Horizon

Wikipedia

Both Cliopatria and Slashdot have had interesting discussions of the limitations of Wikipedia lately. I've been an occasional editor since early 2002, and done a bit of dabbling on the related Wikisource sites as well. In that process, I've formed some opinions about the strengths and weakness of wikis in general, and Wikipedia in particular.

1. Consensus: It might be reasonable for someone to assume that the open editing process encourages articles to be filled with inaccuracies — certainly the stories about Wikipedia that make the news emphasize this. This is rarely the case, however. In practice, a group of people who are interested in an article monitor changes to it on their watchlists, and intervene rapidly to revert vandalism and proofread additions. Their changes are in turn checked by the other editors watching the article, and so a sort of consensus builds around the article's content. This consensus will not permit changes that diverge widely from its editorial viewpoint to stand. For example, an edit to the article on Nineteen Eighty-Four that replaced any negative reference to Stalin's USSR with one about Roosevelt's USA was reverted within an hour.

   One problem with this editorial process is that because the watchers of an article are self-selected, if they are strongly in favor of a subject or strongly opposed to it, an article will drift in that direction. The articles on Noam Chomsky and Attachment Parenting follow a cycle in which a criticism is posted, a rebuttal to the criticism is added, then slowly the criticism itself is edited out. Perhaps more serious is the inability of the article's editors to make useful judgements about the factual validity of statements that do not directly contradict their editorial consensus. See, for example, my long and painful argument about Orwell's inspiration for Nineteen Eighty-Four to understand the limitations of an article's editors ability to judge the validity of certain types of statements.

2. Inaccessability of Sources: Many of Wikipedia's critics seem to be lamenting that an Appeal to Authority doesn't work there. There are no credentials, and editors who attempt to cite their own generally have weak arguments. This really ought to be a strength of Wikipedia; who wouldn't hope for an encyclopedia in which the most coherent synthesis, the most complete description of a subject prevailed?

   This often does occur, but the places in which it doesn't are telling. It's a commonplace criticism of Wikipedia that technical subjects and pop culture are well covered, while the humanities are not. Many critics attribute this to the nature of the internet's users, but I think that accessiblity is to blame. Old Star Trek episodes and RFCs are availible to anyone who has access to Wikipedia. Literature — especially non-fiction — is not. I've seen contentious articles packed with nonsense resolve themselves into decent essays when each editor had access to the relevant literature and was capable of judging whether the article's synthesis was valid. Unfortunately, making this literature availible required a Herculean effort. Perhaps the efforts by Google and others to digitize entire libraries will fix this.

3. Time-wasting Lunatics: Cranks do tend to get edited out of Wikipedia. Unfortunately, this is a never-ending job, and editors get tired of it. Worldwide access means that the lunatics appear in unexpected places. In the Code of Hammurapi article, I've mediated feuds between one crank who insists the Code proves the historicity of Abraham and another who believes redactor theory implies that the Torah was composed ex nihilo in the fifth century. Any article on the northern frontier of Ancient Near Eastern civilization will attract partisans from the Georgian/Abkhazian conflict, for whom the linguistic classification of Hurro-Urartean assumes the importance of our Confederate Flag disputes mixed with the English-Only debate. As none of these cranks are actually vandals, they can only be countered by appeal to literature and persuasion. After a bit of this, even a hobbyist like me begins to wonder if this is the best use of his time.

The three flaws I've discussed all stem from the limitations of the user community and its interactions. That said, one of Wikipedia's great strengths is its early formalization of appropriate conduct, and enforcement of that conduct. There are rules we all might remember when debating on the net: assume good faith, no legal threats, don't bite the newcomers.

I've pretty much quit editing on Wikipedia, and only check back occasionally on the subsections of articles I've participated in. I still feel the pull that first attracted me, though, and that pull is the desire to fix inaccuracies. A sentence in the Nineteen Eighty-Four article by someone apparently unaware of the Molotov-Ribbentrop act. The use of archaic Hebrew names for the verb stems in Akkadian language. They're all waiting, begging to be fixed.