After Close Reading


Review of DISTANT READING, by Franco Moretti and BETWEEN CANON AND CORPUS: Six Perspectives on 20th-Century Novels by Mark Algee-Hewitt and Mark McGurl

Verso, 2013 / Stanford Literary Lab, 2015 (view .pdf)


    Just about every review of his work starts off by saying that Franco Moretti wants to be the enfant terrible of literary criticism, and this one will be no different.  In a number of recent essays, lectures, and books, he’s made it clear that he wants to be that guy: the disruptor who upsets the applecart of the Usual to vend the finer fruit of the Counter-intuitive; the boy who points out that the emperor has no clothes and then morphs into the Grand Vizier who re-clothes him. Moretti has been the leading polemicist for bringing quantitative methods into a field that, until recently, faithfully observed W.H. Auden’s instructions: “Thou shalt not sit with statisticians/ Nor commit a social science.”   Moretti does both, using (variously) statistics, big data, information theory, network analysis, and whatever else he can fit into his counter-intuitive toolbox, to formulate new theories about the novel as genre, literature in general and interpretation at large.  And more scandalously still: Moretti does so in order to put the study of literary texts on what he claims will be a firmer epistemological footing, one in which students of literature will be brought up to standards of rigor more commonly found in other disciplines: “the pursuit of a sound materialistic method[1], and of testable knowledge.”  (Distant Reading, 155) If what emerges in this quest is often less than earthshaking, the brio and versatility that stand behind it allows Moretti to pose powerful questions to the rest of us: nothing less than, what are we doing when we talk about “literature”?  And why are we doing it?

    His intervention is undoubtedly timely.  Literary criticism in the academy has reached a crisis point, and what we mean by “reading” stands at the center of the storm.  For while we academic critics have moved from the hegemony of New Criticism to deconstruction to new historicism to cultural materialism to queer theory to postcolonial criticisms, “close reading”--the attentive inspection of the verbal texture of poems, novels, and plays--continues to be the methodological basis of what we do in our most important venue: the college classroom, especially the Intro to Lit classroom.   There are good reasons for this.  From the rapid postwar expansion of the 1950s forward, close reading became and remained central to the introductory class, where teachers found that students lacking specialized knowledge of the ins and outs of English history or the finer points of Aristotelian logic could still get excited by talking about the form of a Donne lyric or image-patterns in Portrait of the Artist as a Young Man.  The so-called “New Criticism” was thus a perfect tool for the enlarging universities of the post-war era; it allowed public-school trained students at the University of Illinois or Iowa to have as much to say about texts as their preppie counterparts at Yale or Harvard.  And speaking as a teacher, I can testify to its utility.   Moments when students glimpse why Milton uses enjambment in the first lines of Paradise Lost (“Of man’s first disobedience and the Fruit/Of that forbidden tree…”:  Milton’s ambivalence about the cost and the necessity of the Fall is contained in the gap between the last word of the first line and the phrase that begins the second) or how Trollope makes the plot lines of The Way We Live Now rhyme with each other to suggest what it’s like to live in a marketplace world, are among the most rewarding of my career.

   But one by one, the props under the regime of close reading have been knocked aside.  New critical imperatives --the study of gendered difference, imperialisms, of class, of race construed over time—didn’t just challenge our attitudes towards texts, they changed our very ways of reading them.  Close reading became, at best, “reading against the grain” (a formulation adopting Walter Benjamin’s famous injunction to “brush history against the grain”); at worst, close reading was seen to crystallize an ahistorical, undialectical, power-reaffirming, aestheticizing academic practice.  Students, too, have changed: fed on a diet of instant messages and twitter feeds, they seem to be worldlier than students past—than I and my generation were—but to find nuance, complexity, or just plain length of literary texts less to their liking than we did.  (I tried teaching The Way We Live Now again last year, after a fifteen-year hiatus.  It did not go well.)  Reading, with or against the grain, seem anachronistic, out of touch with the times. To add to the dilemma, students are leaving the English major and its ancillary fields at places like Harvard (which constituted a committee to find out why students are abandoning the Humanities) as well as the University of Maryland (which reported a 50% drop in English majors in one year).  Students are voting with their feet as well as their attention spans.  The study of literature has got to change; we all know that.  But how? 


