Abstract
This review of Franco Moretti's Distant Reading
summarizes Moretti’s major arguments within the larger context of recent debates
in the digital humanities. Particular attention is given to Moretti’s uptake of
Immanuel Wallerstein, to his controversial critique of close reading, and to the
variety of digital-humanistic methods that comprise Moretti’s quantitative
formalism. Most valuable as an artifact of literary-critical history rather than
a how-to guide or theoretical treatise, this hodgepodge of essays is at its best
as an audacious and defensive academic memoir tracing Moretti’s transformation
into a digital humanist. As Moretti champions the broad explanatory power of
quantitative literary analysis, he overestimates the scientific objectivity of
his analyses while undervaluing the productively suggestive stories of doubt,
failure, and compromise that lend nuance and depth to his hypotheses. Combative,
absorbing, highly topical, and unevenly persuasive, Distant
Reading embodies both the optimism of early digital literary studies
and its perils.
Buried three-quarters into Franco Moretti’s
Distant
Reading is a moment of astonishment. Breaking for a moment out of the
tone of unruffled calmness that sustains most of the collected essays in this new
release from Verso, Moretti marvels at a cloud of data he produced by comparing no
fewer than 7,000 titles of British novels from 1740-1950: “all you can see are swarms of hybrids
and oddities, for which the categories of literary taxonomy offer very
little help. It’s fascinating, to feel so lost in a universe one didn’t even
know existed”
[
Moretti 2013, 181]. If this passage beautifully conveys that stunning moment in which
quantitative data and qualitative analysis merge, resolving into a single network,
it also conveys a sense of loss, of regret. Tellingly, Moretti’s self-imposed
alienation — he recounts a powerful personal experience in the second person:
you,
one — betrays an anxiety with his own
digital-humanistic method of distant reading, so “very different,” he reflects, “from the reading I used to
know”
[
Moretti 2013, 65]. This mixed emotion may sound out of key upon taking a step back to consider
this moment of ascendance for DH (digital humanities), which has recently attracted
so much recognition from national news outlets, university administrations, and
scholarly journals and presses, and (if for the moment) an array of job openings and
funding opportunities. But if we approach
Distant
Reading through this one conflicted pause, we can use it in turn to
approach the divisive conceptual apparatuses that structure DH meta-discourse at
present. It is exactly this conflict rhetoric that we must re-examine if we want to
avoid becoming the kind of sideshow that can distract institutions as they
perennially search for the next new thing.
The digital humanities are at a rhetorical and institutional crossroads, having just
enough critical mass of public attention to warrant a torrent of speculation. As
Alan Liu has termed it, DH — at first “a
loose suspension of topics and approaches” — has undergone a period of
“integration” and “expansion” that leads to the current “tipping point where it has the potential
not just to facilitate the work of the humanities but represent the state of
the humanities at large in its changing relation to higher education in the
postindustrial state”
[
Liu 2011, 8]. This moment of opportunity is equally a moment of danger, Liu cautions, as
DH must avoid being caught in a web of neo-liberal bureaucratic imperatives — a
danger highlighted, for example, by Wendy Chun, Richard Grusin, Patrick Jagoda, and
Rita Raley in their “The Dark Side of the Digital
Humanities” panel at MLA 2013 [
Chun et al. 2013], and by Matthew
Gold’s recent collection
Debates in the Digital
Humanities
[
Gold 2012]. Such worries have highlighted the conflict rhetoric
characterizing DH meta-discourse — with its familiar debates about distant versus
close reading, about tool builders versus tool users, about rejecting versus
incorporating theory, about using DH skills to be helpful around the office versus
being more protective of our time. What unites these debates is the common tension
between, on the one hand, desiring the fertility and excitement created by the
coexistence of a diversity of approaches and interactions among very
differently-minded scholars and, on the other, fearing that collaboration-related
compromises could devalue DH work, subsequently re-marginalizing DH workers. To
achieve the former and avoid the latter, we must find a way out of this
(increasingly) stale conflict rhetoric, which tends to reify false binaries, thereby
creating obstacles out of uncertainties and talking points out of ambiguities. By
understanding
Distant Reading as a miniature archive
that preserves this process of uncertainty and reification, we can become more aware
of this process — and can help interrupt it.
