Wednesday, August 11, 2010

"The Limits of Reason" in Newsweek

Newsweek just published an article called "The Limits of Reason: Why evolution may favor irrationality", by Sharon Begley.  It's a good article explaining many cognitive biases and fallacies with good concrete examples.

However, I note with some amusement that Begley got this bit wrong.
Consider the syllogism “No C are B; all B are A; therefore some A are not C.” Is it true? Fewer than 10 percent of us figure out that it is, says Mercier.
But the syllogism is wrong!  The syllogism is correct if we assume that there exist some B.  So what if B is unicorns?  Then the syllogism doesn't necessarily work:

No mammals are unicorns (because there are no unicorns).
All unicorns are horses (all existing unicorns anyway, which is to say none).
Therefore some horses are not mammals.

Begley continues...
Interestingly, syllogisms are easier to evaluate in the form “No flying things are penguins; all penguins are birds; so some birds are not fliers.” That’s because we are more likely to argue about animals than A, B, and C.
...Also because we know penguins exist.

But back on topic.  Sharon's main point is that evolution favors irrationality.  This is true.  It's a well-documented fact that humans have some irrational cognitive biases.  This irrationality must have been favored by evolution; nearly all of our biological traits are there because they were, in some sense, favored by evolution.  Empirical fact.

The explanation for why evolution favored irrationality is somewhat more suspect.
Failures of logic, he and cognitive scientist Dan Sperber of the Institut Jean Nicod in Paris propose, are in fact effective ploys to win arguments.
This explanation just raises so many questions.  Does irrationality help you persuade other people, or does it help other people persuade you?  If the latter, doesn't this require some sort of group selection?  What exactly is the mechanism connecting persuasiveness to evolutionary adaptiveness?

In general, I'm pretty suspicious of these evolutionary psychology narratives, because it's difficult to prove one narrative over another.  And yes, there are other narratives.  For instance, Michael Shermer often explains confirmation bias by saying that for early humans, it was more advantageous to mistake the wind for a tiger than mistake a tiger for the wind.  Come to think of it, does Shermer have any evidence for this narrative aside from its initial plausibility?  Another alternative hypothesis is that cognitive biases are not adaptive at all; a flawed brain was just the best that evolution could manage.

Which of these hypotheses, if any, are true?  I cannot imagine any evidence that would distinguish between them.  I wish popular explanations of evolutionary psychology would focus more on the kind of evidence they use, because whatever it is, it'd have to be really clever.  Believe me, I get that scientific evidence can often be really hard to explain to lay people, but this could be downright fascinating, almost too clever to exist.


Larry Hamelin said...

This irrationality must have been favored by evolution; nearly all of our biological traits are there because they were, in some sense, favored by evolution.

This is true only if we take "in some sense" so broadly as to verge on vacuity; if it's not vacuous, then it's the adaptationist fallacy.

It is more illuminating to say that "irrationality" and cognitive bias have not been selected against by evolution.

miller said...

When I say "favored by evolution", I do mean it in a broad sense. Evolution favors not only adaptive traits, but also traits that are consistent with evolutionary history, and consistent with developmental constraints.

I do not think that this is vacuous, because we can definitely think of things that are not favored by evolution. For example, an eye without a blind spot is not favored, at least not in vertebrates. That has nothing to do with adaptation, and everything to do with evolutionary history.

Larry Hamelin said...

It's vacuous in the sense that every feature of every species is a result of evolution, in the broadest sense that evolution consists of all the mechanisms that have shaped species characteristics.

But the statement, "evolution favors irrationality" seems to say that irrationality has a intrinsic positive value, that irrationality somehow "won out" over rationality. This is precisely the sort of thinking that subtly leads people to the adaptationist fallacy, that all the traits we see are not just the result of evolution (everything is the result of evolution), but are in some sense actually optimal, that irrationality is somehow better than rationality.

But we know this is not the case in general: there are all sorts of features that haven't been optimized in even the loosest sense: features that don't face any selection pressure at all, and features where direct selection pressure for optimization is overwhelmed by indirect pressure against other features using some of the same mechanism.

It is always clearer, I think, to talk about evolution and selection in a negative sense: certain features get selected against. Eyes that can't see at all get selected against, but eyes that have a blind spot don't get selected against. It's not that having a blind spot is intrinsically better, it's not optimal in even the loosest sense, but because of other selection pressures, we're stuck with it.