The first problem is that "game theory" as played in the Prisoners' Dilemma isn't science and can't be science, because a "thought experiment" isn't an experiment and can't be an experiment. You really have to believe in Word Magic to believe otherwise: it's an experiment because I call it an experiment. There are absolutely no consequences in a thought experiment if I push the fat man in front of the trolley: the game sheds no light on my likely behavior, but will enable the evolutionary psychologist to taunt me with logical inconsistencies in my stated morality, which a poor old-fashioned social scientist could have told him is just a lot of blah-blah-blah in any case. The first game theorists went wrong thinking that we are rational actors. They were staggeringly naive not to understand that costs and benefits involve moral satisfaction as well as hard cash. Today's game theorists go wrong thinking that they can now systematize all the ways we are irrational. It is worth studying how and why we act illogically and against our own interests, but the valuable work is done only when the subjects of the experiment have real-world gains and losses flowing from their conduct. Thought experiments or monopoly money games are worthless. Read Dan Ariely's "Predictably Irrational" to see how real experiments can be set up.
The rhetorical laxity infecting this whole field should be scandalous to real scientists. "The corresponding cost-benefit calculation need not be conscious, let alone rational." Just listen to yourself – a CALCULATION that is unconscious. "Surprisingly, he hardly mentions the pioneering work of Robert Trivers." Well, that might be because Trivers is purveying joke science. "Kin altruism" (I give up my life for four nephews or eight first cousins) has never happened in the history of the real world. "Reciprocal altruism" is the ultimate word-magic: a new term for "symbiosis," which is then substituted for the phenomenon under observation, namely, non-reciprocal altruism. But even "non-reciprocal altruism" is a redundancy, since altruism by original definition is non-reciprocal; and therefore "reciprocal altruism" is an oxymoron. Trivers "explains" behavior where I do you a good turn and do not expect something in return by re-labeling it (on no evidence and without even the necessity of a bogus "game theory" experiment) as behavior where I do you a good turn and DO expect something in return. This is SCIENCE? Okay, then, taking a cue from physicist Wolfgang Pauli, I'm going to re-label "evolutionary psychology" as "not EVEN wrong." It doesn't rise even to the level of pseudoscience, which sometimes has hypotheses that can be tested.
posted by Stephen Kennamer
April 29, 2010