I just finally got around to reading my June 2006 Communications of the ACM (an academic computer journal) and spotted a little news brief about Britannica trying to sue Nature magazine for this December 2005 Article that noted that the error rate in science entries of Wikipedia is comparable to that in Britannica.
Nature magazine is the premier peer-reviewed high-level publication for science in the world. For an article to get into this journal, it has to be rigorously supported. Their discussion of the controversy is here.
One point about the coverage I hadn’t seen was the different natures of how the information was put together in each of these encyclopedias.
- Britannica is Intelligently Designed. For a dozen decades they hired the best and brightest crafters of techinical prose to cover all the subjects that the intelligent editors can think of as being worthwhile. Yes, the design is always evolving. But it is limited by the need to have a designer watching and controlling its growth and changes.
- Wiki evolved from chaos. It provided an environment, and anyone was invited to add and edit entries. It covers millions of more subjects than Britannica can hope to. It cross references virally, with apparently absurd links to subjects that someone thought were related, and often do turn out to be relevant. People do go in and edit and even mark content to remove. There is some oversight to prevent ideology from trumping understanding. But basically, it is an example of pure evolution with no plan or guidance.
So my point is that over a hundred years of intelligent design has produced something that, in its limited intersecting subset, is slightly better than the half-decade of raw evolution produced.
Okay, the Britannica only consults experts, and Wiki entries are written by everyman. But usually knowledgeable amateurs or even parties to the discoveries themselves. It’s only a century or so since that difference was moot. Chas. Darwin had a degree in Divinity. R. Feynman made a study of the behavior of a Frisbee, and got a Nobel when he applied the math he worked out for flying disks to electron behavior. Interested amateurs often become closet experts in fields for which they have no papers. But, I digress. So here I’ll stop.
"Britannica only consults experts, and Wiki entries are written by everyman."
It seems to me that a 2003 book written by James Surowiecki, The Wisdom of Crowds, might be relevant. Surowiecki argues that under the right circumstances, groups are remarkably intelligent-they are often smarter than the smartest people in those groups.
Surowiecki presented evidence that a group made up of some smart people and some not so smart people almost always did better than a group comprised of experts only. (Page 30).
Surowiecki calls this phenomenon "the wisdom of crowds." Crowds do not really tend to be dumb or crazy, as they are portrayed in popular literature. In Surowiecki's opinion, collective intelligence committee brought to bear on a variety of problems. The first is cognition problems (those that have definite solutions, such as who will win the Super Bowl or how many products will this company sells next month). The second problem is a coordination problem (those that require members of a group to figure out how to coordinate their behavior with each other). The third kind is a cooperation problem (getting self-interested people to work together, such as paying taxes are dealing with pollution).
What characteristics should a crowd have an order for it to be an effective problem solver?
In his book, Surowiecki describes a wide variety of problems in which the semi-ignorant crowds have been more accurate than the experts. The examples range from determining where to find a sunken submarine to estimating the weight of an ox.
I haven't yet made it all the way through Surowiecki's book, but perhaps the system set up by Wikipedia turbochargers the wisdom of the (often non-expert) crowd, accounting for the surprising power of Wikipedia.
As to group decisions, here's a <a title="wiki Grook" href="http://en.wikipedia.org/wiki/Grook">Grook by mathematician <a title="wiki Piet Hein" href="http://en.wikipedia.org/wiki/Piet_Hein_(Denmark)">Piet Hein:
I admire Dan's suggestion that the creationism-evolution controversy is analogous to the Brit-Wiki controversy. Nevertheless, there's an important difference: the creationism-evolution debate involves one "expert" (God) versus no experts (evolution), whereas the Brit-Wiki debate involves a small group of experts (Brit) versus a large crowd of semi-experts (Wiki). For the Brit-Wiki sitution to be truly analogous, we should not choose creationism versus evolution, we should choose Christianity (one god) versus, for example, Hinduism (a large crowd of semi-gods). Erich's comment would then suggest that a crowd of semi-gods (the Hindu model) would out-perform a single god (the Christian model).
This point aside, Dan's argument does a very nice job of confronting the creationist argument that order cannot spontaneously arise out of chaos. As the Wiki example demonstrates, order can indeed spontaneously arise out of chaos, and can do so surprisingly quickly.