National Security

Wed, May 14, 2014

“The Real Nature of Control”

The last text I assigned in my recent “Modernism, Fascism, and Sexuality” seminar was Gravity’s Rainbow.¹ Among its many oddities is a scene where the spirit of Walther Rathenau is summoned through a medium for the entertainment and mockery of an elite “corporate Nazi crowd”:

These signs are real. They are also the symptoms of a process. The process follows the same form, the same structure. To apprehend it you will follow the signs. All talk of cause and effect is secular history, and secular history is a diversionary tactic. Useful to you, gentlemen, but no longer so to us here. If you want the truth—I know I presume—you must look into the technology of these matters. Even into the hearts of certain molecules—it is they after all which dictate temperatures, pressures, rates of flow, costs, profits, the shapes of towers…²

Rathenau, or as we must rationally conclude, the inventive medium Peter Sachsa, ends with these somewhat ominous questions: “what is the real nature of synthesis?” and “what is the real nature of control?” It was many years ago when I wrote an overview of the critical reception of Gravity’s Rainbow for an undergraduate thesis, and it was lonely work for a student with a VAX account. Even then it was clear that this scene had engaged many critics’ attention for its seeming encapsulation of the novel’s many threads.

Among those threads is cybernetics. The idea of an automated and self-propagating technology of control, independent of its computational substrate, lurks in the margins of Gravity’s Rainbow. It’s also interesting that a statistician becomes a member of the “counterforce” in the text because of his distrust of an orthodox behaviorist.³ I found myself wondering whether Pynchon was aware of the then-recent intellectual history of behaviorism and of how the linguist more responsible than anyone else for its demise was also notable for a distrust of empirical data-gathering and statistics himself.

I find it hard to imagine that Pynchon had not read Chomsky’s “The Responsibility of the Intellectuals” and other writings where he had described the university’s involvement in counter-insurgency and propaganda for the U. S. government. He may well have known that Chomsky’s work in devising a computational and rationalist theory of linguistics had also been funded in part by the military-scientific research industry they both so forcefully criticized.

DH’s Secular History

I was thinking about the real nature of synthesis and control while reading through the recent issue of differences devoted to the digital humanities. Or, rather, to the shadows of the digital humanities.⁴ Some of the papers were derived from those presented at the “Dark Side of the Digital Humanities” conference held in Milwaukee in 2013. I was able to read David Golumbia’s article in draft,⁵ and I hope to have some more to say about it in another post. Brian Lennon’s article, “The Digital Humanities and National Security,” argues that there is a dialectic in philology between historical humanism and rationalism, which he finds in Raymond Llull’s “combinatorial unilingualism.”⁶ The rationalist strain of philology inspired cryptological and quantitative approaches to literary study in the late 19th C. Prominent among these disreputables were Baconian cipherists, and Lennon suggests that the conservative temperament of these anti-Stratfordians (along with their rationalist dispositions) affected the development of the intelligence services in the 20th C.

In particular, the utility of literary studies for the national security apparatus grew directly from a mindset favoring “aggregation and documentation” over interpretation and critical debate.⁷ Lennon quotes from an MLA presidential address by John M. Manley in 1927, and it’s a good find. One of the things that JSTOR and other databases have helped with, often in unacknowledged ways, is eliminating the prejudice against the dusty and unindexed tome. I also like to think that the profession is less Whiggish than it was even when I was an undergraduate, and the database deserves much of the credit here. Lennon next follows Robin Winks’s Cloak and Gown⁸ and some other sources on academic complicity with the OSS and later CIA through the first post-war decade. Vietnam, as I alluded to earlier through the example of Chomsky, resulted in some pushback against this widespread integration of academic research with the state security apparatus; but Lennon follows Winks in claiming that this resistance had largely subsided by the 1980s.⁹

In the third section, Lennon turns to the puzzling question of why digital humanists have been reluctant to historicize their research. The question is less puzzling when you consider that almost no digital humanist or humanities computer would recognize or accept Baconian cipherists¹⁰ as legitimate ancestors, though I could see Raymond Llull (and Leibniz) being more congenial to those inveterate combinatorists who haunt us even still. Fewer would consider themselves cyberlibertarians, though this identification is much closer in time and conceptual space. Gloating about catching teachers cheating on standardized tests through statistical analysis and concluding from this exercise that more standardized testing is called for is what I see as the guiding model for some digital humanities research.¹¹

