You are viewing a monthly archive page.


On Wikipedia and "The Truth"

wikipedia logo I’ve been thinking a lot lately about Truth. The impetus was John Siracusa’s two podcasts about Wikipedia, in which he argues that Wikipedia is fundamentally flawed, and that its flaw prevents it from being as good a source of knowledge as it might otherwise be. The crux is Wikipedia’s insistence on verifiability for facts that it includes (not that this rule is rigorously enforced) over what Siracusa calls “the truth.”

These two podcasts are about the closest I’ve ever heard a “normal” discussion come to facing the fundamental questions of epistemology. In the second of the two episodes, Siracusa addresses complaints from the first episode, who repeatedly use the word “verify.” Siracusa’s understood this word from Wikipedia’s perspective, where it often means “to cite another authority.”

I think that the folks writing to Siracusa meant “verify” to mean simply “confirm.” Fundamentally, that’s the problem with first-person accounts—they’re unconfirmed. But, you might say, can’t two first-person accounts confirm each other? Siracusa suggests a system of persistent identities and a system like Stack Overflow’s reputation system. I’m not sure, but I believe that system works because the scope of the site is fairly limited. What if someone on a Wikipedia-like site of knowledge has a high “reputation” earned by writing about software, and suddenly edits an article about politics? And if first-person accounts are allowed, wouldn’t most of the site tend toward being first person accounts (since Siracusa’s whole reason for talking about this is that he wants first-person accounts to be left in)?

Let’s look at an example. I was just looking at the Wikipedia page for John Gruber. It cites an interview with Gruber, along with other sources, for facts that are general common knowledge. They’re common knowledge because Gruber’s website is widely read, and those facts have appeared there. Why not just cite the blog directly?

This gets exactly to the heart of what “truth” means. Siracusa kept saying “Wikipedia is not trying to get to the truth!” But I think that’s putting it too strongly. Wikipedia is after something like “the consensus of human knowledge,” which is a reasonable definition of “truth,” I think. Sirucasa is after the actual, out-there-in-the-universe version of truth. But the problem is that that truth is unknowable. Until Galileo discovered that the Earth is not the center of the universe, there was no way to know the truth about it. What’s more, even after he made the discovery, the truth was knowable only if you believed him. If you found his arguments persuasive. Two kinds of people would have found Galileo’s arguments persuasive: those who had an understanding of science enough to see why he was right, and those who are just inclined to believe bizarre or unusual things they hear: people with a low threshold for truth. The fact that people who understood the science and knew that Galileo was correct had a vested intrest in not saying so doesn’t change the fact that the only way for a member of the general public to come to believe Galileo was for the experts to come around and admit that he’d been right all along.

And so it is with all non-trivial truth: we rely on authoritative sources to tell us which sources to believe. Is global climate change real? The consensus of climatologists is that it is. Does government stimulus policy decrease the effects of recession? The consensus of economists deems it so. Are ruffles the new black? A consensus of fashionistas says yes. “The consensus of human knowledge” is by definition the closest we can get to “the truth” in areas of human knowledge where we’re not experts.

An objection to this is that any person has the theoretical potential to become an expert in any topic. If something is controversial, couldn’t Wikipedia include both sides of the controversy along with the arguments for each? Two reasons to think this might not be a good idea. First, the reason Wikipedia exists seems to me to be to provide general information for non-experts. We can expect Wikipedia to provide you with the basics of the science of climate change, for example. Seems a bit much to expect it to be able to turn you into an expert on the subject. Second, if Wikipedia contains not just what can be cited (“verified,” “confirmed”) but anything and everything that is true (read: anything and everything that might be true), its entries will have no check on their growth. For example, the article on 4chan has over 5000 words and 136 citations. How long would this article be if un-cited facts were allowed? Who’d want to read more than 5000 words about 4chan, and should the level of sort knowledge they’re after be collected, for all possible topics, in one place on the internet. Maybe.

I’m hinting at how it is I think the rules of Wikipedia might have come to be what they are. I think intelligent people got together and spent a lot of time thinking, discussing, and experimenting with these rules, and Wikipedia is the consensus on which they arrived as the best vehicle for arriving at The Truth. It’s, as Sirucasa said, the worst way for arriving at the truth except for all the others. Maybe. I think so, anyway.

Posted: Thursday February 9, 2012 by Alesh Houdek · Permalink · Comment [2]


I don’t think I’m not going to make a whole big post about this every week going forward, but of course the new episode of On The Fence is out today. Go listen!