              Moretti is not the only critic to argue for a new paradigm—the influential journal Representations, founded at Berkeley as the voice of the so-called New Historicism, announced a similar turn from what they called “symptomatic reading” like the kind practiced by many of their predecessors towards “surface reading,” a vague term encapsulating Art of the Book, genre and other non-depth oriented perspectives.  Moretti’s is similar in its intent, but different in practice: he remains in some sense what he has always been, a genre critic, only now one armed with tools that enable him to think about groups of literary texts in a broader, more synoptic way.  “We know how to read texts,” he says at his most provocative, “now let’s learn how not to read them.”   He advocates in its place “distant reading: where distance . . . is a condition of knowledge.  It allows you to focus on  units that are much smaller or much larger than the text . . . . And if, between the very small and the very large, the text itself disappears, well, it is one of those cases where one can justifiably say, Less is more.” (Distant Reading, 48-49)  

    Moretti is on his strongest ground when he reminds us, following the lines of scholars like Margaret Cohen, that the field of literary study has only previously skimmed the surface of the texts of a given period and/or national configuration.   “I work on . . . the canonical fragment  [of] west European  narrative, . . .which is not even 1 percent of the published literature.   . . Some people have read more, but the point is that there are thirty thousand nineteenth-century British novels out there, forty, fifty, sixty thousand-- no one really knows, no one has read them, no one ever will.  And then there are French novels, Chinese, Argentinian, American . . . .(45).   To a certain extent, this is an updated version of the injunction to open up the canon that swept through the American academy the late 1980s and early ‘90s: but Moretti's is canon opening on steroids.   It’s not just a question of moving from valorizing Moby Dick to Uncle Tom’s Cabin or The Wide Wide World by Susan Warner (a re-orientation Jane Tompkins made famous during this period of time); or of admitting the work of people of color into the study of literary statement or admitting popular or mass culture to the critical ken; it’s far more expansive than that[2].   Moretti’s is an attempt to re-envision the literary system in toto, canonical and non-canonical alike, along the lines undertaken by the French Annales school (Braudel, Bloch) or (in an example Moretti adopts and then drops) Immanuel Wallerstein’s World-Systems theory.  And while his ambition is only sporadically realized in Distant Reading, it’s clear that the not-yet tapped potential of the digitalization of libraries will help him realize it more fully in the future.  If close to all the books in the world get digitized, we’ll be able to perform stylistic analysis on a truly global basis (if we can solve the pesky problem that they were written in different languages and published in different alphabets), or compare Asian with African and with Western genres, giving real meaning to the “comparative” in comparative literature. Our notion of the worldliness of literature can finally go planetary, limited only by the data we process. 

   But of course, there’s the rub, to quote the author whom, Google N-grams tells us, still occupies .002 of all the books Google is Googling. The conclusions that emerge from any statistical massaging is only as good as the data going in—GAGO, Garbage in Garbage out, social scientists like to say.  So great care must be taken to make sure that the material you choose to analyze really constitutes a representative or significant sample before you can analyze it, which means that you can’t escape “reading” in some sense: you have to interpret what you have accumulated to turn into data before you can interpret the data that you’ve been given.  And while Moretti insists that he and his fellow toilers in the digital vineyards understand the issue, he gets into problems on precisely this score.   In one admittedly speculative essay, “Planet Hollywood,” he traces the hegemony of U.S. films in the world market; but  while his methodology allows him to comment on Hong Kong, it  precludes him from noting its striking non-presence in India, which has only the largest film audience in the world, 300 million filmgoers avidly consuming Bollywood, not Hollywood, fare. Similarly in another a famous essay from Graphs, Maps, Trees (2007) he argues that literary genres have a delimited lifespan but draws his understanding of genres from an incredibly limited source –30-odd academic monographs and dissertations, some of them understandably, others of them selectively, if not whimsically, chosen.

