First came the debacle of the high-priced “Ada” algorithm, the control center of Hillary Clinton’s ill-fated operation. Next ESPN wonk Nate Silver, after flubbing the 2016 election forecast, defended his numbers by claiming that he was not more wrong than every other statistical predictor since 1968. Finally, consider the kerfuffle over Cambridge Analytica, the British company whose “psychographics” method of data modeling and “emotion analysis” claimed to be the Trump camp’s secret weapon — until skeptics recalled that Ted Cruz and Ben Carson had employed their services as well.
We’re sorry, something went wrong.
We are unable to fully display the content of this page.
This is most likely due to a content blocker on your computer or network.
Please allow access to our site and then refresh this page.
You may then be asked to log in, create an account (if you don't already have one),
or subscribe.
If you continue to experience issues, please contact us at 202-466-1032 or help@chronicle.com.
First came the debacle of the high-priced “Ada” algorithm, the control center of Hillary Clinton’s ill-fated operation. Next ESPN wonk Nate Silver, after flubbing the 2016 election forecast, defended his numbers by claiming that he was not more wrong than every other statistical predictor since 1968. Finally, consider the kerfuffle over Cambridge Analytica, the British company whose “psychographics” method of data modeling and “emotion analysis” claimed to be the Trump camp’s secret weapon — until skeptics recalled that Ted Cruz and Ben Carson had employed their services as well.
The dream that algorithmic computation might reveal the secrets of complex social and cultural processes has suffered a very public and embarrassing results crisis. These setbacks have also led to some soul-searching in the university, prompting a closer look at the digital humanities. Roughly a decade’s worth of resources have now been thrown in their direction, including the founding of an Office of Digital Humanities at the National Endowment for the Humanities, unheard-of amounts of funding from the Andrew W. Mellon Foundation, a parade of celebratory anthologies backed by crossover articles in high-profile magazines, and academic job openings in an era of tenure-track scarcity. So, with all their promise, and all this help, what exactly have the digital humanities accomplished?
It is a hard question to answer. Digital techniques conform beautifully to some types of humanistic work — compiling concordances, for example. They have also proven valuable in deciphering ancient languages. The great code-breaker Yuri Knorozov, who employed a “statistical-positional” method for deciphering Mayan stelae, may have cracked the ancient language without aid of computers, but the computer-friendliness of his procedures made them useful in later work translating Harappan texts.
The digital humanities have also attracted scholars doing distinctly critical kinds of work. A maverick like Jessica Marie Johnson, for example, an assistant professor of Africana studies at the Johns Hopkins University, argues in a Los Angeles Review of Booksinterview series that #BlkTwitterstorians and other uses of social media have “helped people create maroon — free, black, liberatory, radical — spaces in the academy.” Only an ideologue would doubt the potential of low-cost networking for communities typically stopped by financial or cultural obstacles at the college door.
ADVERTISEMENT
Finally, when it comes to modes of presentation — specialized library websites, text/hypertext editions of challenging canonical novels, multimedial history textbooks — the hard labor of digital humanities seems to pay off by establishing smooth or attractive delivery systems for big data.
Computers demand to be asked only what they can answer, changing the questions to conform to its own limitations.
Yet a fair assessment of the field is stymied by a basic confusion of terms: the digital in the humanities is not the digital humanities. Few humanists today are ignorant of Moodles, podcasts, auto-formatting, or deep internet research. Even Luddites are not averse to consulting the Kindle version of a novel in order to search it for phrases without having to page through a physical book. The term “DH,” then, is not about introducing digital technologies where there were none before, but about an institutional reframing. What people mean by “DH” is a program and, ultimately, an epistemology.
DH insiders agree. Ted Underwood, a professor of English at the University of Illinois at Urbana-Champaign, insists that “machine learning is really, honest to God, a theory of learning.” And Marisa Parham, a professor of English at Amherst College, observes that digital humanities does not just deploy e-instruments — it produces meaning “on the level of methodology and instrumentality.” To ask about the field is really to ask how or what DH knows, and what it allows us to know. The answer, it turns out, is not much.
Let’s begin with the tension between promise and product. Any neophyte to digital-humanities literature notices its extravagant rhetoric of exuberance. The field may be “transforming long-established disciplines like history or literary criticism,” according to a Stanford Literary Lab email likely unread or disregarded by a majority in those disciplines. Laura Mandell, director of the Initiative for Digital Humanities, Media, and Culture at Texas A&M University, promises to break “the book format” without explaining why one might want to — even as books, against all predictions, doggedly persist, filling the airplane-hanger-sized warehouses of Amazon.com.
ADVERTISEMENT
Digital humanities has a signature style: technophilia crossed with a love for its own neologisms. Data are “curated” rather than assessed; information is “leveraged”; facts are “aggregated” rather than interrogated. Significantly, those terms are never defined, but there is no question that they expect their readers’ diffidence.
