Magnus and Sanger on Expertise

In my previous entry I commented on one of the articles in the Episteme Wikipedia issue. I thought it would also share my comments on the other two articles that were of interest to me. I read both of these under the influence of Collins and Evans (2008, hereafter "CE:"), which I have also mentioned here.

First, Larry Sanger's piece on The Fate of Expertise after Wikipedia is composed in two parts. First, the author responds to various interpretations of what he calls "The Wikipedia Potential Thesis" (WTP) whereby if Wikipedia fulfills its highest potential in terms of measurable quality, "experts would not need to be granted positions of special authority in order for humanity to have a resource that accurately tracks expert opinion." I think this is a bizarre thesis that no one has actually put forward. After some philosophizing, and given that Wikipedia is dependent upon expert (CE:contributory) knowledge, Sanger concludes this thesis is untenable. I agree. While Wikipedia might be sufficient in providing EC:interactional expertise (knowledge of -- not ability to do -- science) and might threaten other interactional experts (i.e., journalists) it would not obviate EC:contributory expertise. He also argues that Wikipedia is successful not because of anonymity, but because of its freedom -- permitting him to claim Citizendium is just as wiki-like and powerful as Wikipedia, but better in that real-name identities support community, governance, and quality. This is an argument he's made before, and one I largely agree with. Had Wikipedia started with the requirement that people login with an identity that corresponds to some real-world identity -- and this only need be policed in cases of abuse -- I think it would've done just fine.

Second, I most enjoyed P.D. Magnus' On Trusting Wikipedia. After reviewing literature on the reliability of Wikipedia, and arguing that Wikipedia is not like Britannica, the author posits five means by which reliability might be ascertained. The first three means correspond to types of meta-expertise in Evans and Collins: authority (reliable source; EC:local discrimination), plausibility of style (EC:technical connoisseurship), and plausibility of content (EC:ubiquitous discrimination). The second two have no direct corresponding type in Evans and Collins: calibration (testing a subset of the authors claims), and sampling (testing single claim with another expert, i.e., a second opinion). The author concludes that in the case of Wikipedia, none of these indicators are particularly strong. But I find his fault with authority (i.e., check your sources implied by WP:Verifiability) rather weak; he argues sources are unreliable, as are Wikipedia articles, since they are dynamic and can change. That is why one should use the permanent link (dated and versioned) when referring to something on the Web.

Comments !

blogroll

social