Distant Reading is most usefully understood as a
document of development and transformation, a journey that tells the story of Franco
Moretti becoming digital. The chapters have all been presented or published
elsewhere, but each essay’s fascinating prefatory explanation makes the book an
artifact of literary-critical history, rescuing the book from charges of being
little more than a fulfillment of his contractual obligations to Verso. Each preface
provides a key, a narrative kernel explaining the provenance of the essay — its
goals, its rhetorical occasion, its relationships with the other chapters — and,
more interesting still, evaluates its success or failure. He reconstructs his
journey as a comparatist interested in evolutionary theory (
Modern European Literature, 1992) and in mapping (
Atlas of the European Novel 1800-1900, 1998), slowly piecing these
areas into a “quantitative analysis”
[
Moretti 2007]. While
Graphs, Maps, Trees can serve as a
methodological primer, it is in
Distant Reading that
Moretti shifts focus from turning literature into datasets to the affordances of
doing so. Perhaps not coincidentally, it is in this book that he avoids
acknowledging his debt to narratology — making his methods seem more revolutionary
than they perhaps are. But he does offer an alternative context: “In the last few years, literary
studies have experienced what we could call the rise of quantitative
evidence. This had happened before of course, without producing lasting
effects, but this time is probably going to be different, because this time
we have digital databases and automatic data retrieval”
[
Moretti 2013, 212]. Fitting snugly alongside the rise of the quantitative humanities — or at
least an increased willingness to consider that literature is a kind of data and can
therefore be translated into other forms of data — Moretti’s career trajectory
models a kind of alternate academic history in which the digital humanities
developed organically from within traditional methods of scholarship. Though such a
simple, uninterrupted development is at best atypical and at worst disingenuous, it
is compelling as an anti-conflict narrative that emphasizes the shared goals of the
“traditional” academy and DH.
Put simply, Moretti’s distant reading involves sampling from a buffet of “visualization tools” that
convert any given slice of literary history into “a set of two-dimensional signs...that
can be grasped at a single glance”
[
Moretti 2013, 211]. Put more controversially, he claims that eschewing close reading is not only
a positive change for the humanities, but is actually “
a condition of knowledge
”
[
Moretti 2013, 48]. Moretti then defines knowledge as pinpointing and explaining temporal and
national variations in “devices, themes, tropes — or genres
and systems”
[
Moretti 2013, 49]. Curiously, once Moretti does away with close reading, the first thing he
does, as if scared by the resulting vacuum, is usher in Immanual Wallerstein’s
world-systems analysis [
Wallerstein 2004]. Moretti’s preoccupation
with evolution, the “study of formal mutation and cultural
selection”
[
Moretti 2013, 63], finds in Wallerstein’s world-systems theory “a strong conceptual model” even though “no one seemed to use [it]
anymore”
[
Moretti 2013, 43]. Moretti's first two chapters, “Modern European
Literature: A Geographical Sketch” and “Conjectures
on World Literature,” as well as the related chapters “More Conjectures on World Literature” and “Evolution, World-Systems,
Weltliteratur,” use Wallerstein as a glue for combining
evolutionary theory and DH techniques. In these chapters he traces a spatial (rather
than temporal) evolution of literature as an unfolding constellation of “formal novelties”
[
Moretti 2013, 41] radiating out of many points. This approach has been criticized for
overstating the influence of market demands on world literature, but, structurally,
the true weakness of
Distant Reading is that five of
its ten chapters deal heavily with world-systems analysis even though this is,
putatively, not a book about Wallerstein but about distant reading. This
inconsistency could have been cleared up with a theoretical introduction that
explicitly situated (and defended) the relationship between distant reading and
world-systems analysis, but Moretti provides no such introduction, leaving the
reader with the choice either to construct an integrated argument from statements
scattered in various chapters, or to ignore either the Wallersteinian or the
digital-humanistic chapters.
