Contra Deus ex Machina

In Ars Poetica (“The Art of Poetry”), the great Roman lyricist Horace counsels against using gods to resolve thorny plots. The deus ex machina is simply too tidy and unbelievable. When gods swoop in to save the day, the mundane becomes sacred. Metaphysics to the rescue.

I was reminded of Horace’s enduring wisdom by two recent studies; the first on cooperation and second on punishment. Both are major contributions to our understanding of human altruism and collective action. Neither invokes the magic of gods.

In the “Evolution of Direct Reciprocity,” Andrew Delton and colleagues demonstrate that humans are naturally generous even to strangers and that such generosity is evolutionarily advantageous. A co-author of the July 25 PNAS study, Leda Cosmides, explains why humans can afford to be generous (i.e., incur costs) even when interaction might be a one-time affair:

There are two errors a cooperating animal can make, and one is more costly than the other. Believing that you will never meet this individual again, you might choose to benefit yourself at his expense –– only to find out later that the relationship could have been open-ended. If you make this error, you lose out on all the benefits you might have had from a long-term, perhaps life-long, cooperative relationship. This is an extraordinarily costly error to make.

The other error is to mistakenly assume that you will have additional interactions with the other individual and therefore cooperate with him, only to find out later that it wasn’t necessary. Although you were “unnecessarily” nice in that one interaction, the cost of this error is relatively small. Without knowing why, the mind is skewed to be generous to make sure we find and cement all those valuable, long-term relationships.

This is the restrained and mathematical kind of evolutionary psychology we can believe in.

In “Punishment Sustains Large-Scale Cooperation in Prestate Warfare,” Sarah Mathew and Robert Boyd find that profane punishment solves the free-rider problem that so exorcizes some evolutionary theorists of religion:

Understanding cooperation and punishment in small-scale societies is crucial for explaining the origins of human cooperation. We studied warfare among the Turkana, a politically uncentralized, egalitarian, nomadic pastoral society in East Africa.

Based on a representative sample of 88 recent raids, we show that the Turkana sustain costly cooperation in combat at a remarkably large scale, at least in part, through punishment of free-riders. Raiding parties comprised several hundred warriors and participants are not kin or day-to-day interactants. Warriors incur substantial risk of death and produce collective benefits. Cowardice and desertions occur, and are punished by community-imposed sanctions, including collective corporal punishment and fines. Furthermore, Turkana norms governing warfare benefit the ethnolinguistic group, a population of a half-million people, at the expense of smaller social groupings.

These results challenge current views that punishment is unimportant in small-scale societies and that human cooperation evolved in small groups of kin and familiar individuals. Instead, these results suggest that cooperation at the larger scale of ethnolinguistic units enforced by third-party sanctions could have a deep evolutionary history in the human species.

Large-scale cooperation, in other words, can revolve around something other than systematic religion or supernatural punishment. Shared language and ethnicity — along with earthly rewards (and beatings) — seem to work just fine.

We don’t need a group level or adaptive deus ex machina to explain the extraordinary success of ordinary humans. Parsimony to the rescue.


Delton AW, Krasnow MM, Cosmides L, & Tooby J (2011). Evolution of direct reciprocity under uncertainty can explain human generosity in one-shot encounters. Proceedings of the National Academy of Sciences of the United States of America PMID: 21788489

Mathew S, & Boyd R (2011). Punishment sustains large-scale cooperation in prestate warfare. Proceedings of the National Academy of Sciences of the United States of America, 108 (28), 11375-80 PMID: 21670285

Did you like this? Share it:

16 thoughts on “Contra Deus ex Machina

  1. Psycasm

    I’m disappointed in the first part of the post, particularly the reference from twitter that led me here.

    “This is the restrained and mathematical kind of evolutionary psychology we can believe in.”

    They demonstrated, in a model, that generosity is favoured under certain circumstances.

    Those circumstances did not include a notion of reputation (very frequently linked to theories on generosity), nor did it include any sense of input for estimating/assessing the likelihood that a given single exchange would be one-shot, or repeated. The program developed a value, based on past experience with other agents within the program, of the general likelihood that ‘this interaction will similar to past interactions’.

    This is not great Evo Psych; being ‘restrained and mathematical’ may be virtuous – and certaining this is thought provoking and may lead future enquirey- but I see no reference that these data, these artificial data, are mirrored in the real world. The model predicts a certain pattern of behaviour, yet fails to demonstrate (or even allude) that these data are evident amongst population with trait generosity, nor a proposal of how one might explore this within a population – be they humans, apes, vampire bats, or … anything.

    It would be difficult to deny that a computer model, as well programmed and informed as it may be, is any less depraved of nuance and complexity as some other prominent Evo Psych theories… So just as a point of discussion I’ll ask the following question.
    I assume that ‘unrestrained’ Evo Psych is a reference to ‘just so’ theorizing; and yet’just so’ stories are not all equal. Some may be pure navel-gazing, and some are based in contemporary understandings of social and comparative data. Some have more value than others. And so my questino is – Where on the spectrum of ‘just so’ stories do impoverished econimic models fit when attempting to explaining human behaviour?

