Creating Technology for Social Change

The Politics of Verifiability

In “Too Big To Know”, David Weinberger correctly characterizes our pressing epistemic crisis. How do we know what we know what we know on the Internet? How do we create, locate, and trust that authority?

As Weinberger points out, the authority of print publications is in part conferred by the metadata of material scarcity. The cost of producing and distributing marginal copies of Nature tells us that someone must have had good reason (and backing). Material capital precedes reputational and social capital. As Clay Shirky has put it, in a world of print, the dynamic is filter, then publish; on the web, it inverts to publish, then filter.

Weinberger correctly notes that this epistemic crisis has always been with us. Knowledge always has been a social construction; what is true, and what is false, is what defines (and is defined by) a culture. The shift in material limitations has merely revealed the contest of construction. But the process of construction also changes when it shifts from the darkness to the light. What are the politics of this new age of knowledge construction?

Consider the case of Wikipedia. The author Philip Roth recently wrote in the New Yorker about opposition he encountered while revising an article about one of his books. Wikipedians reverted Roth’s edits on the rationale that his word alone was insufficient evidence. When Roth argued he, as author, was the last best source for information about his book, he was told that Wikipedia was not a repository for what was true, but rather a repository for what was verifiable.

Those who celebrate unreliable readers and the death of authorial intentionality were no doubt charmed by this response (and the subsequent revelation that Roth’s edits may have been wrong on the merits). But their pleasure may be premature. To paraphrase the Marxist critic David Harvey, Wikipedia never solves its authorial problems, it just moves them around geographically.

When Wikipedia says that it seeks not what is true, but what is verifiable, it does not actually discard truth. Instead, it disguises the idea of truth behind the mask of verifiability. Rather than rejecting authorial intent, it merely relocates it, by distributing it more generally.

This conflict can be seen in a close reading of the Wikipedia:Verifiable page. Wikipedians have struggled (admirably) with the difficulty of defining “reliability” in sources. Many of its considerations are legacies – almost intellectual skeumorphisms – from an era of costly printing, as when Wikipedia characterizes “self-published” material as “usually unreliable” because anyone can do it. Instead, Wikipedia admonishes its editors to rely on “established expert[s] on the topic of the article, whose work in the relevant field has previously been published by reliable third-party publications.” How do we know when a source is reliable? When it has already been reliable, which is to say, when it’s reliable all the way down.

Of course, there are exceptions to the reliability rule, and these come in the area of explicit opinion. Where criticisms and opinions are concerned, these need not come from established institutions. After all, it is verifiably correct that some Americans believe President Obama is a Muslim, and we can cite the primary sources to prove it.

This is where the politics of verifiability become more interesting. As the blogger Jay Rosen has noted in his commentary of the present Presidential campaign, the uncritical passing-along of statement is not a neutral act. A press which strives for an independent “truth”, as epistemically incoherent as that may be, will come to very different results, with very different effects, then a press which simply passes along statements which it can verify a campaign uttered.

As Rosen has pointed out, this dynamic can be gamed in different and interesting ways. For example, a clear substantive distinction (for example, whether to transform Medicare into a voucher system or preserve it more or less in its present form) can become deeply muddied, as when both campaigns accuse the other of planning to end it. But this muddiness is not neutral. The equal-but-opposite claims to “end Medicare” removes the policy differences from their pedestal and submerge them deeply in the muck of partisan squabbling. A potentially defining difference becomes recast as “just more partisan squabbling.” This displacement benefits whichever campaign would have “lost” a debate on the merits of plans by disarming its opponent of an advantage.

This is not to say that the consensus-driven (and driving) era of evening newscasters speaking in unison was some panacea. The “truth” then was just as false as it seems now; we are just able to imagine (unable to avoid seeing?) otherwise in an Internet era. The point is that verifiability has a politics too, a politics which both relies on some sites of traditional authority while it dismantles or dislocates others.

This entry adapts an essay for FAS 297 at Harvard.