Of the digital-humanistic chapters, the ambitiously titled “The
Novel, History and Theory” most dramatically suggests quantitative
formalism’s potential. In it, Moretti uses functional magnetic resonance imagery to
displace European realism’s literary-critical hegemony, making the literary field “historically longer, geographically
larger, and morphologically deeper”
[
Moretti 2013, 161] — an effort admirably connected to recent attempts to destabilize
nationalistic models of literary periodization by scholars like Susan Stanford
Friedman, Rebecca L. Walkowitz, and Eric Hayot. The following chapters are also
impressive in their ambition; “The Slaughterhouse of
Literature” describes how Moretti and his graduate students read
all British detective narratives within a certain period to discern
which stylistic features distinguish the canonical from the non-canonical, while
“Planet Hollywood” claims to prove how “plot and style [are] becoming
manifestly de-coupled as a result of their movement (or not) in
space”
[
Moretti 2013, 92]. And the final chapter, in what is clearly
Distant
Reading’s signature interpretation, converts
Hamlet into a series of character-networks to discover that the hidden
protagonist of the play is the nascent European international state system. Moretti
makes no secret of his basic gambit: the reader can easily supply the silent “before me” in his observation that “we have never really tried to read
the entire volume of the literary past”
[
Moretti 2013, 210]. Indeed, making the impossible seem possible (and desirable) is what Moretti
does best. And his tone of relaxed omniscience is at times seductive. But when he
claims that his grandiose stories are “resting solidly on facts”
[
Moretti 2013, 44] or that distant reading yields “the clarity of the empirical
confirmation,”
[
Moretti 2013, 92] the generalization-spouting bravado by which Moretti skates over impossibly
broad terrain finally comes across as glib, revealing one of the primary dangers of
digital literary studies to be the adoption of an aggrandized, even hubristic
attitude toward literature as so much inert
stuff being poked at.
At the same time, this inertness seems to connote a fresh impartiality, a promise of
neutrality and newness, even as it suggests another, more familiar controversial
tendency of quantitative analysis: that it can become “a patchwork of other people’s research,
without a
single direct textual reading.” For Moretti this is no
problem, as he argues that the “ambition” or scope of the literary-critical project is “directly proportional
to the
distance from the text: the more ambitious the project, the
greater must the distance be”
[
Moretti 2013, 48]. Haters left unconvinced are dealt with in a clever but catty passage that
accuses close readers of pro-canon elitism:
The United States is the country of
close reading, so I don’t expect this idea to be particularly popular. But
the trouble with close reading (in all of its incarnations, from the new
criticism to deconstruction) is that it necessarily depends on an extremely
small canon. This may have become an unconscious and invisible premise by
now, but it is an iron one nonetheless: you invest so much in individual
texts only if you think that very few of them really
matter.
[Moretti 2013, 48]
Though Moretti’s all-or-nothing rhetoric is extreme — and his “
only
” far from earned — he is not alone contrasting new DH methods with close
reading; Lev Manovich has called databases and narratives “natural enemies”
[
Manovich 2001, 225], while Edward Whitley coolly assumes that literary critics “value close reading...over the broad
brushstrokes of information visualization”
[
Whitley 2011, 188], and Matthew Wilkens serenely owns that “we’ll almost certainly become worse
close readers”
[
Wilkens 2012, 256]. Yet David Hoover has complained in
DHQ of DH’s
“marginalization of textual analysis
and other text-centered approaches” despite “the tremendous potential of electronic
texts for literary studies”
[
Hoover 2007, para. 4]. And when Stephen Ramsay observes that “[t]he digital revolution, for all its
wonders, has not penetrated the core activity of literary studies”
[
Ramsay 2008, 477], his real claim is that it has not
yet been done, not that it
cannot happen. Even distant reading does not work by
not reading: more accurately, Moretti means selective reading
(reading only the titles, only the first paragraphs, or scanning for certain
patterns), delegated reading (recruiting his graduate students), or mediated reading
(using search tools to generate statistics and charts). Furthermore, though Moretti
uses DH to generate grand systems of literature, these methods can also be used to
return to the text — particularly if we link them to the “surface reading” of Heather Love, Stephen Best, and
Sharon Marcus, as seen in the special issue of
Representations devoted to it and by Love’s “Close
but not Deep”
[
Love 2010]. Best and Marcus’s introduction links new strands in
literary criticism (including DH, actor-network theory, and new approaches in
narratology) as evidence of an emerging preference for descriptions of “the complexity of literary surfaces”
instead of a psychoanalytically, historically, or politically inflected “symptomatic” method that “assumes that a text’s truest
meaning lies in what it does not say”
[
Best and Marcus 2009, 1]. It may be more accurate to say that surface reading, distant reading, and DH
do not oppose close reading
per se, but rather certain
kinds of interpretation habitually associated with (but not inherent to) close
reading. Pursuing these new forms of analysis does not automatically exclude close
attention to the sentence-level details of form and style; indeed, moving back and
forth between the microscopy of close reading and the wide-angle lens of distant
reading would enrich both methods, creating a dual perspective that boasts both
specificity and significance.
