Field of Science

Religion as a costly signal: why the idea is bunk

In the previous post I mentioned the idea of costly signalling. And that's prompted this post, which has been gestating for a while, about the 'costly signalling' explanation for religion. I think the idea is fundamentally flawed, and to explain why I'm going to lean on an essay by Jeff Schloss, who's an evolutionary biologist and ex-member of the Discovery Institute! (You can read more about that bizarre story here).

A costly signal is a cunning evolutionary device, and the classic example is the male peacock's tail. The elaborate tail imposes a cost, but (so the theory goes), it also demonstrates to potential mates the male's genetic fitness. So the guys with the big tails get the girls, and the investment in the tail pays off.

The crucial feature of a costly signal is that it's hard to fake. Keep that in mind...

There is a theory that religious rituals evolved because they're a costly signal. To understand how this works, first you have to accept that religious beliefs encourage people to be honest. (This isn't really backed up by the evidence - the evidence is that environmental primes are effective but not supernatural beliefs in themselves. But anyway...)

So, the theory goes that being altruistic is a potentially a good thing, because people will treat you better. But they can only do that if they can trust you. And that's where costly signalling comes in.

The idea is that all the rituals involved with religion are actually a kind of costly signal. Only people who truly have supernatural beliefs will devote the time and energy to religious rituals, and so you can tell the true believers by their outward show of devotion.

Anybody spot the flaw in that one?

OK, so the obvious problem is that it's a signal that's easy to fake. If going to religious services and pretending to be pious gets you and advantage, then that's what cheats will do.

So, says, Jeff Schloss, we can move up a level. He suggests that deep-seated, involuntary actions are the true costly signals:

Another way—sometimes attending ritual but often contrasted with it—is the widespread, varied, and in many respects distinctive existence of highly visible, involuntary, dramatic manifestations of religious experience: Glossalalia (“speaking in tongues”), convulsive weeping (“veil of tears”), contagious laughing or singing (“holy laughter” or “singing in the spirit”), fainting (“slain in the spirit”), trembling/shaking (“under the power”), religious trances, spontaneous bleeding, etc.

The existence of these ecstatic human behaviors, especially in the religious context, warrants both proximal (neurophysiological) and ultimate (evolutionary) explanation. Unlike involuntary displays such as blushing or piloerection, which merely signal emotional arousal, or vasomotor fainting/epileptic seizures, which are not associated with particular cognitions—these autonomic manifestations are taken to reflect the experience of a very specific (and sublime) reality.

To put it bluntly, when religious people freak out they are giving a hard-to-fake religious signal. The analogy is with smiling - a smile is hard to fake.

Frankly, I'm skeptical. Firstly, these kinds of behaviours are relatively rare, and there's no evidence that people who act like this are regarded as more trustworthy. There's certainly no evidence that they are more trustworthy.

Secondly, smiling might be hard to fake but people who have an incentive can certainly do it. There's no shortage of con men out there who can do it. If you trust people because they 'have an honest smile' then you are a ready-made dupe.

Lastly, and more fundamentally, I think the whole concept is fundamentally flawed because there is no evidence that religious beliefs are linked to altruistic behaviour. Religious priming is, certainly. If you put religious images up, or prompt people with religious messages, then they behave better. But this works with atheists just as well as with the religious. Beliefs have nothing to do with it (Shariff & Norenzayan showed this back in 2007).

What's more, what's more, people tend to justify and rationalize their bad behaviour. Since they also tend to create God in their own image, they can easily co-opt their God into their own rationalizations.

If they can do that, then the whole idea of costly signalling is fatally skewered. If religious beliefs are not linked to altruistic behaviour, then engaging in religious rituals can't possibly be a signal of good intent.

Schloss himself makes this point, and I think I'll leave the last word to him:

It is possible that these highly contagious religious displays are not adaptations for human flourishing at all, but are viral memes parasitizing reward systems that have been selected for other purposes or distorted by various deprivations. Although I have been arguing that this is not the case and that religious affections along with their distinctive manifestations play an important role in promoting cooperative commitment, still, they are notoriously vulnerable to a final, quinternary level of cheating: self-deception. Unlike intentional hypocrisy or consciously manipulative employment of costly signals (Cronk 1994), the best way to fake a hard-to-fake signal is to be sincerely, though inauthentically persuaded of ones own commitment.

Creative Commons LicenseThis work by Tom Rees is licensed under a Creative Commons Attribution-Share Alike 2.0 UK: England & Wales License.


  1. Costly signal is costly indeed.It goes against the evolutionary theory in one sense.
    It increases the short term (that generation) reproductive success and thereby succeed in transmitting genes.But then a peacock's beautiful but relatively useless tail decreases life expectancy.
    The chances of repeated reproduction is not possible in that case,so in a way decreases overall reproductive success.
    Also remember that the genes which code for 'beautiful' tails keep getting transmitted.So in the future generations the distinctiveness that beauty offers would not remain but the disadvantage does.This could jeopardize the whole stock of peacocks and peahens.Isn't it?