   But choose one must, or else one will be swept away by the sheer plethora of material that the Googlization of everything brings with it.  Perhaps (àla Moretti) one could mount an anatomy of titles from the hundreds of thousands or even millions of books potentially part of this inclusive archive but finding anything more significant than that seems problematic in the extreme; counting titles, or “ands” and “thes,” or assembling words designating sounds or emotions will only take you so far.  This not accidental; in a literary version of the mathematical law of large numbers--that the more repetitions of a random event, the more the results will revert to the mean—it seems that the bigger the data-set, the more limited the conclusions one can draw from it, because the more any idiosyncratic variations will disappear under the weight of the repeated.   But then how to make the selection meaningful? 

    The Stanford Literary Lab, working with critic Mark McGurl, faced this latter problem resolutely when they set out to study contemporary fiction.  Quailing at the prospect of dealing with a mere 279,000 titles, they sought to create a “subset” of the “most reasonable, interesting and useful” 350 books of the twentieth century in order to perform their analyses.   The criteria (vague but loaded—reasonable according to what standard? Interesting to whom? Useful for what?) forced them to eschew either the elitist option—draw up the list themselves, or choosing an expert professor to do so—or the obvious alternative choice—random sampling, this latter option rejected for the reason that it would probably exclude the likes  of Joyce, Woolf, Faulkner, et al.    They set out to compose their own canon on the basis of lists of the greatest books of the 20th-century compiled by Random House and one proposed by its readers; a list of the favorites composed by participants in the Radcliffe Summer Institute designed to prepare people for jobs in the publishing industry; a Publisher's Weekly survey of its readers on the same question;  a list of the century’s best-sellers; and a list of the top works of the century composed by college professor and experimental fiction buff Larry McCaffery.  Fearing an underrepresentation of works by women and writers of color, they then reached out to groups of professors focused on postcolonial letters to suggest other works, which they added in to make roughly 350. 

   Compared to Moretti's frequently unconvincing attitude towards the composition of his data-sets, this is undeniably a step in the right direction.  But what does this list tell us?  Is it really so different from the option of having one professor choose what he or she thinks is the most “reasonable, interesting and useful” books of the century?  Putting your thumb on the scales in such a way as to over-represent  elite opinion and under-represent genre fiction is fine—as a former girlfriend once said of herself just before dumping me, at least they’re being honest--but what does it mean that a huge-selling genre, Westerns, is represented by two books? Works I would call "airport reading"--the novels of Tom Clancy and John Grisham, are on the list, but romance novels are not--these compose roughly 17% of all  books sold in the U.S.-- despite much attention that the genre has received from feminist critics, like Janice Radway, whose  Reading the Romance changed the way we think about the genre some thirty years ago.   Ayn Rand, Stephen King  and L. Ron Hubbard are there but not Christian fiction (like the Left Behind series, which has sold 65 million copies).  And so on.  Seeking both representativeness and distinction, popularity and stylistic diversity, what they really offer us a reflection of the tastes and lifestyles of upper-middlebrow readers  (who are, sadly, all too influenced by Ayn Rand, as the horrid state of U.S. politics would suggest[3] ) and their professorial kin.   The results are, I suppose, not entirely uninteresting, but they’re also not entirely unexpected--and a far step from composing  the subset of the 350 “most reasonable, interesting and useful” books  of the century.   What we have instead, as in so much digital humanities work, is statistical reification of the assumptions that went into the making of the data from which statistical inferences will then be drawn. The hermeneutic circle is inescapable--and, at times, vicious.

