Wednesday, October 5, 2016

behavioral genes are preprogrammed heuristics

I've been thinking more and more about the whys of prosociality. The "identify/model a behavioral consistency" game[1] hasn't entirely played out, but its size has reached punchline proportions, so the whats seem rapidly less interesting than the whys to me personally.

"Why" leads to evolutionary and rational irrationality explanations[2]. Evolutionary explanations say there is some fitness advantage to having a particular type of social preference written into our genes. Rational irrationality[3] explanations say there is a reason to act prosocially given the contextual details and constraints of the game of life. I have a fetish for rational irrationality models, but these sort of beg the evolutionary questions because these models are mostly meant in an as-if sense. I don't consciously calculate the exact type of information to seek or the optimal heuristics to use, so what evolutionary pressures existed that led me to act like I do?

So, now I'm interested in evolution, which I know nothing about from a sociobiological standpoint and very little from the evolutionary economic standpoint, but I'm learning. And from this ignorant starting point, I'm thinking about why we evolved prosocial genes. Hopefully in a few months I will be able to identify most of the wrong or stupid steps in the train of thought I'm going to spell out next.

Prosocial norms/behavior/preferences are useful because we can overcome coordination problems with them. We're all better off if we cooperate but once everyone cooperates each individual wants to freeride. We maintain cooperation through punishment, a la folk theorems. But for this to work in a group the punishment must be cheap (which I don't think is a problem; social exclusion is pretty much free to the punishers and very harmful to the punishee) and accurate monitoring. The accurate monitoring is a much bigger problem. There might be a lot wrong or oversimplified in this paragraph because I still don't know very much about the anti folk theorem literature (another thing I'm learning about currently).

So, how to overcome the anti folk theorems? The first impulse might be to claim group selection, but there are problems with that that the evolutionary biologists/psychologists have thoroughly documented (I dare you to suggest it off-hand in a seminar with those folks in the audience!) that I think I finally understand and agree with[4]. Instead, I suspect imperfect monitoring isn't a fatal problem if the punishment is free/beneficial to the punisher and very bad for the punishee. Which is plausibly true in the setting of human evolution: I can get away with privately gorging on my hunted rabbits or gathered blueberries for awhile, but if I'm caught or if enough suspicion builds up against me, I'm ejected from the group, and then I'm really in trouble. From another group member's perspective, they want to eject me from the group because I'm free-riding, and they want to continue cooperating with the rest of the group because it's still the equilibrium that's being enforced, so by definition they want to perpetuate it. Punishment is therefore neither altruistic nor a coordination problem in itself.

But if cooperation is once again a self-enforcing Nash equilibrium, why don't individuals simply choose to cooperate when it's rational and stop behaving prosocially in contexts in which it's no longer the equilibrium? I'm not going to rehash the evidence here, but people clearly have prosocial impulses even when there is no selfish incentive to. It's worked its way into our genes. But how? A gene telling you to do what you already want to do may drift in but will just as easily drift out.

Individuals aren't perfectly smart, though. We learn what works over time and with experimentation. Someone with a mutant gene that tells them to cooperate will never learn their lesson the hard way. Someone without this mutant gene is likely to try cheating a little bit, notice they've gotten away with it, and continue cheating until they get caught and pay the price.

So prosocial genes are just preprogrammed heuristics. They help us avoid very costly mistakes and they lead to minor "mistakes" when we, for example, share with strangers in double-blind dictator games. And it makes perfect sense that social image would be such a strong motivation relative to e.g. pure and impure altruism, because a gene that prevents you from being caught as a cheat is more useful than one that tells you to be more concerned about others than yourself.

That's enough hand-waving for today. I'm sure the parts that aren't wrong are also not new, so suggestions as to where credit belong and what to read that's related are very welcome.

~~~

[1] in which I am, yes, a player. And yes I intentionally said "behavioral consistency" rather than "bias".

[2] I suppose there might be a third possible level of ingrainedness, in which we learn or are taught or habituate (perhaps intentionally) prosociality to such a deep degree that it's hard to override even in situations in which it's clearly not beneficial. This would act like a genetically programmed motivation, but would not need to have been evolutionarily advantageous in the setting in which human genes evolved (i.e. hunter gatherer society, so I am told by various evolutionary social scientists who are not actually biologists or geneticists so I don't know how much stake to put in that story). But does this kind of subconscious-level learning even exist? Any biologists out there?

[3] I wrote this phrase and immediately disliked it because it uses "rational" first in the economic, utility maximizing sense, and then in a more colloquial, bad-by-some-external-judgment-of-bad decision making sense. Then I thought someone must have used this phrase before and found Bryan Caplan's usage, which is almost but not exactly what I mean by it. Despite both of these counts against the phrasing, I'm too stupidly amused by the oxymoronity not to use it anyway. It's the simple things in life, ok?

[4] It's not actually that complicated, I just didn't make a point of learning about it until relatively recently. The stupidly obvious point that I was missing is that genes:individuals::individuals:groups is an invalid analogy because a single mutation in a gene is all it takes for a mutant individual to form, while a single mutant individual is not enough to create a mutant group.

1 comment:

Roger Sweeny said...

Morris B. Hoffman says that "not defecting" could be the difference between survival and non-survival in the early days of humanity so we evolved "preprogrammed heuristics" to blame and punish. First party punishment: conscience and feelings of guilt. Second party punishment: direct retaliation. Third party punishment: posses, juries, a court system, etc.

https://www.amazon.com/Punishers-Brain-Evolution-Cambridge-Economics/dp/1107038065/ref=sr_1_1?s=books&ie=UTF8&qid=1485883691&sr=1-1&keywords=punisher%27s+brain