What Makes A Good Explanation?

31 Jan

I posted the following over at Jerry Coyne’s blog in the comments, and I like having a record of my comments so I’m posting it here as well :). If you’re sharp and have been reading my blog regularly, you’ll notice that my “precision” example is really just Bayes Theorem and the “simplicity” example is Occam’s Razor:

The thing we have to get out of the way first, when dealing with Creationists (or anyone with wacky ideas, like 9-11 truthers), is to determine what exactly constitutes a “good explanation”.

I think that there are certain criteria that all good explanations share, yet god-belief fails at all of them.

1) Mechanism. A good explanation explains more of the underlying mechanisms than bad explanations. If your faucet is leaking and you call a plumber over to fix it, the plumber will be able to explain the underlying mechanism behind what causes the faucet to leak. There’s no mechanism for positing god, other than “goddidit” or “sin”.

2) Testability. A good explanation lends itself to being testable. Your plumber will be able to reproduce the leak at command if he actually understands the underlying mechanism. And if the leak happens again and your plumber told you the underlying mechanism, you should be able to test his explanation and fix the leak yourself. God-belief is entirely untestable (well, it is, but it fails every single test).

3) Simplicity. Good explanations use fewer ad hoc claims — i.e. claims that are not testable and have no mechanism — to support itself. A plumber that does all of the above but then posits that the reason behind the leak is that you haven’t arranged the furniture in your house in a manner that resonates with the frequencies of the Crystals of Andraste is a worse explanation than one that leaves that out.

4) Precision. Good explanations exclude more possible evidence than bad explanations. Let’s say that you have two friends who collect marbles. One friend collects only black marbles while the other collects every single color marble he can get his hands on. If your plumbing problems started after both friends were over for a few hours, and a black marble was found in your pipes, it’s much more likely that your friend who only collects black marbles caused it than your friend who collects all marble colors; even though it’s known that both friends own black marbles. God-belief does not restrict the type of evidence would be seen as opposed to naturalism so god-belief would be analogous to the friend who collects every marble color imaginable. The more evidence god-belief allows, the less likely it is that it explains this one particular piece of evidence.

I would like to see Creationists come up with their own criteria for what constitutes a “good explanation” using examples from real life which also supports their Creationism. Usually they fall prey to simplistic thinking like the tornado-in-a-junkyard strawman, claiming that evolution breaks the 2nd law of thermodynamics, the watchmaker fallacy, appeals to ignorance (“no one knows how this happens, therefore goddidit”) or the fallacy of composition (e.g. everything we know of in the universe was made by someone, therefore the universe was made by someone).

IOW, they usually make what they suppose is a negative case against evolution. They don’t make a positive case for creation.

I’ve seen all the negative arguments above, but I’m sure I’m not particularly well-versed in creationist arguments; I’d be interested to know if they do try to present positive evidence.

Which basically boils down to: Just because you suspect something is a lie, doesn’t mean you know the truth.

Posted by on January 31, 2013 in Bayes, creationism

2 responses to “What Makes A Good Explanation?”

NeuroLogica Blog

Slate Star Codex

SELF-RECOMMENDING!

Κέλσος

Matthew Ferguson Blogs

The Wandering Scientist

What a lovely world it is

NT Blog

PsyBlog

Understand your mind with the science of psychology -

Vridar

Musings on biblical studies, politics, religion, ethics, human nature, tidbits from science

Maximum Entropy

Skepticism, Properly Applied

Criticism is not uncivil

Say..

Research Digest

Disrupting Dinner Parties

Feminism is for everyone!