If the text of
Distant Reading is confidence, its
subtext is failure. Moretti repeatedly, generously, though fleetingly faces being
wrong. He admits some of his work “may well have overstated its
case”
[
Moretti 2013, 119], muses at another juncture “I may be exaggerating here,”
[
Moretti 2013, 228], and laments that his lack of technical skills drove him to hand-draw his
visualizations in the early stages of distant reading. “This is not a long-term solution, of course,” he remarks of
the hand drawings, but instead a characteristic of “the childhood of network theory for
literature; a brief happiness, before the stern adulthood of
statistics”
[
Moretti 2013, 215]. But such “failures” are only a
problem if we deprecate the “brief happiness”
and make post-haste for that “stern
adulthood,” if we demand absolute truth and methods sanitized from any
taint of non-DH methods. Unfairly deprecating to himself and his graduate students
in this section, Moretti undervalues how his book successfully records select “home grown” digital literary studies, and as a
result,
Distant Reading builds fences around them. What
if instead we privilege openness and adopt the ad hoc playfulness of open-source
programming, recognizing that systems and theories and methods are but temporary
scaffolding, not the grand erection of something permanent and totalizing? Digital
projects are often ends in themselves, not means — sometimes forgotten, fragmented,
bug-ridden, and incomplete — but what is open is less likely to get bogged down in
the past. To be open does not automatically entail the “technological hipsterism” that Amanda Gailey and Andrew
Jewell warn of, that neurotic coolness that regards “the quality of work...not so
important as staying at the edge of innovation”
[
Gailey and Jewell 2012, 5].
Ultimately, Moretti needs to take his own advice: he reasons, “since no one knows what knowledge
will mean in literary studies ten years from now, our best chance lies in
the radical diversity of intellectual positions”
[
Moretti 2013, 89]. As he concludes the book, he ends with a warning that before quantitative
analysis can begin in earnest, “an enormous amount of empirical data
must be first put together. Will we, as a discipline, be capable of sharing
raw materials, evidence —
facts — with each other? It remains
to be seen”
[
Moretti 2013, 240]. While the creation of (and access to) complete literary archives and data
sets will certainly meet with obstacles,
Distant
Reading conjures a very different specter: overestimating the power of
data and in doing so losing the texture and feel of the objects we study, setting
ourselves up as so many Carmen Sandiegos: megalomaniacs hoarding what is impossible
to hoard. It is better to seek the moments like the one opening this review, in
which Moretti felt awed, “so lost in a
universe” filled with “swarms of hybrids and
oddities”
[
Moretti 2013, 181] — moments of suspension that suggest that it is not the
“knowledge” that results, but rather our
wonder
while encountering a new dataset that makes DH methods so attractive. Similarly, the
conflict rhetoric of DH meta-discourse tends to suspend, rather than resolve, its
favorite debates, putting off the agony of saying that DH is one thing rather than
another, preferring instead to work inside the “what if” and the
maybe, the conditional and the subjunctive, rather than inside absolutes and
interdictions. Perhaps we should also preserve that moment of suspension when we
manipulate our data.