My PhD supervisor once told me, after reading a chapter draft of mine on testing universal common descent, “You’re too hung up on falsification, Paul. That’s not how evolutionary theory works.”
Well, okay. One doesn’t want to force a theory into a philosophical straitjacket, if the theory is doing just fine, as is.
One should want SOME consequences, however, if a theory asserts that “According to our theory, A is the case except if A is not the case.” If the rooster stands on the manure pile, as the folk saying goes, it will rain tomorrow or it will not rain tomorrow. We’ve got all the possibilities covered: our theory is fully general and comprehensive.
Not a Healthy Theory
That’s not knowledge. And that is not a healthy theory.
An open access editorial (whose title I have borrowed for the headline), by the Dutch biologist Dave Speijer, is worth a look in relation to this issue.
“[E]volutionary complexity,” he writes, “invites us to tolerate exceptions.” I think this is exactly wrong. A theory that predicted the exceptions to its rules and generalizations would convey knowledge. A theory that “tolerates exceptions,” however, will end up in a 1:1 mapping with whatever one observes — in which event, the theory is doing no work at all, simply wandering along behind the data like a puppy on a leash.
Rationalizing Failed Predictions
Stephen Jay Gould grew very fond of the notion of “contingency,” which he would deploy to rationalize departures from prediction. If that seems a harsh judgment, consider the last sentence of this article, commenting on Gould’s position:
Organisms tend to achieve similar solutions to similar problems, but give it [i.e., evolution] enough time (or a small enough population), and anything is possible.
“Anything is possible” may be true — but then, don’t pretend you have a theory which is doing any real work. You don’t.
Editor’s note: This post was updated on May 14, 2020.