Monday, May 19, 2008

Tyranny of the discrete

1. Anne Applebaum reviewed Nicholson Baker's new book Human Smoke for the New Republic last week. She takes issue with Baker's narrative method of writing short disconnected accounts of various people associated with the path to world war during the late 1930s. In Baker's book, these accounts are no more than a few hundred words long; his method is to let them aggregate into an implicit argument against World War II as a "good war." I'm not in a historian's position to assess Baker's argument about good wars or bad wars--other than to say that I don't think implicit (contrarian) arguments made by aggregation are the best way to write history. Applebaum doesn't think so either. One way of beginning such a critique would be to notice that Human Smoke looks similar Baker's other works. They proceed by minute repetitions of a narrative tic to create fractals of obsessiveness. By testing a perhaps untestable claim on so much primary source data, Baker gets an artifact of a repeated procedure with no good way to check the assumptions or what the repetition of the procedure is actually generating. That might be the way I'd look at the book.

Instead, Applebaum makes, I think, a bad move in connecting Baker to some larger zeitgeist of worrying about the decline of context and the rise of small chunks of unrelated data. She underscores her argument by mimicking Baker's box-car structure of discrete items which should, in Applebaum's mind, aggregate to show the folly of such a post-Gawker, post-Wikipedia arrangement of and attitude toward facts. Applebaum writes of Baker's short accounts:
Presumably, these items were selected because Baker finds them important, or perhaps because, like a Gawker post, they are meant to "turn conventional wisdom on its head." But ripped from their respective contexts, each item has the same weight as the next. There is no hierarchy, no sense that one enigmatic anecdote might be more important than the next equally enigmatic anecdote, or that one source might be more reliable than the next.

The entire last half of the article extends this argument to semi-ridiculous comparisons:
Yet the dull truth is that we arrived at the topic of Nicholson Baker not because we were talking about the war, but because we were talking about the contemporary cult of the non-expert, or rather the anti-expert: the bloggers who assume that the "mainstream media" is always wrong, the Wikipedia readers who think that a compilation of random anecdotes is always preferable to a learned study, and of course the college students who nowadays prefer to get their news in emails from friends because it is too bothersome to read a newspaper. And the even duller truth is that Human Smoke belongs to this cult, and not to the more exotic outer reaches of the historiography of World War II. One cannot properly understand Baker's book by comparing it to, say, Martin Gilbert's Auschwitz and the Allies or to the latest work on the fire-bombing of Dresden. To understand Human Smoke properly, one needs to read Gawker, Wikipedia, and above all The Da Vinci Code. The latter comparison might sound odd, but the resemblance is actually quite striking. Like Baker, the author of The Da Vinci Code is not a historian. And also like Baker, Dan Brown is a man apparently obsessed by his belief in the existence of a widespread historical conspiracy. (For those lucky enough to be unfamiliar with it, Brown's theory goes like this: the church hierarchy, along with the world's religious historians, art historians, and church historians, have been hiding the fact of Jesus's wedding to Mary Magdalene, as well as his subsequent children, from the public for centuries, using a massive cover-up perpetuated by Opus Dei, and so on, and on, and on.)

2. I first read about George W.S. Trow in his Gawker obituary, which noted that his most work, an extended essay published in the New Yorker in 1980 called "Within the Context of No Context," was their raison d'etre, in a weird way: "Trow's major thesis, that mass media and a cultural obsession with celebrity were ruining society as we know it," the editors at Gawker wrote, "is borne out pretty much each day on this website and every other." Trow's essay is an aggregation of discrete comments about television, print, and decline-decline-decline. ( The New Yorker has an excerpt available online which gives an idea of how the structure works.) The style was nothing weird for the New Yorker in those days; that Trow's work from 1980 about the decline of history looks so similar to what Applebaum is criticizing is one clue that decline narratives have the generic tendency to look to newer forms of technology as a reason for the decline. Trow names his reason for decline:
The work of television is to establish false contexts and to chornicle the unraveling of existing contexts; finally, to establish the context of no-context and chronicle it.

I became obsessed with Trow's essay because I think the eighteenth century is a great century to think about context--in literary miscellanies, newspapers, animadversions where quotations are assessed in new contexts of critique and selectivity, and novels, and secret histories of gossip removed from context (gossip is another favorite subject in Trow's work: "Gossip is small, shameless history. It sets out to tell the trivial about the great or about those connected to the great. It thrives on awkwardness--because it assumes dignity somewhere: 'Somewhere else, you're getting another story,' gossip says with a knowing look, 'but this is what you wanted to know.'"). How does the genre of the secret history in the eighteenth century look similar to or different from the conspiracy novel as Applebaum is describing it in Dan Brown?

But different forms of mediation have their own, unique issues of context, so I wanted to see what kinds of claims I could make about technological change and worries about decontextualization. Print, television, and digital media have different ways of creating and recreating context. Applebaum seems to ignore this issue of mediation in making such a leap between Baker's book form and her criticism of Gawker as medium for the short attention span.

