved them and they were part of the first publication, though not of the first manuscript, and both versions are still on the market, published, read, and accepted as Frankenstein? Is it Mary's approval that makes both of them still Frankenstein? Is one of them more 'the real one' than another? What about things that come to us out of the oral tradition in various versions? I mean, I can understand not getting entirely on board with the death of the author, but aren't there cases where textual variations don't obliterate the identity of the work? I think it's in finding the limits of where the work is no longer itself that we'll find something interesting...

    2. One might also think of stagings of plays -- when does MacBeth cease to be MacBeth, in radical stagings? And then there's the question of readings aloud of poetry, especially when different accents make rhymes full or slant -- at what point does sounding matter? What features don't count as changing a text? I mean, it all gets pretty complicated pretty quickly, I think. And I like that.

  7. It's only a matter of definition in terms of giving the thing the snappy title of "Frankenstein" or "Macbeth" without explaining the differences at length. A performance is a performance and exists in time in a different way from a text, and a variant edition is a variant edition and ought to be advertised as such to anyone who's going to know the difference, but mostly it's convenient enough to give it the recognisable title. The business does get complicated, but only if all the multitudes it contains have to be juggled with at once - and most of the time, they don't.

    1. I suppose where we differ is in that term "only."

    2. Ah yes, your job is to juggle them all into a unified field, and mine is to take things as they come. The luxury of having dodged the academic life.

    3. Walking around the grocery store I was kind of lost in questions about when a text ceases being itself. Auden's revisions were extensive -- is it the same poem? I what sense? I mean, I get that a librarian will label them as different versions and be done with it. But I don't find the question "where do we draw the line between a text being itself and not being itself?" irritating. It's easy enough to say "the thing the author wrote is F. Scott Fitzgerald's The Great Gatsby," but so much of what we admire in that novel was put in by Max Perkins -- would it still be the same novel if we reverted to FSF's text?. Change forty words in Tess of the D'Urbervilles and I doubt I'll notice -- it'll be the same novel to me, even though I read it twice a year. And so on. Perhaps you find this more banal or ridiculous than I do, but I'm sort of used to people thinking things I care about are ridiculous.

    4. Even when one is qualifying things by saying "these are VERSIONS of the novel," we presuppose the notion of the novel, that it exists as an essential entity with various accidental variants. But what is of the essence? Well, perhaps this does seem academic. But so am I, for better or for worse.

    5. It isn't banal or ridiculous, but most of the questions are practical rather than philosophical. Great Expectations has two endings which make a big difference. I'm glad I got to know it in an edition that gave me both. That's worth knowing about, as a specific version. "Director's cut" versions are usually pretty well labelled as such, other kinds of edition often not. The effect of the missing forty words in Tess might depend on which words they are, but the main thing is that you've been short-changed on the goods, whether you noticed or not. It's "the goods as advertised" that are the thing we're discussing, surely? In which case the subject is really the advertising more than the goods. With the Kenny Goldsmith version of the New York Times, it's his publicity that's the whole point (cf Leavis on the Sitwells belonging to the history of publicity rather than poetry). As long as he transcribed it accurately (and frankly who cares), the information is the same. Does anyone get an ontological panic when they see two identical copies of Tess in a bookshop? No, they pick up one and ignore the other because it's "the same book" even when it isn't the same object. So far I haven't seen anyone offer that as an example of such a problem of definition: it's when one of the copies has a page missing you go back to the shop and they check whether they were all like that.

    6. Fair enough -- the first thing I think when I'm hit with a snowball isn't "did the impact modify the snowball to the point where it is no longer the snowball," it's "which member of the anthropology department threw that thing." One can get through life without ontology. But one can also get through life without Chopin.

    7. Even when one is qualifying things by saying "these are VERSIONS of the novel," we presuppose the notion of the novel, that it exists as an essential entity with various accidental variants.

      I wonder. Under the pressure of Harman's observations it may seem that way, but it's not clear to me that that's what's going on. Maybe we've just got informal conventions for talking about sets of closely related texts held together by acts of authorship but we're not necessarily presupposing any strong notion of essence holding them together.

      * * *

      What happens to these questions if we pose them in the context of story-telling in a pre-literate culture? I don't mean any and all story-telling, but the telling of culturally important stories, myths?

      The author as such disappears. The stories are traditional, passed down from one generation to the next. The storyteller is (just) a performer, not an author. So the Max Perkins/F Scott Fitzgerald problem disappears.

      Each performance of the story is different. Whatever the identity criteria are for a given myth, they don't depend on the exact wording. Presumbably they have to do with required characters and incidents and perhaps an obligatory phrase here and there. And it IS a performance. The teller is in the presence of an audience and presumably is responsive to that audience.

      Some years ago there was an interest in "story grammars" in several closely related disciplines. The idea was to specify formal requirements that make a given discourse into a proper story. As far as I know, this enterprise collapsed. But perhaps something like that is what's a stake in worrying about the identity conditions of a story (or a poem).

  8. "I think it's in finding the limits of where the work is no longer itself that we'll find something interesting..."

    But who makes that determination? What are the criteria?

