Yet lab studies show that people will punish misbehaviour even if all the transactions are anonymous and "single-shot". That's generally considered to be a good thing, because society benefits (even if the individual doesn't). The mystery is why that kind of behaviour should persist.
When tested in the lab, the typical set-up goes something like this. Person A gets given some money, which he can share with person B either fairly or unfairly (keeping most for himself). Person B then gets the option to spend some of her money to 'punish' person A (by taking some or all of their money away).
Ryan McKay, at the University of London, along with colleagues at University of Zurich, set out to see how religion affects this kind of costly punishment.
They tested a group of 300 Swiss students, mostly Christian (30% Protestant, 28% Catholic, 42% no affiliation). They all took part in round 1 (allocation of money) and all took part in round 2 (option to punish the person they were playing with). But, before round 2, they were subliminally exposed to 1 of 4 different sets of priming words:
- religion: (divine, holy, pious, religious);
- punishment (revenge, punish, penalty, retribution);
- religion–punishment (divine, revenge, pious, punish); and
- control (northeast, acoustic, tractor, carton)
Overall there was no effect of the primes on the amount of punishment handed out. The religious were no more likely to punish than were the non-religious, and religious primes had no effect on either the religious or the non-religious.
However, religious primes did affect one group. Those people who had donated to a religious organization in the past year were significantly more likely to punish after they were exposed to religious primes.
That's an interesting result, because previous studies had found that religious primes affect everyone (religious and non-religious) and previous researchers have suggested that religious primes work by making people feel that they are being watched by a supernatural observer (and so they behave better).
What McKay thinks, however, is that these primes are activating the social conditioning among the 'engaged' religious. When people attend religious services, ideals of costly punishment (i.e. sacrifice for the good of the group) are drilled into them. The religious primes in this study activated that social conditioning, resulting in heavier levels of punishment.
Mckay goes on to make a wider claim: that this is evidence that 'religions' were developed as a way to increase the survivability of those people and groups who adhere to them. The essence of this argument is that religions are cultural constructions that make use of the errors inherent in our thinking (seeing things that aren't there, for example) to promote and reinforce beneficial behaviours.
Well, maybe. But in fact this study doesn't support that claim.
The problem is with the assumption that costly punishment is a good thing (for the group, if not the individual). Recent research suggests that isn't actually true. It seems that costly punishment is actually a bad strategy for individuals, and also a bad strategy for the group as a whole (the best strategy for all concerned is actually to turn the other cheek).
From this perspective, costly punishment doesn't promote co-operation (since it sets up cycles of retaliation). What it does, however, is allow hierarchies and dominance to be established.
So that's the theory. Is there any evidence that this is a problem in real life? Well possibly. You see, it turns out that anti-social punishment (i.e. retaliation against people who engage in 'costly punishment' of cheats) seems to be lowest in Westernised, secular cultures. Anti-social punishment is the evil twin of costly punishment, and is the reason costly punishment does not, in fact work too well.
Could it be that religion reinforces a behaviour which actually lowers group fitness?
McKay R, Efferson C, Whitehouse H, & Fehr E (2010). Wrath of God: religious primes and punishment. Proceedings. Biological sciences / The Royal Society PMID: 21106588
This article by Tom Rees was first published on Epiphenom. It is licensed under Creative Commons.