For it seemed to me that I could find much more truth in
the reasonings that each person makes concerning mat-
ters that are important to him, and whose outcome ought
to cost him dearly later on if he judged badly, than in
those reasonings engaged in by a man of letters in his
study, which touch on speculations that produce no effect
and are of no other consequence to him except perhaps
that, the more they are removed from common sense, the
more pride he will take in them.
—Rene Descartes, Discourse on Method1
SUPPOSE you grant that voters are irrational. Can you stop there? Voters are people. If they are highly irrational on election day, one would expect them to be equally irrational the rest of the year. Do individuals magically transform into a lower form of life when they enter the voting booth, then revert to their normal state upon exit?
The thesis of global human rationality is internally consistent. So is the opposite thesis that humans are irrational through and through. Is there a coherent intermediate position? Without one, the practical relevance of voters' folly shrinks or vanishes. If people are rational on Monday and irrational on Tuesday, it is a good idea to shift decisionmaking to Monday. But if people are irrational twenty-four seven, you just have to live with the fact that all decisions will be worse. By the same reasoning, if people are rational as consumers but irrational as voters, it is a good idea to rely more on markets and less on politics. But if people are irrational across the board, we should expect less of every form of human organization. The relative merits of alternative systems stay roughly the same.2
Even if an intermediate position is coherent, is it consistent with what we already know? One could postulate voter irrationality as an ad hoc exception to the laws of human behavior. But ad hoc exceptions to well-established principles understandably provoke skepticism.3 Is there any way to subsume established patterns and anomalies under a single rule?