Trust but verify: the problem of knowledge entropy

NPR has a photo essay, “The Jobs Of Yesteryear: Obsolete Occupations.” The story was researched and produced by Carolyn Beeler and edited by Tanya Ballard Brown of NPR, and it’s a lovely reminiscence for those of us old enough to have seen people doing these jobs.

However, the story exemplifies a problem with the transmittal of knowledge across decades and centuries. Knowledge degrades. There’s noise in the signal.

The NPR essay includes an entry for typesetters (click the arrow to the right of the photo until you get to it). Now that’s something I know about with a degree of intimacy that I would not expect from a photo archivist or a journalist. So it’s obvious to me that the descriptive paragraphs are full of misunderstandings and misstatements (never mind what they are, as the details are irrelevant to the point I’m making). Perhaps there are similar kinds of errors in the other descriptions, regarding occupations I know less about.

This is not a new phenomenon. I’ve been involved in events that were reported on the next day in a newspaper or the next week in a newsmagazine, and I’ve frequently had the experience, as you may have, of finding many factual errors and gross misinterpretations of events in the news story. I have no reason to think this discrepancy between fact and the reporting of the fact is larger or smaller today than it was a hundred or five hundred or a thousand or thirty thousand years ago. Surely you’ve read about psychological experiments involving the accuracy (or lack thereof) in eyewitness accounts of staged events. We’re not wired for reporting facts; we’re wired for remembering emotional states and rationalizing them by attributing them to things we think we saw.

I think this idea has begun to seep into the public consciousness, at least to the extent that literary critics now talk about the unreliable narrator as a feature not only of fiction, where it is an intentional device, but also of biography, autobiography, and memoir, where it is often not intentional at all.

However, I’m not sure the idea has fully penetrated to the extent that everything becomes suspect.

Over at Language Log, Mark Liberman writes often about bad science journalism, in which journalists jump to conclusions completely unsupported by the research they are reporting on. This is just one more area where we should question what we hear and see in the news (irrespective of how reputable a media outlet is or what political slant we believe it has).

But what about books? Books are edited, researched, fact-checked…some of the time. But there are also an overwhelming number of bad books, as there have always been (even before Gutenberg). And even in the most carefully done books, less-than-omniscient authors (all of them, in other words) write things they do not know from personal experience but read about or heard about somewhere else.

Well then, what about primary sources, the documents historians put so much trust in? Those, too, reflect a human interpretation and are subject to error.

What all this comes down to is the idea that maybe we shouldn’t be too sure we know what we think we know; we should be tolerant of others’ differing understandings (unless they’re total lunatics, of course). And just because we heard it on NPR doesn’t always mean it’s true.