Never Caught a Rabbit, Aint No Friend of Mine

The Daily Standard's Jonathan Last is disappointed with self-proclaimed media watchdog Corey Pein of Columbia Journalism Review:

THE COLUMBIA JOURNALISM REVIEW has a long, proud history of ignoring the story of the forged documents used by Dan Rather and CBS News. You'll recall that the scandal first broke on September 9, 2004, when a group of bloggers publicly questioned the validity of the four CBS memos. In the ensuing media scramble to get to the bottom of the story, blogs and big journalistic outfits such as ABC News, the Dallas Morning News, and the Washington Post took turns breaking news about the forgery, and then about the real source of the documents. (For a full history, click here.)

[...]

BUT NOW CJR'S flagship magazine has waded into the fray, albeit four months late, with a long analysis of the story by Corey Pein.

Pein's article, "Blog-Gate," posits, somewhat counterintuitively, that the lesson of CBS's "forged" documents is that the media are allowing themselves to be manipulated by a throng of right-wing bloggers. Says Pein, "on close examination the scene looks less like a victory for democracy than a case of mob rule."

The case Pein makes against bloggers rests largely on one point: That the CBS documents were not forged. Pein says that the memos "it turns out, were of unknown origin."

"We don't know enough to justify the conventional wisdom: that the documents were 'apparently bogus,'" Pein says. He adds, "We don't know whether the memos were forged, authentic, or some combination thereof." (Authentically forged, perhaps?) And finding proof for Pein may well be impossible. "The bottom line," he says, "which credible document examiners concede, is that copies cannot be authenticated either way with absolute certainty." Which suggests that, to Pein's mind, it is actually impossible to prove that the documents are forgeries.

Last's criticism focuses on the radical agnosticism of Pein's article, the epistemological equivalent of moral relativism, concluding with the biting punch-line:

So goes it at the Columbia Journalism Review. The university's motto may still be "In lumine Tuo videbimus lumen," but over at the j-school they have a new slogan: You can't prove anything.

But I think there is more to be said here. Pein is not merely sceptical about the possibility of knowing whether or not the documents were forged. He has seriously mischaracterized the argument that Newcomer and others involved in the debate were making. Consider this excerpt, which is the most coherent part of Pein's article:

Newcomer gave the press what it wanted: a definite answer. The problem is, his proof turns out to be far less than that. Newcomer’s résumé — boasting a Ph.D. in computer science and a role in creating electronic typesetting — seemed impressive. His conclusions came out quickly, and were bold bordering on hyperbolic. The accompanying analysis was long and technical, discouraging close examination. Still, his method was simple to replicate, and the results were easy to understand:

Based on the fact that I was able, in less than five minutes . . . to type in the text of the 01-August-1972 memo into Microsoft Word and get a document so close that you can hold my document in front of the ‘authentic’ document and see virtually no errors, I can assert without any doubt (as have many others) that this document is a modern forgery. Any other position is indefensible.

Red flags wave here, or should have. Newcomer begins with the presumption that the documents are forgeries, and as evidence submits that he can create a very similar document on his computer. This proves nothing — you could make a replica of almost any document using Word. Yet Newcomer’s aggressive conclusion is based on this logical error.

This would be a logical error if that were what Newcomer had been claiming but it was not. His claim was that Killian could not have produced the documents given the existing technology at the time. The fact that they could be easily produced on MS Word is not in itself damning, but it corroborates the theory that it was a modern forgery. By analogy, if the glove had fit OJ's hand, it would not prove the was the murderer, but it would have supported the prosecutions theory in combination of the other evidence. (Meryl Yourish evidently has a similar defense of Newcomer here, but her site seems to be overwhelmed by the Daily Standard's link.)

Pein then tries to support his agnosticism by calling into question the suggestion that the typography could not have been produced on period machines:

The specific points of contention about the memos are too numerous to go into here. One, the raised “th” character appearing in the documents, became emblematic of the scandal, as Internet analysts contended that typewriters at the time of the memo could not produce that character. But they could, in fact, according to multiple sources. Some of the CBS critics contend they couldn’t produce the specific “th” seen in the CBS documents. But none other than Bobby Hodges, who was Colonel Killian’s Guard supervisor, thinks otherwise. He told CJR, “The typewriter can do that little ‘th,’ sure it can.” He added, “I didn’t think they were forged because of the typewriter, spacing, or signature. The only reason is because of the verbiage.”

The problem is that no one has actually been able to reproduce such a document on a period machine. Jeff Harrell at The Shape Of Days did a thorough analysis of the IBM Selectric Composer typesetting machine that is the most viable possibility suggested and could not experimentally reproduce even a close approximation of the memo in question. The fact that Bobby Hodges or any of the other "multiple sources" thinks otherwise is not really impressive in the absence of actual evidence to support his assertion.

So the argument comes down to three basic propositions:

1. We can show that the documents could be produced on modern software.
2. No one has shown that the documents could have been been produced on period hardware.
3. Even if they could have been so produced, it is not likely that Killian would have had access to such hardware.

The most likely conclusion, therefore is that they are late forgeries. Would this argument satisfy Rene Descartes? Possibly not. But who cares? It isn't strict logic but an application of our intuition of what is likely that drives our day to day conclusion that we are not merely butterflies dreaming that we are men.

Some watchdog!

UPDATE: Yourish's site is now back up. In addition to the post noted above, she also responds to critics here. She isn't making precisely the same argument as I am, but it is well worth reading. Here is a brief excerpt:

When I first started in typesetting, Harper's Bazaar was still being set by a hot lead type shop in New York. I know this, because shortly after I joined Publisher's Phototype (now Applied Graphics Technologies), we got the Harper's account. It tooks us weeks of painstaking work to match our computer typefaces to the ones in the magazine.

I was assigned to the team that set up new accounts. I would sometimes spend an entire shift matching the type in a single article. I never—never—got a clean match on the first try. Nobody ever did. Matching type was and is the most frustrating, exacting, painstaking, time-consuming process that exists in any aspect of publishing. Imagine having to take the same few paragraphs and incrementally increase or decrease the spacing between characters, words, even between kerning pairs such as ff or WA. (If you ever want to torture me, just sit me down and make me match type. You'll get anything you want from me in an hour or two.)

Remember this when you realize that Charles Johnson typed the Killian memo into Word using the default settings and came up with a near-perfect match—the first time.

Her punch-line is also memorable: "Shoddy, slipshod research, false accusations, and ignoring important facts—isn't that what caused Memogate in the first place?" Heh.