That is, I think Applebaum is right that the Internet has changed the way that people think about doing research--indeed, how people think in general--but I don't see it as the same decline narrative as she does. Nor do I think Human Smoke, a booky book about print archives if there ever was one, is the best text to try out that argument about technology and thinking.

3. Indeed, Trow does historicize his essay in the introduction to the book version of Within the Context of No Context, in which he discusses his family's "demographic history" in establishing the first address directories in New York, the Trow directories. His family had always been in the print business:
I first encountered the world of printing in 1952 in New York City in the City Room of the New York Post, where my father worked... As to the City Room of the Post, I found it unimaginably powerful, unimaginably netherworld in its implications, and I felt it belonged to me... In the Composing Room, a man from a Friendly Hell sat at a dangerous machine and turned my father's text into a metal object. It was a process ancient and modern both. Back in the City Room, I could feel a vibration in the floor. My father took me, then, to see the presses which in their violent turning set the whole building ahum. Those were rotary printing presses; the rotary printing presses had made the modern daily newspaper possible.
...I realize now that I am a man from a broken tradition who was convinced by the theater of a moment that his tradition was unbroken and that he was heir to it. Working as though from an unbroken tradition I have made sense of my life by developing an ability to analyze Mainstream American Cultural Artifact, and this I also urge my fellow citizens to do.

4. I have some affection for Nicholson Baker. Not the phone sex, the John Updike stuff, the shoelaces, but he did write one of my favorite New Yorker short stories ever, "Subsoil," a horror tale set at a potato museum. When I read his essay about Wikipedia in the New York Review of Books this spring, I was really charmed. I thought, There is no one else who should be editing Wikipedia but Nicholson Baker. Here is someone whose temperament is directly suited to such a task: the obsessive attention to additions and deletions, the mind that can range over the very weirdest things and home in on small details, only to repeat that procedure at another entry minutes later. It works on Wikipedia! Anne Applebaum may be the only person who read that article and wasn't won over.

I also like him because he likes newspaper archives. Too much, perhaps, and he doesn't see the possibilities of digitization of printed material. Or maybe he would like it better than microfilm--anything's better than microfilm. Discussing his previous book project, The Double Fold: Libraries and the Assault on Paper, he told a reporter at the New York Times that he had perhaps applied his obsessive personality to an unmanageable project when he decided to buy the collections of late 19th and early 20th century American newspapers which were being deaccessioned in favor of microfilm or digitization.

5. With Trow and The Double Fold in mind, I think the better explanation for Baker's project is not Wikipedia or Gawker-based thinking, but his belief that print, specifically newsprint, is a medium for documenting history may be on its way out. Or at least challenged in the hierarchy of mediation. The project originated from his time in newspaper archives working on projects related to The Double Fold. He writes of his research methods for Human Smoke in the afterword:
This book ends on December 31, 1941. Most of the people who died in the Second World War were at that moment still alive. Was it a good war? Did waging it help anyone who needed help? Those were the basic questions that I hoped to answer when I began writing. I've relied on newspaper articles, diaries, memos, memoirs, and public proclamations, each tied as much as possible to a particular date, because they helped me understand the grain of events better than the secondary sources did.

He elaborates on his research methods and hopes for reader responses in an interview with Baker in the Nation:
It's up to you to rescue it! Let's see. I left out a good deal of what happened in Russia and what happened in France. And even in the draft that I had, when my wife read it, she reached her limit with bloodthirsty statements from Winston Churchill. She actually was in a kind of rage, and she said, "This page has to go!" She was so right. But I think that anybody who wrote a book like this would do it a different way. People used to have books rebound, and they would double in size, and they'd put a blank page between each page. And then they'd put in their own things. This book is all written using libraries. It's not using any secret sources, anything unpublished. Anybody who wants to go ahead and fill in pages between my pages or cross out my pages should, because I think that a war this important to the history of humanity needs lots of amateur historians.

6. There are some interesting assumptions in Baker's afterword: Is it possible to write a history with those two questions as the framework? Why would looking at newspaper articles in aggregate help answer those questions? I hope it's enough to say that asking them reflects my deep skepticism for the project. I'm thinking of the taxonomy of errors--from moralizing history, to histories of "how x was a good thing," to overly selective research base--that David Hackett Fischer set out in his classic book from 1970, Historians' Fallacies: Toward a Logic of Historical Thought. For one thing, what if a journalist's sense of periodization can serve as its own form of context-of-no-context in a larger work of history:
Another kind of periodization error is one in which the historiographical medium provides the method. A classic example is that species of quasi-historical writing which appears in newspapers and newsmagazines. Academic historians are often contemptuous of the historical interpretations of journalists--and properly so. Those failings are attributable not tho the cultural barbarism of the fourth estate, but rather to the scheme of periodization which is forced upon it. Time, for a newspaperman, is measured in the intervals between editions. His often desperate effort to find some significant happening in each of these periods explains his shallowness, rather than ignorance or illiteracy or the company he keeps.

