Experiencing the Bust

  • Posted on: 30 October 2017
  • By: mandell

Let’s imagine that Timothy Brennan’s article “The Digital-Humanities Bust”   1   is a poem in William Blake’s Songs of Experience. What would be the corresponding poem in Songs of Innocence? It could be any of the utopian essays that opened the digital revolution--candidates that come to mind are Ed Folsom’s essay in PMLA heralding “database” as a liberating genre. Many Blake scholars, but not all, think the “Experience” poems do not represent Blake’s final view--that the poems gesture toward a third state, a synthesis of Innocence and Experience. That version of Blake criticism also believes that the experience state of mind is caused by an original faith in innocence: that is, the cynical attitude pervading the Experience poems reflects disappointed hopes that were too idealistic in the first place: “For if all do their duty, they need not fear harm,” one Innocence poem intones, and we wish that were true--progressive politics can be seen as an active response to the failure of duty. (I remember, when lobbying for a feminist issue, a particularly conservative state legislator gazing at me in wonder that I expected anything to be fair, and of course we demand that it be.) The moral of re-situating Brennan’s article inside Songs of Innocence and Experience, however, is this: digital humanities is only busted if you expected it to be salvific.

Bringing Blake into the discussion should put pressure on the word “much” in Brennan’s question, “What has the Digital Humanities accomplished? Not much.” As good Blakean dialecticians, we ought to rejoin, “much what?” What if I were to stand up in front of Timothy Brennan’s, or even my own, print scholarship, lining up the books, and then asking them: what have you accomplished? Their answer will be, I hope, “it’s not that simple.”

To put this Blakean vision more bluntly and prosaically, Timothy Brennan has taken a view of digital humanities propagated among enthusiasts who see its grant-getting potential, attributed it to digital humanists themselves, a fallacy that has been addressed by Melissa Terras. Paradoxically, Brennan then attacks them/us--digital humanists--in exactly the same way that a conservative state legislator might attack a humanist: What exactly have you accomplished?

So let me describe the significant work that has been generated by the very people and projects dismissed by Brennan, and, after that, I wish to offer a bit of a diatribe against the new spectator sport of attacking DH--Brennan provides only one instance, so he is not the proper object of my mounting frustration, but I’m tired of “traditional” humanists discounting what "we" digital humanists are trying to do when in fact you, we, desperately need this work to be done. Can we take a break from “us” vs. “them” long enough to see what we have in common, or even, perhaps, try to work on some kind of synthesis? For “experience” people who see synthesis the cooptation of revolutionary energies, synthesis will not be an option, but I hope to convince you that it is the utopianists who deserve your justifiable critique, and not those DHers who are trying to build an infrastructure adequate to humanities research.

Brennan mentions Ted Underwood’s data-mining work as having accomplished nothing. For those of you who do not know Underwood’s work, he publishes regularly in major journals in the field of literary and cultural studies (PMLA, Representations, MLQ, New Literary History) as well as in all the major digital humanities journals (DHQ, Digital Scholarship in the Humanities [formerly Literary and Linguistic Computing], Journal of Digital Humanities). He also keeps a cutting-edge blog that makes major inroads into theorizing digital humanities, all the while accomplishing infrastructure work that benefits other researchers: extracting fiction from the HathiTrust dataset   2  , creating algorithms for correcting OCR, to name a few. His Stanford UP book Why Literary Periods Mattered is widely cited because it rigorously analyzes literary historical disciplines, without using any DH methods, but one could ask of it, “What has it accomplished?”

I will mention just one instance of Underwood “reading at scale” that I found very surprising and compelling, from “The Emergence of Literary Diction”   3  . Underwood and Sellers tracked the usage of Latinate words on the one hand and words of Anglo-Saxon origin on the other in field-specific discourses throughout the long eighteenth century. We all know that the gambit beginning in 1798 with Wordsworth and Coleridge’s Lyrical Ballads was to use “the language spoken by men” in poetry. Underwood and Sellers confirm that poetic diction becomes late in the century more Anglo-Saxon in etymology, therefore closer to “common language” in England, but also more specialized. The language of poetry deviates substantially from language used in published books and articles in other disciplines as well as general interest. That changes drastically my vision of the romantic “revolution in poetry,” knowing that “the language of men” turns out to be a kind of specialized discourse. I didn’t know that, did you?