    Anything could bring such signalling into being except the non-aesthetic blindly functional force of evolution.

  2. Tom, the costly-signalling theory has been developped into different directions (i.e. by Sosis and Zahavi respectively) and it has been tested empirically. The peacock-model has been weakened, but models incorporating social communication (that is: benefits for signaler and receiver) have been strenghtened. See for example Sosis, Richard and Eric Bressler. Cooperation and commune longevity: A test of the costly signaling theory of religion. Cross-Cultural Research 37:211-239

    Free download here:

    Best wishes! Michael

  3. Michael, yes, I've read Sosis' study on longevity of religious communities. And I totally agree that costly signalling is an effective way to increase group commitment.

    This has been shown for gangs in general, of course. It might be reinforced by religion, as Sosis' work suggests (although it's very preliminary).

    That would help explain why religious ritual evolved. But not religious beliefs.

    What I'm talking about here is not group commitment, but metaphysical beliefs. I simply don't see how you can have a costly, honest, hard-to-fake signal of supernatural beliefs.

    And I think the number of fakers bears me out!

  4. In a sales seminar that I attended the professional sales instructor (he had been in sales for 40 years) stated that in order to convince other people that your product is the best deal, the very first thing you have to do is convince yourself that your product is the best. Really good professional salesmen (e.g. politicians, religious leaders, military leaders, etc.) are masters at self-deception.

  5. Tom, the point seems to be not about "metaphysical beliefs", but about the acceptance of a supernatural observer. (You certainly know the studies of Bering et al. on that subject. Or saw a totem pole. :-) ) For example, we have a range of studies i.e. about Jewish commandments concerning food, clothing and the Shabbat. All of them are costly, signalling belief in God and obedience can be observed by fellows, neighbours etc. Throughout the centuries, these religious commandments secured the identity of the minority, whereas liberals and seculars tended to assimilate in the broader public. Although there have been fakers, of course, the average level of cooperation among the religious remained higher (i.e. in comparison of the (thriving!) religious kibbutzim in Israel compared to the (imploding) secular ones). And, of course, the orthodox surpass the seculars in the average number of offspring by far.

    In "Judaism in Biological Perspective" (Ed. Rick Goldberg, 2009) are a range of strong studies and a wonderful article from Zahavi (the ornithologist who originally brought in the peacock-modell, but has delved deeper since then) on that topic.

    Best wishes!

  6. Michael, triggering thoughts of a supernatural observer can prompt more honest behaviour. And religious communes have greater longevity than non-religious ones. However, this isn't evidence that religious rituals are a costly signal of belief in a supernatural observer.

    There are several reasons for this. It isn't belief in a supernatural observer that counts - it's environmental prompts. In the absence of a prompt, religious people don't behave better. In the presence of a prompt, both religious and atheists behave better (e.g. this study).

    Second, it's easy enough to explain the data in terms of conventional ideas about costly signalling and group commitment.

    Lastly, signalling about internal beliefs regarding a supernatural observer is easy to fake. So it doesn't work even in theory.

    That's the key consideration. Costly signalling of group commitment is impossible to fake. Costly signalling of supernatural beliefs is easy to fake.

    Why should religious communes show greater longevity? I believe because the benefits of group membership (life everlasting, as far as the believers are concerned) are greater than the benefits that could be offered by any secular group.

    I'm going to do a post on this, I think!

  7. Tom, I am looking forward to your post! Thank you very much for the interesting perspectives and discussions!

    In testing Sosis in distinctively different settings, Soler came to conclusive results. See "Commitment Costs and Cooperation: Evidence from Candomblé, an Afro-Brazilian Religion" by Montserrat Soler, in: The Evolution of Religion, 2008.

    Personally, I find the studies of Jesse Bering most convincing, which highlight the strong influence of perceived observance. See

    Note that supernatural agents are frequently narrated as watching, loving and punishing, their depictions (from skulls to totem poles to hieroglyphs and statues) centering on the subject of observing the faithfull. By ritually submitting to this supernatural watching, religious affiliated tend to evolve and signal rule-abiding lifestyles (without becoming 'better' people in general). I am looking forward to reading more of your thoughts!

  8. Dear Tom,

    The truth is that practicing religion is costly.One pays for his convictions if he is in a minority because of his faith.

    The payoffs are for the ones who identify with stronger group be it the religious(like Muslims in Arabia)or non-religious( like the communists in communist countries).

    Considering this,let us see who would actually stand for his true atheist or a theist.

