Friday, March 23, 2007

Investigation and paranoia

I strongly second Ben's recommendation to see The Lives of Others. I also strongly recommend Zodiac: I saw it a few weeks ago and haven't stopped thinking about it. One of my friends was initially sheepish about how excited he was to see it: "I know it sounds weird, but I sort of had this serial killer obsession when I was in high school..." I think my version of it was my Tonya Harding obsession, which I admitted in similarly sheepish detail here. Back in 1994, I was convinced that if I could just read every piece of "evidence" about Tonya and Nancy Kerrigan, I could figure out the real culprit behind the knee-clubbing. It was a flawed premise, to say the least. The comments on that post from from Jenny D and Kate lead me to believe that it's not an uncommon belief to have about how obsessive reading will pay off.

Zodiac is a brutal movie about something like that problem. The scenes of violence are shocking, but they occur at the beginning of the movie, and the next two-plus hours are an account of how detection and information-gathering techniques do or don't work the way they're supposed to. I wasn't surprised that the ads for the movie show Jake Gyllenhaal mumbling that he "likes puzzles": basically, he's the stand-in for Fincher and the audience. Gyllenhaal's character, Robert Graysmith, isn't even the one to break the Zodiac killer's symbol code--a schoolteacher who does the daily crossword solves that puzzle--but the code stuff is a distraction. It's the micro version of the larger problem of how to interpret data that's incomplete or even irreconcilable.

Fincher, a notoriously exacting film director, even put himself into the role of obsessive investigator in making the film. He grew up in Marin County and remembers being afraid of the Zodiac killer's threats to kill schoolchildren. In a NY Times profile of the movie, he discusses this personal connection to the investigation:
But the source of his dark-hued lens on life, Mr. Fincher suggested, might be as simple as that original bogeyman. "It was a very interesting and weird time to grow up, and incredibly evocative," he said. "I have a handful of friends who were from Marin County at the same time, the same age group, and they're all very kind of sinister, dark, sardonic people. And I wonder if Zodiac had something to do with that."

Mr. Fincher was first approached about "Zodiac" by Brad Fischer, a producer at Phoenix Pictures, with a script by James Vanderbilt. It was based on two books by Robert Graysmith, a former San Francisco Chronicle cartoonist who became obsessed with the Zodiac, and who built a case against one suspect, now dead. Mr. Fincher said he wanted Mr. Vanderbilt to overhaul the script, but wanted first to dig into the original police sources. So director, writer and producer spent months interviewing witnesses, investigators and the case's only two surviving victims, and poring over reams of documents.

"I said I won't use anything in this book that we don't have a police report for," Mr. Fincher said. "There's an enormous amount of hearsay in any circumstantial case, and I wanted to look some of these people in the eye and see if I believed them. It was an extremely difficult thing to make a movie that posthumously convicts somebody."

Mr. Graysmith said Mr. Fincher's team found evidence that investigators had missed. "He outdid the police," Mr. Graysmith said. "My hat's off to them."

And of course Fincher's been down this road before in Se7en: Morgan Freeman and Brad Pitt take a bizarre side-trip to the library because Freeman is convinced that they'll find the killer if they can figure out what books he's read. It's an odd digression in the detective work that seems like it could only work in an unnamed rain-drenched city with artfully filthy windows and pristine Gothic libraries, although it does lead to the great line from Pitt about that smooth operator, the Marquis de Sade. Gyllenhaal's Graysmith uses the same trick of coordinating library check-outs in Zodiac.

I'm partial to any data-mining techniques practiced by young Robert Redford in Three Days of the Condor, but I'm skeptical of how often the gambit gets used in movies and books because it frequently seems to provide exactly and only the information that the detectives need. I know I'm supposed to suspend disbelief at these things--I always do in 24, where data-mining seems to take seconds--but it strikes me as an interesting epistemological question (no, really) about crime procedurals: how do we know what we know in the procedural? How do we distinguish good and bad information? Why do we believe that more evidence-gathering will lead to greater clarity? The great thing about Zodiac is that it takes up many of these questions: the film is so long because there's so much contradictory information to be sorted out, and there are significant lags in coordinating the investigations among the three police departments involved in the case. Owen Gleiberman mentions this fascinating problem in his review of the film.

Two other things: of course Zodiac has a fantastic cast, but I was struck by how many actors with instantly recognizable voices are in the movie. Robert Downey Jr. and Mark Ruffalo are obvious examples--they're reliably great--but I recognized the voices of Philip Baker Hall, Brian Cox, and Elias Koteas before I saw them on the screen (I think they're each introduced on the phone, but I'm not sure about that). The other is that Robert Graysmith also wrote the book that Greg Kinnear's Auto-Focus is based on; that movie has something in common with Zodiac in that it's about a man who's done in by obsessive documentation.