An early foray into “cultural analytics” that Brennan does not mention was published by Brad Pasanek and D. Sculley in Literary and Linguistic Computing, “Meaning and Mining: the Impact of Implicit Assumptions in Data Mining for the Humanities.” Like Richard So and Hoyt Long whose Critical Inquiry essay is attacked by Brennan, their conclusion is not striking. After testing out socio-linguist George Lakoff’s theories from Moral Politics about liberal versus conservative uses of metaphor on Whig and Tory texts of the eighteenth century, they conclude, “our experiments do not position us to make any final pronouncement concerning Lakoff’s theories” (420). But just as the most interesting part of So and Long's essay is the "30 pages" of technical details that Brennan skips over, the conclusion of "Meaning and Mining" is not the most interesting part.

In reading Pasanek and Sculley, the conservative legislator / anti-DH diatribe could say, “you have accomplished nothing.” But “nothing” has only been accomplished if “something” is a scientific proof   4  . In fact, Pasanek and Sculley at first get results that confirm Lakoff’s hypothesis and spend most of their article describing the methods they used for attempting to determine whether other factors not imagined by Lakoff could account for the results -- the article tries to determine whether they got a false positive. And indeed, as they test for other possible factors distinguishing Tory from Whig discourse, those too give positive results. This summary is uninteresting because the details of what they tested are the accomplishment of this work, not the quotable conclusions.

Julia Flanders has argued, successfully in my view, that it is always the details that are interesting to humanists   5  . The goal of interpretation is precisely to confront head on wayward, ambiguous, sometimes conflicting, sometimes cohering details and to offer multiple hypotheses for a phenomenon’s overdetermined character. What I mean by that is that every human endeavor has many causes and effects, some conflicting, some not, and humanists work to trace and understand how cultural objects work.

If science reduces ambiguity to come up with universally applicable principles, humanities work is always impoverished if it is read only to reveal general laws. This is as true of digital humanities work as it is of work in the humanities: Frye argues that genres are composites of archetypes, Derrida that texts are undecidably undecidable, Timothy Brennan that African religious elements persist in contemporary popular music. These generalizations are only mildly interesting: the journey is what matters. The journey is what matters in digital humanities work as well, but the kinds of insights derivable from operationalizing hypotheses and assumptions in machine language are invisible if dismissed as “30 pages of highly technical discussion.”

And now for my diatribe.

Andrew Piper’s forthcoming book Enumerations refuses at the outset to participate in the “debates genre” that characterizes so much DH work. Over-use of this genre is explained by the fact that DH is always under attack, but it definitely warps arguments made by digital humanists, necessarily driving them toward discussing why DH is worth doing   6  . What would any scholarly monograph look like if it had to convince its readers that its own interpretations were worth making while offering those interpretations at the same time?

Can we answer the question, “Why do DH?” once and for all, and move on to the real work?

Katie Trumpener argued in the pages of Critical Inquiry against Franco Moretti’s reduction of books to title pages. She said, basically, that creating a database is a horrible way to look at books, that she instead looked at a few of them individually . . . by googling them. Google is a very, very large database. (And, may I just add, anyone who has read Trumpener’s Bardic Nationalism also knows that she has a database mind, capable of storing far more data than most human beings. For the rest of us, using a database as extended cognition only levels the playing field of scholarship.)

How the digital record is organized and preserved, and what tools are used to analyze it, database and search engine design, how cultural artifacts are findable when archived, how contextualized, how documented through metadata: these are all questions and problems currently under discussion in DH as well as currently being worked out in practical ways so that we do not need to depend solely on Google to mediate culture for us. If you don’t want to help, fine. But can you, instead of penning another anti-DH screed, simply step out of the way?

-Laura Mandell
Director, Initiative for Digital Humanities, Media, and Culture
Interim Director, Glasscock Center for Humanities Research
Texas A&M University
  1. "The Digital Humanities Bust" (content available only to Chronicle subscribers)

  2. All of Google Books plus ~500,000 more items held in libraries.

  3. Journal of Digital Humanities 1.2 (2012), [link]

  4. I am grateful to Philip Galanter for this insight.

  5. “Detailism, Digital Texts, and the Problems of Pedantry,” TEXT Technology 2 [2005]: 41-70.

  6. One exception to this warping effect: the magisterial Debates in Digital Humanities volumes edited by Matthew Gold and Lauren Klein [link]. The articles in these collections confront issues head on and consequently are not caught in the rhetorical stranglehold of having to do something and defend what they are doing at the same time.