    For an atheist the life on earth is all that there is.There can be no reason why he should not deny his atheism if he were at gun point.A theist on the other hand looks forward for heaven etc so there is motivation to speak the truth at such a gunpoint.

    I realize that there are weak theists and strong atheists.It is who values truth.At least I can say a Christian should value truth.As for him truth is absolute.A Hindu can take a pragmatic view and say anything, with his belief that truth is only relative.

    Of course the above considerations also depend on if other members of the social group are spectators to the responses.This could affect responses either way.If the ones seeing are sympathetic to beliefs or have hatred towards such beliefs,motivations to stick to or defect from original viewpoint may be triggered,respectively.

    There is another perspective I like to offer regarding changing behavior.Priming of mind with 'good' can motivate each of us to do that good deed.This is a behavioral approach.It is quite superficial.Behavior disappears with removal of priming stimuli.

    Change that religion offers is much deeper.It takes time to 'change' a person's schemas and personality.Though it is tough,such change is more enduring.

    No one argues that religious ones are better than others.They are ones who know that they are not good enough(at least the Christians).They hope to be better by God's grace.They know it takes time.They call it sanctification.

  9. Hi Dheeraj, you're quite right that an atheist would probably rather say 'There is a god' than be shot dead. We could argue about whether or not that's a sensible thing to do. Personally, I think that if you're confronted by a homicidal maniac, the best thing to do is humour them :)

    At the end of the day, though, discussions like that are totally hypothetical. A better question is to investigate what the practical effects of belief in life after death might be.

  10. As for those demonstrations of highly religious feelings such as glossalia, crying, etc. they might be the result of sociological phenomena. Mass hysteria of some sort. I've never personally seen any of these displays, but from what I've read, they only occur in groups. Valerie Tarico, a psychologist and former Evangelical Christian, describes this event as "breaking"; she has a short post over at another blog where she gives a short description of it.

  11. Tarico's 6-part series "Christian belief through the lens of cognitive science" (also available at Huffington Post) is excellent.

  12. I'd like to add that while ritual participation is easy to fake, commitment to religious studies and work is very hard to fake. A professional priesthood serves as a good way to ween out people not committed to the group's ideology and, if it includes such, charity and honesty values, especially in religions or sects without much political power. This in turn allows them to function at a higher level of effectiveness and lower level of corruption than the local government, which gives rise to organizations such as the Hamas or Hezbullah. As they gain power, however, it becomes more and more advantageous to fake the religious commitment necessary to become a member of the priesthood, and in a few decades the system becomes corrupt.

  13. I think that commitment to religious studies is also easy to fake. Anybody can do it, to be honest. It's simply a question of whether the benefit is worth the faking. Most people can't be bothered, because the benefits aren't that great (unless, of course, you happen to be a believer...).

    A true 'costly signal' should be impossible to fake. Even if you'd benefit from faking, you shouldn't be able to. Otherwise it's not a reliable signal for the attribute.

    I'd also be a little bit cautious about supposing that religious diligence is linked to better behaviour. The Catholic Church is finding that out to it's cost. Who knows what horrors go on in the hidden worlds of Hamas and Hezbollah!

    In this light, a study showing that people who score high on 'moral certainty' are also more likely to cheat.

  14. Tom - how difficult religious studies and work are to fake depends on the surrounding social constructs (cheating your way through a yeshiva is rather difficult), but the point is not that it is impossible to cheat but rather than it's costly. When the religious group is small and not influential this creates self-selection - nobody bothers to cheat, because there is no benefit. This leads to people in the cult being (largely) those actually attracted to its teachings, and if these teachings include honesty and justice then they'll be more likely to uphold them. Again, as the cult gains power this self-selection effect dissipates and temptations rise, leading to rampant corruption - but the movement has already risen, it has momentum that will not easily be swayed.

  15. Just coming to this but isn't one possible defense that as long as the costs create an incentive to the non-believer to break religious laws in secret, then the person's religious community will pick up on this and oust them?

  16. So it depends on whether you can lie effectively. If it's possible to lie and get away with it (not necessarily all the time, but often enough to break the system), then the signal can't work.

    And it's clear that it is possible to lie about your religious beliefs. Remember, the lie doesn't even have to be a conscious ones. The most potent lies are ones where the liar has convinced himself that he is faithful.

    There are, of course, a lot of examples of self-deceptive religious people who do not live up to the the standards set for them. Enough to break the system? I think so, certainly.

  17. If the liar has convinced himself - that is, the source of the original assertion, belief, idea, etc. - then it ceases to be a means of deceit or defect. It has effectively, then, become an "honest" and credible signal.

    It appears therefore that the lines are blurred ...


Markup Key:
- <b>bold</b> = bold
- <i>italic</i> = italic
- <a href="">FoS</a> = FoS