No claims have been quite as exorbitant as those in “Quantitative Analysis of Culture Using Millions of Digitized Books,” a 2011 article in Science, whose authors include Jean-Baptiste Michel and Steven Pinker. The widely read essay, which inspired papers published in academic journals, boasted that its method, “culturomics,” unearthed data with “profound consequences” for the evolution of grammar and the study of fame, public memory, and etymology. It claimed to have discovered tens of thousands of words found in no published dictionary. Upon closer consideration, it turned out that the authors had mistaken morphological variants of known words for new words, speciously reasoned that what happens after an event is caused by what preceded it, discovered in the fame index what anyone in media culture already knows (that the famous are quickly forgotten), relied almost exclusively on Western libraries, and failed to define the term “book,” the central data-set of their study.
A similar shortfall is evident when digital humanists turn to straight literary criticism. “Distant reading,” a method of studying novels without reading them, uses computer scanning to search for “units that are much smaller or much larger than the text” (in Franco Moretti’s words) — tropes, at one end, genres or systems, at the other. One of the most intelligent examples of the technique is Richard Jean So and Andrew Piper’s 2016 Atlantic article, “How Has the MFA Changed the American Novel?” (based on their research for articles published in academic journals). The authors set out to quantify “how similar authors were across a range of literary aspects, including diction, style, theme, setting.” But they never cite exactly what the computers were asked to quantify. In the real world of novels, after all, style, theme, and character are often achieved relationally — that is, without leaving a trace in words or phrases recognizable as patterns by a program.
Perhaps toward that end, So, an assistant professor of English at the University of Chicago, wrote an elaborate article in Critical Inquiry with Hoyt Long (also of Chicago) on the uses of machine learning and “literary pattern recognition” in the study of modernist haiku poetry. Here they actually do specify what they instructed programmers to look for, and what computers actually counted. But the explanation introduces new problems that somehow escape the authors. By their own admission, some of their interpretations derive from what they knew “in advance”; hence the findings do not need the data and, as a result, are somewhat pointless. After 30 pages of highly technical discussion, the payoff is to tell us that haikus have formal features different from other short poems. We already knew that.
For all its resources, the digital humanities makes a rookie mistake: It confuses more information for more knowledge.
The digital humanities ignores a nagging theoretical dilemma: The interpretive problems that computers solve are not the ones that have long stumped critics. On the contrary, the technology demands that it be asked only what it can answer, changing the questions to conform to its own limitations. These turn out to be not very interesting. As often happens in computational schemes, DH researchers shrink their inquiries to make them manageable. For example, to build a baseline standard of what constitutes quality, So and Piper posit that “literary excellence” be equated with being reviewed in The New York Times. Such an arbitrary standard would not withstand scrutiny in a non-DH essay. The disturbing possibility is not only that this “cheat” is given a pass (the aura of digital exactness foils the reproaches of laymen), but also that DH methods — operating across incompatible registers of quality and quantity — demand empty signifiers of this sort to set the machine in motion.
ADVERTISEMENT
Distant readers are not wrong to say that no human being can possibly read the 3,346 novels that Matthew L. Jockers, an associate professor of English at the University of Nebraska at Lincoln, has machines do in Macroanalysis: Digital Methods and Literary History (University of Illinois Press, 2013). But they never really say why they think computers can. Compared with the brute optical scanning of distant reading, human reading is symphonic — a mixture of subliminal speaking, note-taking, savoring, and associating.
Computer circuits may be lightning-fast, but they preclude random redirections of inquiry. By design, digital “reading” obviates the natural intelligence of the brain making leaps, establishing forms of value, and rushing instinctively to where it means to go. Scour DH literature all you want, but you will find no head-on treatment of the theoretical problems that such methods entail. DH offers us powerful but dull tools, like a weightlifter doing a pirouette.
The outsized promises of big-data mining (which have been a fixture in big-figure grant proposals) seem curiously stuck at the level of confident assertion. In a 2011 New Left Review article, “Network Theory, Plot Analysis,” Moretti gives us a promissory note that characterizes a lot of DH writing: “One day, after we add to these skeletons the layers of direction, weight and semantics, those richer images will perhaps make us see different genres — tragedies and comedies; picaresque, gothic, Bildungsroman … — as different shapes; ideally, they may even make visible the micro-patterns out of which these larger network shapes emerge.”
But what are the semantics of a shape when measured against the tragedy to which it corresponds? If “shape” is only a place-holder meant to allow for more-complex calculations of literary meaning (disburdened of their annoyingly human baggage), by what synesthetic principle do we reconvert it into its original, now reconfigured, genre-form? It is not simply that no answers are provided; it is that DH never asks the questions. And without them, how can Moretti’s “one day” ever arrive?
