Continuing the thoughts in the earlier post, Bogusity, (or should that be bogosity?) about research findings and their interpretations. Well designed experiments and observations yield data that has intrinsic value, but the interpretation of that data seems to be a sand box where superstitions and biases often play.

Consider this:

Experimental evidence for reciprocity comes from behavioral game theory, which uses economic games in which subjects make choices under varied social conditions. For instance, Herrmann et al. employ a public goods game in which each of four anonymous subjects is initially given 20 tokens, and each is told he can place any number of these tokens in a public account. The tokens in the account are multiplied by 1.6 and the result divided evenly among the four. At the end of the experiment, the tokens are exchanged for real money.

In this game, each individual helps the group most by placing his 20 tokens in the public account, and if all do so, each earns 32 tokens. However, if a single individual is selfish, he will place nothing in the public account, and his earnings will be 20 + 60(1.6)/4 = 44 tokens. But, if all four are selfish, each earns only 20 tokens. Because the four subjects are strangers, the standard view of human nature suggests that there will be zero contributions. However, in the many times this game has been played in a variety of social settings, the older view is virtually never supported, and the average contribution is about half the initial endowment (8). . .

However, mean contributions to the public account generally fall over many trials, reaching a very low level after 10 repetitions. By varying the rules of the game, researchers have concluded that the principle of reciprocity is responsible for the observed decay of cooperation: Subjects who contributed more than average on one round contribute less on the following round, showing their disapproval of the unfairness of their fellow players. Indeed, a single selfish individual in the group can lead contributions to spiral down to almost zero.

I have a number of doubts about the experimental design since it is highly unrealistic, but they can wait for a bit.

An innovation of Fehr and Gächter (9), used by Herrmann et al. as well, was to add punishment after each round of play. Each player A could specify that the player B associated with a particular contribution have three tokens deducted from his payoff, for each token deducted from A’s payoff. Under these new conditions, the high contributors punished the low contributors who, in succeeding rounds, increased their contributions, so that in the 10th and final round, there was almost 100% cooperation. The behavioral propensity to cooperate with others at personal cost, and to punish non-cooperators even when this is personally costly in the long run, has been called strong reciprocity. The punishment meted out is considered altruistic because it increases the payoff of group members at a personal cost to the punisher.

Altruistic punishment has been discussed here in the past^1,2, citing research evidence that "cooperation can become the default behaviour in large groups provided punishers are willing to punish not only those who cheat, but also those who fail to punish cheats". But is this a cultural artifact, something that isn't a human universal?

A few subjects, when punished, rather than contributing more, suspected that it was the high contributors who punished them, and responded with antisocial punishment: They punished the high contributors in future rounds, leading the latter to reduce both their contribution and altruistic punishment (11).

Herrmann et al. collected data in 15 countries with widely varying levels of economic development. The subjects were university students in all societies. The authors found that antisocial punishment was rare in the most democratic societies and very common otherwise. . .

The countries with a high level of antisocial punishment and a low score on the WDA [World Democracy Audit] evaluation included Oman, Saudi Arabia, Greece, Russia, Turkey, and Belarus. The most likely explanation is that in more traditional societies, the experimental setup represents a clash of cultures. . . When punished, such subjects are likely to respond with anger rather than guilt. Punishing the high contributors is thus a means of asserting one’s personal values, which take precedence over maximizing one’s payoff in the game.

I wonder how it would go if the earlier notion of punishing non-punishers had been an option, if that information had been made available to players?

A lot of confusion is introduced into the game description by the use of biased language. Why is it antisocial for players to return fire at those who have taken shots? How is any of this about public goods? The belated realization that there was a culture clash going on may be an indication of far deeper confusions resulting from biased expectations. Simply using students as test subjects introduces confusion.

If IQ really correlates with the ability to flourish in an industrial society (and I'm quite prepared to believe that), then it is, as I said last time, a measurement of the ability to navigate paper-pushing bureaucracies — to learn to manipulate arbitrary abstract explicit rules, and to do so on command.

Students, despite their propensities for rebellion, are among the most docile members of society. The longer they remain students and the more advanced their attainments the more this is true. There are individual exceptions, but as a group they grow increasingly docile.

It seems to me that the concepts of reciprocity and socialization need to be clarified in light of the demonstrated effects of varied socializations. Those who are damned as being antisocial by one culture could be considered to be prosocial by another. Reciprocity, then, is a far more complicated behavior than assumed and asserted by the researchers. Experimental design would change to reveal any underlying behavior that might be more plausibly called innate.

A loosey-goosey expression of that insight is that the experiments as designed reward a sort of priggishness and conformism that is not universal in human societies and that contra the assertions in the press release about this research does not "show that the advanced market societies with democratic institutions produce an ethic of spontaneous cooperation, with a strong altruistic dimension, that likely accounts at least in part for their material success and legitimacy".

Much credit for the material success of such societies goes to those who are less docile and escape the priggish pressures of the herd, and so are more creative and innovative. These are often people who bristle when oppressed by prigs rather than knuckling under in guilt. Such people are perfectly capable of truly cooperative behaviors, especially adhoc, task oriented associations. They fit well into the sort of heuristically diverse groups that Scott Page and Lu Hong have shown to be so very adept at solving the most difficult sorts of problems.

Another consideration is the very arbitrariness of the experimental design. What real world situation does the magical increase of tokens by 60% map to? An alert person would be suspicious of such a claim and discount it appropriately. This adds risk to any move in the game and casts a different meaning on the behavior of punishing prigs. Other research into economic behaviors has shown that this sort of distrust of the claims of authorities explains a great deal of what simpler economic and political models have called irrational behavior. It's perfectly rational when you consider that the authorities are often mistaken and likely lying in any event.

[NB - this is a draft post that I am not satisfied with, but a friend wanted to see it and perhaps will help fix it]

Update:

Speaking of biased language and expectations, it's Fantasy Land.

Throughout the ages people have enjoyed expressing, maybe even believing, moral discourses bearing little relationship to their everyday behavior. And, as people tend to be more rational in their actions than in their talk or in their thoughts, we are fortunate that morals have largely remained in the realm of fantasy.

It seems to me that human living standards have improved faster than our moral discourse. In this sense, the contrast between our well-being and our fantastic ideology has never been greater than at present. The "consensus" moral of our time is a potpourri of socialist, environmentalist and puritanical ideas that stand in stark contrast to almost everything that makes our lives enjoyable.

I've complained that politics looks backwards, and so always fails to address current realities, still chewing on old historical bones rather than the meat of present or future issues. This seems most clear when politics chases moral discourses, which tend to be even more backward.

I wonder if it could be otherwise? Can morality and politics be reformed in some way to engage with reality?

Update:

Richard Posner on student docility

From the standpoint of most teachers, right up to and including the level of teachers of college undergraduates, the ideal student is well behaved, unaggressive, docile, patient, meticulous, and empathetic in the sense of intuiting the response to the teacher that is most likely to please the teacher. Those are traits less characteristic of boys than of girls.

College students make exceedingly poor test subjects when the objective is to understand society. By the time they reach college they have been selected for unusual docility.