I got in the mood to see Zodiac by reading Ken Alder's The Lie Detectors: The History of an American Obsesion. It's pretty good: I was most interested in the last chapter, "Box Populi," which details how the lie detector has become a part of popular culture and we make certain uncritical assumptions about how information can and should be attained in investigations. In an earlier chapter he talks about how the lie detector was used to intimidate suspects, even as it didn't generate reliable data about truth-telling. And how would you really be able to tell whether it was successful in distinguishing a truthful statement from lie if there was no control variable?
The lie detector, it turned out, was not so much a thing-in-itself as a mirror that magnified the context in which it was used. Change the context, and the meaning changed. The machine could just as easily amplify the intimdating mob violence of the Ku Klux Klan as the sanctioned investigation of the model station house. Like any mirror (like any placebo), it was not so much an agent in itself as a question: What do you believe?

This problem recurs several times in the book, when Alder notes the lie detector's use to manipulate suspects in a few high-profile cases, but I would have liked to read more about that issue of knowledge production in a police investigation and a little less about the odd biographies of the inventors. That is, I think the biography section could have been streamlined, and the implications sections could have been fuller.

I have an ongoing research interest in what happens when investigations turn up noise that's interpreted as significant data. This problem shows up in one of my favorite scenes in the movie, when Graysmith thinks he's found a lead and ends up getting totally creeped out in a house that seems to be full of clues. Another way to think about it is what happens when the evidence turned up in an investigation turns out to be a mere artifact of the investigatory procedure itself (like Charles Kinbote's annotation work in Pale Fire)?

I've been thinking lately about Don Foster's Author Unknown, an account of his career as a literary detective, or "attributional theorist." Foster argues for his ability to determine authorship based on close analysis of literary style. He's used those close reading skills to determine the identity of Joe Klein, once-anonymous author of Primary Colors; to attribute a funeral elegy to Shakespeare; to determine the identity of an alt-weekly correspondent in Northern California whom some people believed was Thomas Pynchon; and to claim that Clement Clark Moore was not the author of "'Twas the Night Before Christmas." (I do love the Statistically Improbable Phrases for the book: among them are huge liar, funeral poem, William Gaddis, questioned document, anonymous document, mad bomber, talking points, and text archive.) In his review of the book (from the London Review of Books, archived at the Complete Review), John Lanchester argues that these investigations can pose as many problems about close reading as evidence as they bring up:
Still, the dedication started Foster off on his career. It led him to a funeral elegy of 1612, printed and published by the same team of Eld and Thorpe, and signed by one "W.S.". He became convinced that the (not very good) poem was a late, unattributed work by Shakespeare, and set about proving it. The difficulty was that all the evidence for the authorship was internal; it consisted for the most part of the poem's strikingly high incidence of Shakespearean phrases. When Oxford University Press turned down a proposal for a book on the elegy, partly on the grounds that internal evidence could not make a reliable case for authorship, Foster cheekily wrote to the anonymous author of the reader's report, on the basis that his style was indistinguishable from that of Samuel Schoenbaum. He was right, though no one was especially impressed by the stunt; Schoenbaum had the English Department secretary reply to the letter. When Foster finished the book, he submitted it, was turned down again, and pulled the same trick again, this time identifying the anonymous reader as Stanley Wells. Foster wrote to Wells, without letting on how he had worked out who wrote the reader's report, and Wells wrote back, thanking Foster for his letter and expressing surprise that his editor at OUP had let on who he was. "I wrote back, explaining to Dr Wells with imperfectly concealed glee that . . . it was by relying on methods employed in Elegy by W.S.: A Study in Attribution that I had established Dr Wells's authorship of that anonymous reader's report. Dr Wells was not amused." No surprise there. What is amusing is that Foster would think this piece of detective work might in any way advance his case over attributing the elegy. Wells was, as Foster points out, co-editing the OUP's new complete Shakespeare, and establishing him as the author of the anonymous report was close to a no-brainer; a very different business from establishing a highly contestable 17th-century attribution with no external evidence. Here as elsewhere, Author Unknown tells a story a little different from the one it thinks it is telling. The reader sees Foster as a much more bumptious, aggressive, disingenuous, insensitive, on-the-make figure than the country-mouse-cum-fearless-quester-after truth he presents himself as being.

There's also considerable debate about the Clement Clark Moore chapter, as Stephen Nissenbaum points out in this thoughtful essay from the Common-place. I can't remember whether Foster ever points out that some of these projects seem Pynchon-esque in and of themselves--less strange versions of Oedipa Maas' investigations into the Tristero mystery in The Crying of Lot 49. Here's this passage from the final pages of the book:
For there either was some Tristero beyond the appearance of the legacy America, or there was just America and if there was just America then it seemed the only way she could continue, and manage to be at all relevant to it, was as an alien, unfurrowed, assumed full circle into paranoia.