Wednesday, September 21, 2005

Understanding Evolution

Berkely has launched their new and improved Understanding Evolution page. It's colourful, inviting, uses a tasteful number of flash animations, and still looks professional. It's very easy to navigate, but they really ought to put the beef near the top. A guided tour through evolution is what people need to see.

Unfortunately, there are a couple problems I see with this site. It spends too much time presenting the formulation of evolution. That is, it presents too much about how "we think" evolution works. I agree this is important, because creationists have done such a fine job of creating straw men out of age-old public misconceptions. Unfortunately, much of what I see will re-inforce creationist prejudices that we are just spoon-feeding them an "alternate truth" to the one they were initially fed. Which "truth" do you think they'll go for?

The presentation of homology as evidence for evolution is, I'm sorry to say, bad, bad, bad. I'm going to give some credit here to creationists in order to point out why this is bad. First, the Berkely site starts with the classic example of tetrapod limbs. While I agree it's a very ostensible example, one must remember that an evolution-weary mind is approaching this with doubts and reservations. Creationists already have a response to this and this new presentation makes no pre-emptive strikes..

Here's the problem, the skeptical student (doubful, but interested in knowing more about evolution) is faced with a dichotomy:
similar structures are caused by a common ancestor
similar structures are caused by a common designer

Why is one of these statements better than the other?

To most biologists, it's clear why the first is a better explanation. It actually offers a reason to expect the pattern, whereas the second relies on some statement about the capriciousness of the deity. To the general (and perhaps somewhat skeptical) reader, this is not so clear. In fact, both of the above statements could be viewed as begging the question (i.e. "circular reasoning").

The important issue is not simply one of similarity, but rather of similarity distributions. To which animals do these sorts of limbs belong? Why don't flies or starfish have limbs with these bones in them? Such limbs always belong to special subset of vertebrate animals. The emerging picture is one of a nested hierarchy which really is the key issue. Since nested hierarchies are a mathematically necessary consequence of any branching process (i.e. common ancestry) we have a priori reasons to expect such a pattern to emerge. This is the actual theoretical framework that is the other half of the agument. The reasons for expecting this.

In the context of a nested hierarchy, the distribution of limb skeletons is precisely and explicity predicted by a theory of common ancestry. What students are never presented with first is a picture of nested hierarchies. They're thus not presented with a picture of the evidence for common ancestry. Instead, they're first presented with a picture of a tree (the inference) and a bunch of characters mapped on (the evidence). To a naïve and dubious young student, this looks as tenuous as any crackpot seminar or tract booklet. It's easily charicatured as selective data usage and has become a favourite of creationist attacks on evolution education. Show the kids the evidence first, then show them how the conclusions came from that evidence.

If you're going to present limb homologies to students, then you have to have other evidence to re-inforce your conclusion. Otherwise, you're just going to get the "common designer" crap in return. The theory of evolution is far too elegant to be in competition with such an illogical argument.


BrianJ said...

Thank you for your clear, concise writing on complex issues. I particularly enjoy your description of the problem with teaching results rather than evidence.

Anonymous said...

Does anyone have an animation of a lancelet? Specifically the wheel organ? A prof. said there would be extra credit if we find one. Thanks Guys!