   Problems grow we turn world literatures.   As far as “distant reading” is concerned can one really perform it in the same way globally as one can with, say, European literatures? Setting aside questions of methodological colonialism—do we really want to enter a world where the un-read writing of the non-Western world becomes just a set of data points along with its over-read partners in the West? --how is the archive composed for the non-European world? This is  a question not just for Moretti, of course, but for the whole of the world-literature phenomenon heinvokes.    Consider the Chinese example.  Over the course of the last dozen years or so,  Moretti has devoted a good deal of attention to Chinese fiction:  he selected two pieces on the Chinese prose narratives in his anthology, The Novel, including a fascinating essay by Sinologist Jeffrey Plaks on how Chinese prose-narrative might be thought of as growing out of debates very similar to those which provided the matrix from which the European novel grew; Moretti writes, as we shall see in more detail below, about the Chinese “novel”  The Story of the Stone (a.k.a.  The Dream of the Red Chamber ) in a provocative essay that compares the structure of its characters’ relations with those in Our Mutual Friend and Hamlet;  elsewhere, he provocatively juxtaposes  the question of the development of the novel in China and that of the West (why did one grow faster than the other?) (176)  What, I wondered to myself, is the status of the archive of Chinese prose fiction?  Could we create a list of Chinese fictions like those the Stanford Literary Lab is composing for modern American novelistic fictions?  If so—or if not—what would it mean? I turned to a Chinese-American scholar to help me with this question and she responded with a brief history lesson in Chinese list making:


The first compendium/list of Chinese classics (including but not limited to longish fictional narratives ) appears during the Kangxi reign (Qing dynasty) in a volume called 四库全书 Sìkù Qúanshū, aka Complete Books of the Four Silos. There is no list of Chinese novels. From the late Ming to the mid- Qing, "the Four Classics 四大名著 Sìdà Míngzhù" entered common parlance. This list included Dream of the Red Mansion, The Water Margin, Romance of the Three Kingdoms, and Journey to the West. It was based on late Ming dynasty critic Jin Shengtan金圣叹's previous list Four Miraculous Works 四大奇书 Sìdà Qíshū. In the c18 & c19 Western missionaries/sinologists made lists of Chinese classics. They include Elijah Coleman Bridgman & Arthur Waley and really too many to name.

    The important point is: Chinese people did not really go about making comprehensive lists of their great literature on a regular basis. Chinese libraries used to be called Chambers for Hiding Books 藏书楼cángshūgé because books were precious things to squirrel away, not externalize. The raison d'être for the c18 _Complete Books of the Four Silos/libraries_ cannot be separated from the emperor Kangxi's imperial desires. Such canonizing impulses usually are--- and their reproduction usually come as a result of "international " contact.

   Also Chinese lists of novels such as they are only really useful for thinking about regime legitimation, nationalism, or communist propaganda, and canonicity. There's a reason why the Four Classics are allowed to remain as they are (with some notable periods of censorship especially for Dream of the Red Mansions): three of them can be read as loyalty to old regime AND inevitability of revolution & new regime. The fourth one is about a monkey. 


      If this scholar is right, to admit Chinese literature into our purview is also be forced to admit that colonialism from without and internal politics from within shaped and continue to shape the Chinese fictional prose-archive; which is to say that there can be no study of the literary object without also the studying history, culture and circumstance which shaped very constitution of the object in question.  Twenty years ago, this would be a truism, maybe even ten; at the current moment of extravagant faith in data—not just Moretti’s but  also that of the digital humanities at large--it’s useful to reiterate the lesson.

   All of which leads me back to my initial question: are we really beyond close reading?   Or do we in some sense need to return to it--in a way responsive to historical and cultural circumstance--chastened, perhaps, by contrast between the grandeur of the possibilities of a Unified Field Theory of Literary artifacts and the stubborn questions posed by individual works to defy such reification?  Can we go back again to a place we know it untenable?  Can we move forward to a future that looks less tautological? 