(No one escapes unharmed from Fischer's putdowns--but I think it's an interesting possibility to consider in light of Applebaum's charge about blog-thought: I remember writing plenty of history papers in college which were structured around reading print runs of newspapers and doing close-readings of the changes from week to week or year to year. I never really thought about any other way of doing it, and anyway I was obsessed with whether the next day's issue of the Spectator would actually get done.)

7. There's also an interesting question of mediation in Baker's evocation of old forms of books which allowed readers to recombine materials into new forms. Indeed, if Baker has delightfully embraced Web 2.0 Wikipedia, he still evinces many more habits of a print archivist from The Double Fold. The story of books and reading that he's telling in the Nation interview is an archivist's fantasy of recombining printed materials into new forms, finding new pieces of paper, and ordering them in a book. It's just not the same form of mediation as the Internet--nor is it really the same as the conspiracy novel as Applebaum described it. So you could ask, what does the Internet do to our procedures for checking facts, but I don't think Baker was doing his research on Wikipedia--he was doing it with a hedgehog's focus on a very small set of print archives.

Baker's relationship with his research base is constituted on repeating the same procedure of constructing a brief account--which owes something in its construction to a news clipping (not a blog post). J.D. Marshall makes a similar point to Fischer about research bases in his attack on local histories in England, The Tyranny of the Discrete:
If the student remains absorbed in original sources, he or she will be content to reproduce information from these sources, which he or she will regard as having special historical validity. In other words the information itself will be come a substitute for history; the discrete fact itself becomes pseudo-history.
Students ... are more likely to develop the notion that crucial or important historical information comes from manuscript sources only. This leaves them unable to appreciate an argument which uses a variety of different sources to provide evidence. They are also prone to develop a prejudice to the effect that facts are more important than arguments, and the verification of facts represents almost the sum total of local historical activity. ... Local history, like any other kind of history is meaningless without coherent, immediate, imaginative and above all telling context, an it is precisely this which is lacking in many museum collections and their presentations.

8. The thing that was striking to me about Baker's book was that the majority of the anecdotes began with the name of a person. His interest is always at the local level. That seems like a comfortable level for someone who's interested in iterations of small procedures, but what if one zoomed out to see the fractal pattern at a different scale? I recently read about Lewis Fry Richardson, an English Quaker mathematician from the mid-20th century who devoted the second half of his career to developing statistical models to study arms races and world conflicts. He had started his career as a mathematician at the British Meteorological Service and founded the field of computational weather prediction. In his book, After Thought: The Computer Challenge to Human Intelligence, James Bailey spends several pages discussing Richardson's vision for statistical modeling of data which had never been crunched at high volumes before. For example, to develop mathematical formula for weather forecasting, he search through volumes and volumes of weather information from the nineteenth century in order to compile enough information to refine his statistical measures. He then moved on to study world conflicts from the nineteenth century in Generalized Foreign Politics: A Study in Group Psychology. In the introduction, he rejects the day-to-day level of analysis in favor of a much larger scale:
Foreign affairs as they appear day by day in the newspaper: the text of the despatch, the facial expression of the ambassador as he comes away from the important interview, the movement of warships, these may be likened to the eddying view of the wind. Whereas the theory here presented may be likened to an account of the general circulation of the earth's atmosphere.

So that's another way to write a pacifist history, although Bailey notes that Fry's attempts to crunch everything into prediction curves on graphs was not entirely successful and failed to predict situations such as the Arab-Israeli arms race or the fall of the Soviet Union. Elsewhere in Historians' Fallacies, Fischer has many words for cliometrics and politico-metrics of the sort that Fry wanted to innovate. But what interests me about Fry's work is that commitment to a too-large scale; he's like the bizarro version of Baker.

9. So one returns to scale, between the discrete and the demographic, as the problem in the context of no context, exactly as Trow imagined it:
That movement, from wonder to the wonder that a country should be so big, to the wonder that a building could be so big, to the last, small wonder, that a marketplace could be so big—-that was the movement of history. Then there was a change. The direction of the movement paused, sat silent for a moment, and reversed. From that moment, vastness was the start, not the finish. The movement now began with the fact of two hundred million, and the movement was toward a unit of one, alone. Groups of more than one were now united not by a common history but by common characteristics. History became the history of demographics, the history of no-history.

History had been the record of growth, conflict, and destruction.

The New History was the record of the expression of demographically significant preferences: the lunge of demography here as opposed to there.

Labels: , , , , , , , , ,

Blogger Peter on Thu May 22, 12:47:00 AM:
Great post. But -- small detail -- "Double Fold", not "The Double Fold".