  2. admin Post author

    You don’t like Horace? Or you don’t appreciate his advice on the use and abuse of the deus ex machina? I don’t twitter so don’t know what twitter reference led you here.

    It seems to me that their model has “reputation” built into it; reputation is encapsulated by the parameters and subsumed by them. And although it is just a “model” with all the simplifications and assumptions that accompany all models, this one happens to resonate with what happens in the real world. I have often acted the way the model predicts, knowing full well the encounter would be one-shot. Haven’t you acted this way also?

    I can’t really answer your question because I don’t sit around evaluating evo psych models and fitting them to some kind of spectrum. Some are good, some are bad, and many are in between. I evaluate them on a case by case basis. This one looks good, and fits squarely within a longstanding tradition of using game theory (and population genetics math) to evaluate human behavior.

  3. J. A. LeFevre

    Nice start, but recognizing flour in the mix does not solve the secret of baking bread.

  4. Psycasm

    Out of Context science (@oocscience) was promoting the quote on twitter.

    I disagree – reputation was not in the model. Each interaction was determined as a likelihood based on all previous experience with other agents… not with a given agent in particular. It seems that there was no Agent A, Agent B, Agent C by which the model could discern whether the interaction was one-short, or not. I don’t think the reputation of the group (which is kind of what they’ve got here) is meaningful if you’re making a claim about individuals.

    I’d like to make my position clear, however. Models are just fine, they can be incredibly useful. But they inevitably lack full sophistication (as does does so much research, there’s always a lot of variance in the results). Apples and oranges, in my opinion.

    I’m not sure you can claim ‘this one resonates with what happens in the real world’. The paper didn’t provide any evidence to support that idea, and your anecdote is not useful. I could just as easily say ‘no, this does not match my experience’ and we’ve got a stalemate. N=1 is not useful here, only comparable measures in larger samples are.

    Furthermore I agree that some evo psych models are good and bad (and everything in between). However, if you’re capable of making a judgement on a case by case basis surely you’re capable of making a judgement with regard to my question?

    I think LeFevre makes a good point. They’ve identified something of interest (and it is interesting). It doesn’t explain much at the moment… but it might explain something with future research – particularly if they apply it to a population that does engage in generosity-related behaviours.

  5. admin Post author

    It is ironic that “Out of Context” science would pull this particular statement from the blog without knowing the context. This context would be the long series of posts I have been doing over the past year in which I take serious issue with the more egregious forms of evo psych storytelling. Several of these posts are recent. I have also been at pains to explain that just because (or “just so”) there is a bunch of really bad evo psych out there, this does not mean that evo psych does not hold promise or that it is a doomed enterprise.

    As far as resonating in the real world, my N is quite larger. I can’t think of a single friend of mine who hasn’t, at some point in her/his lifetime, acted generously or altruistically to a complete stranger in a situation where my friend knew, or could reasonably suppose, that it would be a one-shot encounter. I see this daily. I am not sure what kind of friends you have, or where you have been in the world, but this is my experience. And guess what? The model partly explains this kind of behavior. Unlike many who champion the idea that “religion” was the magic elixir which transformed humanity and paved the way to earthly heaven, I am not a mono-causalist.

    As you and the good baker LeFevre note, this study is but a single strand of evidence among many. Whatever its limitations, it adds another brick to the wall.

    My comment about this being restrained evo psych — the “kind we can believe in” — was a tongue in cheek way of referring to all my other posts on evo psych (my regular readers would have known this), and a subtle dig at evo psych types who are “true believers,” the faithful if you will. Your sarcasm meter must be running low.

  6. J. A. LeFevre

    Consider the level of cooperation uncovered in this review: Young, testosterone charged men of the Turkana collecting together as an inner-city street gang, building ‘cooperation’ through threats and beatings, then setting out to raid their neighbors for their cattle. The raids last several hours to a couple of days, then everyone goes home to their own families. They may make one to a few of these raids a year. The total ‘cooperation’ identified in this study lasts a few days a year, hardly the stuff stable, successful communities are built of. Not mentioned in this study is that the Turkana share a common religion. The role of this was not considered nor isolated from other variables. You are claiming conclusions that were absolutely not even considered in the study, and not supported by the evidence presented.

  7. admin Post author

    Given that one of my advisors (Terry McCabe) lived and worked with the Turkana for nearly 20 years, and I have studied them extensively, I can tell you that their “religion” (if one can call it that) is less than coherent, unsystematic, fluid, changing, and is not cited by any of them as the reason for being “Turkana,” or as the reason for their cooperation. You can begin your study of the Turkana by reading Terry McCabe’s award winning book, Cattle Bring Us to Our Enemies.