   Perhaps without knowing it, perhaps consciously, Moretti starts us on the path to providing an answer.  In a really interesting chapter, the last one in the book, he uses network theory to compare and contrast the social networks, measured by nodes of contact between various characters in HamletOur Mutual Friend, and The Story of the Stone a.k.a. The Dream of the Red Chamber.  These works he claims, look different when subjected to this kind of analysis, and he proceeds to draw diagrams that demonstrate this difference.  Intuitively I would tend to agree with the importance of such an exercise, and on the basis of what I see as the ultimate test of any reading-system: what happens in the classroom.  I remember the spectacular efficacy of tracing on the blackboard the relations among characters when teaching Great Expectations a century-and-a-half ago, and Moretti’s seems a much more sophisticated version of that.  (All relations in Great Expectations, it seems, radiate out from the central node of the relation of con-man Compeyson and Miss Havisham, the primal betrayal that sets all the novels many plots in motion.)   Here’s a particularly striking moment in Moretti's reading:


     Take the characters that are connected to both Hamlet and Claudius [in Moretti’s network]: except for Osric and Horatio, whose link to Claudius is however extremely tenuous, they are all killed. Killed by whom, it is not always easy to say; Polonius is killed by Hamlet, for instance—but Hamlet has no idea that it is Polonius he is stabbing behind the arras; Gertrude is killed by Claudius--but with poison prepared for Hamlet, not for her; Hamlet is killed by Laertes, with Claudius’s help, while Laertes, like Rosencrantz and Guildenstern before him, is killed by Hamlet but with Claudius’s weapons.  Individual agency is muddled; what is truly deadly, is the characters’ position in the network, chained to the warring poles of king and prince.  Outside of that bold region, no one dies in Hamlet.  The tragedy is all there. (217)


       Setting aside poor Ophelia, as neglected here by Moretti as she was by Hamlet, this is a really interesting analysis of the play.  The network analysis really does show us something new: that this is a play which on a structural level enacts a conflict over sovereignty, over who is to rule the Kingdom, and, more generally, from where sovereignty derives its authority:  the court (represented by Hamlet and Claudius—their hegemony riven by a conflict between blood and power as sources of legitimacy—a conflict relatively neglected by Moretti, but one which he shows somewhat malgrélui to be crucial;) or the emerging post-Court State (represented by Horatio—conspicuous in his flat language and absence of any nodes  of connection to the Court society, though Fortinbras, equally flat and even more disconnected, would seem to make his point better).   Like my chart of Great Expectations, but in an infinitely more complex way, Moretti articulates in a pedagogically useful fashion what we used to call deep structure of the play and the ways its characters’ relationships define that structure. 

    But that is not all his argument does.  What gives it its kick is the simultaneous and not un-contradictory recognition of the power of mistaken death throughout Hamlet, revealing a different pattern of possibilities underlying the conflict over sovereignty.  Pace Moretti himself, he shows that “individual agency isn’t just “muddled” in the play, it’s utterly inefficacious in contrast to . . . accident? fate? Darker powers than we humans can know? Random happenstances that defy the odds to work themselves out disastrously?  Seen from the perspective of ubiquitous, seemingly accidental death, the conflict over earthly power looks inessential; instead, the play renders a world that is murky, drenched with dread, magic and fate—in short, uncanny. Networks aren’t necessary for Moretti to get to this latter place.  But sheer intuition and power as a reader aren’t enough to get hin to the former, to seeing the play as a struggle over sovereignty, either. He needs—and offers—both, giving  us a reading of the play as a struggle between these two different sorts of power.  I leave it to you to choose whether earthly power or the uncanny power of death is primary in the play if not our lives; so, I think, does Shakespeare.   The rest is silence.

   Two sorts of power, then, but also two ways of reading.   We need them both--and as I have tried to show, Moretti is conversant in each.[4] We need to embrace the new technologies Moretti and the Stanford Literary Lab are exploring because they do offer incredibly powerful tools for understanding the phenomenon of literature, and will clearly enable us to say certain things about it which are just at the dawning of being explorable.  Who knows what Moretti will come up with next?—his most recent publication treats the language of the World Bank, in what seems to be a promising extension of stylistics into the  analysis of neoliberal doublespeak.  Who knows what McGurl, one of our finest critics of contemporary fiction, will do with his lists?  Who knows what other readers and critics will be able to imagine and create out of the mass of agglomerated data that confronts us in the brave new world of the digital everything?