    At the moment I'm working and blogging about Gojira/Godzilla, King of the Monsters. Gojira is a Japanese film that came out in Japan in late 1954. Godzilla, King of the Monsters, is an Americanized version of the film that came out in the USA in 1956. It's significantly different from that Japanese film. Roughly 30 minutes (out of 95) in the Japanese film was scrapped and replaced by 20 minutes shot for the American version. The new footage centers on a new character, an American reporter named Steve Martin, who tells the story. The Japanese film has a romance plot involving a conflict between arranged marriage and marriage by personal choice; that was almost completely eliminated from the American version.

    Are they the same film? I can't say that the answer much matters. What's interesting are the differences between the two films. I assume that romance plot was weakened considerably in the American version because arranged marriage was (and is) not a live issue for an American audience, while it was relevant to a mid-20th-century Japanese audience (and still turns up in manga and anime). And I figure that the American reporter was added so as to give an American audience a more comfortable point of entry into the film, an on-screen American they can identify with.

    What's really interesting is THE crucial decision in the stories, the one that allowed the creation of a device capable of killing the monster. In the Japanese version that decision was made by Emiko Yamene, who knew about the secret research of her childhood betrothed, Dr. Serizawa. Her decision to communicate that knowledge to others more or less forced Serizawa to build the weapon (and to commit suicide). In the American version Steve Martin played a role in prodding Emiko into making that decision. Is that a merely contingent fact of the film or is in linked into the underlying causal structure?

    However interesting object-oriented ontology may be – I've blogged quite a bit about it a year or so ago – I don't see that it provides any tools for addressing these questions. Nor, for that matter, do I know of any form of criticism that does.

    1. In answer to the first question -- we're probably driven back to Stanley Fish and the idea of the "interpretive community."

    2. Or the community of librarians who have to catalogue these things as "Gojira [film]. Japan, 1954" and "Godzilla [film]. USA, 1956". That's the tool that addresses the question. Then the scholars use the catalogue and find the two different films, whose differences are thoroughly documented, and turn them into something mysteriously philosophical.

    3. And it seems that the 1956 American film was dubbed into Japanese and then released in Japan with Japanese subtitles. It seems to me that whether you think of them as two versions of the same film or as two different films with extensive similarities is a matter of mere definition rather than one of deep substance.

      And the same goes, for example, the published version of Coleridge's "This Lime-Tree Bower My Prison" and the verse letter Coleridge wrote to his friend Southey. And so on for many other examples.

  9. I modified original text in one work to integrate it into a work of literature I composed. I wrote a narrative poem in blank verse on the life of philosopher Aristokles Platon. In a scene near the end, I picture him in the Akademia giving a lecture to his students. I took the long passage in The Republic widely known as the Allegory of the Cave, and rewrote it in blank verse as his speech to his students rather than a dialog between Sokrates and Glaukon brother of Platon. I tried to reexpress the exact concepts in more elaborate blank verse form. Since I have not yet published it, that text can be read here:

  10. About the time I got to the fourth paragraph I was inspired to write this fragment I intend to insert in my epic poem about the life of Lucretius.

    "Objects in this world are composed of atoms
    that exist as matter structured in patterns
    whether or not our brains perceive their forms,
    while our brains construct virtual reality
    from concepts of objects based on perception
    which we express through ideas in words,
    so we dream this already existing world
    then sing what we see in religious hymns."

  11. "...Moretti has the technical resources to make massive substitutions: he could change Crusoe's verb forms around electronically, and produce exactly the kind of modified text Harman describes..."

    Come to think of it, I'm not at all sure this would be possible with current technology. It's not as simple as doing a global search and replace to find all instances of, say, "big" and replace them with "large." Here's your example sentence, in context:

    In about three miles or thereabouts, coasting the shore, I came to a very good inlet or bay, about a mile over, which narrowed till it came to a very little rivulet or brook, where I found a very convenient harbour for my boat, and where she lay as if she had been in a little dock made on purpose for her.  Here I put in, and having stowed my boat very safe, I went on shore to look about me, and see where I was.

    I soon found I had but a little passed by the place where I had been before, when I travelled on foot to that shore; so taking nothing out of my boat but my gun and umbrella, for it was exceedingly hot, I began my march.

    What are you going to change that sentence to? How does the computer decide? Will the same kind of decision work in every case? I doubt that we've got off the shelf technology that can do this.

    1. We'd really have to talk to someone in Moretti's lab. They've got Stanford's Silicon Valley connections and a fairly strong set of resources, so it's not just a matter of Google book search. I don't know what they can and can't do, but I do know these things are progressing rather rapidly. Let's imagine five years out...

    2. I've got some training in computational linguistics (my dissertation was a quasi-technical exercise in knowledge representation), I know something about the history of the discipline, and I know a bit about the kind of technology used in Moretti's lab. It's one thing to automatically do syntactic analysis of texts. But taking existing texts and transforming fragments of them into syntactically different forms that nonetheless preserve meaning and are grammatically correct, that's a much more complex task.

      Moreover, it's not simply that you have to re-organize an individual sentence in isolation. But you have to do so in a way that preserves discourse relationships with preceding and following text.That's likely to be tricky stuff, the sort of thing that takes a bit of skill. That's the sort of thing editors do.

      I'm skeptical about computers being able to do that now and in the forseeable future.

    3. Well, I'll check in with you in five years. Meanwhile, here's a little experiment I think of as interesting from the Harman point of view...