“The Shapes of Towers”

There are no proverbs for paranoids sufficient to navigate the surveillance state we now inhabit. A ready example: a sociologist and internet researcher recently documented her efforts at hiding her pregnancy from the internet, which required the use of anonymous browsing technology (Tor) known to be penetrated by intelligence services (if not yet the marketers) and eventually resulted in a report of suspicious economic activity.¹² We fill pre-existing internet forms, as the saying goes, and when we fill them we change them and are changed.¹³ This filling depends on machine learning and textual analysis techniques whose methodological assumptions are often at odds—if not outright inimical—to those of literary analysis or cultural studies.

So, what happens when literary scholars begin to experiment with these technologies?

Can a ballyhooed turn in the humanities, especially in literary studies, that promotes a putatively novel computational textual analytics including textual and other data “visualization” possibly be or remain isolated from the cultural-analytic and specifically textual-analytic activities of the security and military intelligence organizations that are the university’s neighbors—especially when such a turn is represented as a historic opportunity made possible by historic advances in information technology?¹⁴

Lennon’s answer is “It seems unlikely.” I have myself dabbled in computational textual analytics, even going so far as to answer a CS graduate student’s question about a variant of the LDA algorithm.¹⁵ Though I toil in deserved obscurity, I follow more visible efforts with attention and compassion. I understand the “historic opportunity” rhetoric of mass digitization, and I’m legitimately excited by the potential for disciplinary self-study and generic evolution afforded by the digital archive. Citation analysis is transparently invoked by those who seem to despise humanities scholarship as evidence of its inconsequence, and yet I have spent a great deal time creating co-citation network visualizations.¹⁶ Network analysis works wonders for counterinsurgency and population control. Kieran Healy’s humorous post on Paul Revere¹⁷ illustrates this well.

I am tempted to answer Lennon’s challenge with an appeal to technological neutrality.¹⁸ Topic-modeling is an information-retrieval technology that offers a higher-order of search capability. No one needs this more than the intelligence agencies, just as they would benefit immensely from automated translation. The ominous role of machine translation and the way that it interacts with the discipline of comparative literature in particular is the subject of much of Lennon’s research, and I can see how the argument would carry over. A sustained analysis of a “Digital Humanities Questions and Answers” post about the ethics of accepting military research funding follows in Lennon’s argument. He notes that a forum respondent mentioned a statement by anthropologists concerned over their research being used in counter-insurgency contexts. The relative lack of concern about this issue compared to the anthropologists is taken as further evidence of the complicity of DH as a research practice with the larger national security state. It’s not clear to me that the situations are directly comparable, however, largely because of DH research being mediated through the more technical fields that receive massive amounts of funding for these purposes.

A Nice Future

In his final section, Lennon makes the piquant claim that the “inability to define ‘digital humanities’ means that anyone willing to be sufficiently cheerful in the act can don the digital humanities hat at will.”¹⁹ Here is a dig at the legendary “niceness” of the DH community, which sometimes is perceived as passive-aggression to those who do not feel wholly a part of it. At least other facets of the humanities are forthright about their elitism, perhaps someone has thought at one time or another. He counsels DH “enthusiasts” to turn to comparative literature and its debates in recent years as a guide to understanding their present historical dilemma.

Though it is not mentioned in this essay, Lennon has sometimes referred to a Harvard Business Review feature²⁰ that documents decreasing job prospects for information sector workers, defined broadly. I think that Lennon attributes this to progressive (perhaps “recursive”) automation in the industry, which will gradually reduce labor demand in this area. I suspect that outsourcing is more immediately relevant as an explanation for this trend, but the overall point is worth contemplating. It is relatively easy to imagine a turn away from computation as an engine of cultural and economic activity, though the biological turn which is the most easily imaginable substitute would be derived from a heavily computational science itself. What does the post-computational future look like? Does philology (or any other branch of cultural analysis associated with the humanites) exist in it?

These questions may seem a bit ridiculous, but I remember hearing Katherine Hayles predict that the traditional English Department would go the way of Classics. That was fourteen years ago, and I admit that I found it an implausible claim at the time. But perhaps the process was already underway.

Notes