    But at the same time, Moretti’s own performance shows that we also need the intuitions, the critical reflexes, the sharp attention that can only come with reading, reading, and more reading--and attention to the meaning-making capacities of the tests we read--to guide us to what is important in composing our field of data, and then to help us make sense of it once we find it.   You have to have some knowledge of the parts—and this means of the words as well as the narrative structures—of the literary system to begin to assess its work as a system; you need to have a sense of the system in order to put its parts in order. But more: distant reading doesn’t just have a guilty, complicitous secret-sharer relation to soi-disant close reading: it depends on it.  Drawing on the techniques of intrinsic analysis of literary texts becomes all the more necessary if we are to keep from drowning in the sea of undifferentiated and undifferentiable data.

       The paradoxical effect of Moretti’s polemic for distant reading, then to remind us of the continuing importance of its double, twin, and dialectical companion, close reading.  Whether an approach incorporating both would bring undergraduates back to our classes or not—I tend to doubt it (my wan hope is that the iron laws of supply and demand will soon render STEM fields unattractively overcrowded, and ours an exciting opportunity for the brave of heart)—it may help us overcome the legitimation crisis to which all wings of the literary profession seem to be prey at the current moment.  It may, in other words, force those of us who are willfully or characterolgically immune to the charms of statistics, data-crunching and chart drawing, to up our game, to articulate why, in an age in which  data is king and close attention to the nuances of language and form passé, we need to pay attention to the meaning that is made by such meaning-making mechanisms.. Moretti is quite helpful here; his polemical arguments for a practice that he does not—I would suggest, cannot--follow, reminds us that we need the virgin and the dynamo, as Henry Adams might put it, or both the ghost and the machine, as Gilbert Ryle, Arthur Koestler, and the X-files would insist, in order to grasp both the facticity and the uncanniness of literary statement.  And isn’t it that doubleness that drew us to it in the first place?


[1]  Moretti’s use of this loaded term is no accident: he sees the turn towards the empirical as a logical outgrowth of the empirically-oriented Italian Marxist tradition he cleaved to in his youth.

[2]  One reason Americanists such as myself are less impressed by the audacity of  Moretti’s work is that the inventorying, “distant reading” impulse has long been a powerful one in U.S. literary studies. See among so many others,  William Charvat’s The Professon of Authorship in America, 1800-1870 (New York: Columbia University Press, 1968), John William Tebbel’s four-volume History of  Book Publishing in the U.S.  (New York: R.R. Bowker,  1972-1981); Mary Kelley, Private WomenPublic Stage: Literary Domesticity in Nineteenth-Century America (New York: Oxford University Press, 1984); it’s interestingly combined with a case-study or micro-history in June Howard, Publishing the Family (Durham: Duke University Press, 2001). 

[3] The Libertarian bent of this audience’s response is demonstrated by the presence of not one but four Rand novels and three by Robert Heinlein, whose characters often spout wisdom to each other like “TANSTFAAL”—an acronym for “there ain’t no such thing as a free lunch.”  The relation between Libertarianism and upper-middle-class class self-conceptions is an underexplored topic, one this reading list brings to the fore. (There are very few impoverished Libertarians.)  It’s a genuine  service that McGurl and the Stanford Literary Lab have brought it to our attention.  But they don’t need to digitalizing project to go there, and in general the focus on stylistics and genre analyses suggests that they might not explore it any further.

[4]  In his book,  The Bourgeois, written and published at the same time as Distant Reading,  Moretti continues his mixture of close reading and historically based genre criticism, with only the slightest touches of Literature-Lab-like analysis which he feels free to accept or reject on a case-by-case basis.

JONATHAN FREEDMAN is a professor of English and American Studies as well as a faculty associate in the Frankel Center for Jewish Studies at the University of Michigan.