Ted Underwood, the Illinois professor, finds “interesting things” when tracing word frequencies in stories but does not say what they are — a typical gesture in DH literature. Similarly, Alexander Galloway, a professor of media, culture, and communication at New York University, accepts that word frequencies in Melville are significant: “If you count words in Moby-Dick, are you going to learn more about the white whale? I think you probably can — and we have to acknowledge that.” But why should we? The significance of the appearance of the word “whale” (say, 1,700 times) is precisely this: the appearance of the world “whale” 1,700 times.
ADVERTISEMENT
For all its resources, the digital humanities makes a rookie mistake: It confuses more information for more knowledge. DH doesn’t know why it thinks it knows what it does not know. And that is an odd place for a science to be.
Given all of this, why the digital-humanities excitement? More than the obvious culprits — the fetish of science, neoliberal defunding — are at work. DH is at least partly a revolt of the academically disenfranchised. With shrunken hopes for a tenure-track job, younger scholars set out to make necessity a virtue, and instead of protesting their disenfranchisement by attacking the neoliberal logic of the university, they join the corporate attack on a professoriate that has what they want. The academic divide between poor and rich is effaced in DH initiatives, which have thrived in smaller, less well-funded liberal-arts colleges or second-tier research universities.
The resentment is sometimes frankly expressed. A staunch DH advocate, William Pannapacker, of Hope College, in Michigan, points out that the discipline’s enthusiasts define themselves outside the “status economies” of the profession, which they see as mostly having “chosen to exclude them from the tenure track”; instead, they opt for the “collaborative possibilities opened up by the internet.” Pannapacker sees DH as an insurgent tool to “liberate students from lectures and exams, and involve them in their own learning.” Who needs universities?
Already raised in a milieu of saturation social media, where the digital push and pull overwhelm slow-moving critical consciousness, many younger scholars find the disciplinary transition to DH almost natural. Some even begin to feel an affinity with opinion-forming elites in government and the media who seek to demote the stature of the “intellectual.” A revealing section title in a 1985 research paper from the CIA titled “France: Defection of the Leftist Intellectuals” reads: “There Are No More Sartres, No More Gides.” In a sharp reading of the paper, Gabriel Rockhill observes that the agency charted, with some satisfaction, how “the precarization of academic labor contributes to the demolition of radical leftism,” in part by putting a bull’s-eye on intellectual stars whose authority comes from their brilliance and theoretical finesse in heroically shattering corporate social codes.
The frequent appeal of DH to “crowdsourcing” and to “building user communities” in order to circumvent professorial authority sounds at first very democratic. But it has a dark side — weaponizing an already potent techno-discourse in the general culture and cutting the legs out from under innovative malcontents and nonconformists. The scattered aggregate of atomized individual contributions in supposedly “collaborative work” inevitably migrates to the safety of social norms. The crowd remains a crowd.
ADVERTISEMENT
Once no more than a slogan, DH now has many anthologies detailing its program, and almost always proposing DH, somewhat threateningly, as the future of the humanities, not just a distinct school within it. One of the best is Digital_Humanities (MIT Press, 2012). In a field of overstatements, this volume is more modulated and concessive, and so it is a fair gauge for the skeptical to examine DH in the words of its participants. Still, the book is relentlessly upbeat. Despite attacks on the humanities by cost-cutting administrators and the culture warriors of right-wing think tanks, the authors see DH as coming to the rescue. The anthology “envisages the present era as one of exceptional promise for the renewal of humanistic scholarship.” But to realize this outcome, the humanities must get in line by learning from communication and media design how to “create hierarchies of reading, forge pathways of understanding, deploy grids and templates to best effect, and develop navigational schemata.”
Once again, the knife-edge of the operation is epistemological: The authors envision a “generative enterprise” in which students and faculty alike “perform” research that goes beyond textual analysis, commentary, and critique, focusing instead on “cross-media corpora, software, and platforms.” The authors summarize the intellectual activities they promote: “digitization, classification, description and metadata, organization, and navigation.” An amazing list, which leaves out that contradictory and negating quality of what is normally called “thinking.” It would be a mistake to see this banishment of the concept as the passive by-product of a technical constraint. It is the aim of the entire operation.
Rather than a revolution, the digital humanities is a wedge separating the humanities from its reason to exist — namely, to think against prevailing norms. DH instead brings the humanities over to the outlooks of those administrators and legislators with programs that, on the one hand, put a scientistic gloss on method and, on the other, create a framework for lucrative tech deals in classrooms with the promise of the vast automation of teaching. The “results” of DH, then, are not entirely illusory. They have turned many humanists into establishment curators and made critical thought a form of planned obsolescence.
Timothy Brennan is a professor of cultural studies, comparative literature, and English at the University of Minnesota-Twin Cities.