    I was talking to Terry recently about Turkana “religion” and this is what he said: “You can ask them about religion, and most of them will tell you they have no idea and those who offer opinions candidly claim that everything is vague. They don’t care.”

    Despite this fact, the Turkana are a distinct ethnolinguistic group of nearly 500,000 that engage in group level cooperation that extends far beyond raiding; group grazing rights and communal obligations to assist other Turkana with food and shelter in times of crisis being foremost among them. The Turkana are a discrete functioning group, engaged in large scale cooperation, that don’t rely on communal rituals or religion for bonding or prosociality. They have been so for well over 100 years.

    If you read any of the ethnographic books or articles on the Turkana (which I recommend you do before engaging in thought experiments about “Turkana gangsters”), you will realize they are bonded and cooperative because they share a language, ethnic identity, history, territory, and because of extended/fictive kinship.

  8. admin Post author

    I would also add that if you had read the study instead of just the abstract, you would know the authors in fact noted that the Turkana “share a common religion” and claim a “common high god.” The important point is that this “common religion” is known to very few (i.e., the diviners or shamans) and most don’t care. It just isn’t important.

    As for the study not supporting the inference I draw from it, pray tell what you mean. A blanket assertion simply won’t do.

  9. J. A. Le Fevre

    Sorry, I did mean abstract, as the study requires a subscription, and I was simply continuing my original comment. I did not suggest ‘inference’ but ‘claim’, and I recognized only one claim you made, which I was addressing: that the cooperation of the people was independent of religion.
    Whether true or not, it was not even hinted at in the abstract, nor was there any suggestion in what you’ve posted (originally or your following comments) that religion had been isolated as a variable. It had simply been ignored as a factor, as you continue to do. Its effect is probably small, though I suspect significant, but simply dismissing a variable does not demonstrate that that variable is not a factor.

    The Turkana religion (yes, I do call it that) is similar in ‘fluid form’ to most aboriginal religions, but having an identified ‘god’ makes it more ‘formal’ than typical hunter-gatherer ‘spirit guides’ centered religions. Pre-state religions do not resemble state, or ‘organized’ religions, and pre-state societies place much less interest or attention to religion than do state societies. They also have much lower levels of cooperation.

    I think the important aspect of basic religion (if not all religion) is not what the lay individuals think it is about but that they participate in the rituals. What does your advisor have to say about their rituals?

    Lets jump back to snippets from your post:
    From Mathew & Boyd (rearranged for clarity, I hope not to change the meaning – original text in original post above): .’ . .the Turkana sustain cooperation in combat at a scale of several hundred, at least in part, through corporal punishment.’
    And: ‘Turkana, a politically uncentralized, egalitarian, nomadic pastoral society . . ‘

    So to paraphrase Chris, cooperation of up to hundreds is seen to be enabled by beatings.
    That sounds like the social structure of the chimpanzee, except that the chimp community is stable for years to generations, where this human ’community’ (of warriors) is stable for the duration of the raid, and then it’s back to their families.

    Mathew & Boyd go on to say: ‘. . . these results suggest that cooperation at the larger scale of ethnolinguistic units enforced by third-party sanctions could have a deep evolutionary history in the human species.’

    Yes, right back to the apes. Evolution is like that.

  10. admin Post author

    After you read the actual study and bone up on your Turkana ethnography, let’s have a discussion. This one is going nowhere.

  11. J. A. LeFevre

    I have been to the Rift valley, met the Turkana, seen their camps and their cattle. They are a tough people who live on land few want and fewer still could survive on. I have also stayed in villages at the edge of the valley where my sister taught school. Where they are teaching their children to cooperate in communities of thousands to billions, to hold jobs over spears (and increasingly small arms). Villages where they live in mortal fear of tribal raids. Sorry if my emotional attachment to the people of Kenya is distracting, but blind reverence to romantic notions of ‘traditional’ culture break down when your friends are its victims.

  12. admin Post author

    I think your attachment to the people of Kenya is endearing and valuable. I was only saying that your conception of the Turkana as a whole and their high-level interaction is mistaken. It does not sound like anything you saw there would have given you insights into their large-scale cooperative enterprises. I think you would really like Terry McCabe’s book on them.

  13. J. A. LeFevre

    I might enjoy his book, I know very little more of them than already explained. My original point is that ‘large scale’ must be taken in careful context. The Turkana are a band-level society, the most (politically) primitive of modern humans. In a tribal level society (the next step up), 200 cooperating adults would be expected, not noteworthy. My (earlier thread) comment holds: Cooperation and community among humans has proven to be a very difficult achievement.

  14. admin Post author

    The old typology of band-tribe-chiefdom-state-empire was discarded by anthropologists several decades ago. It just doesn’t fit the the facts, and there is far too much variation. There were so many exceptions they swallowed any supposed “rule.” Having said that, and even assuming this typology applies, the Turkana are not a band level society. The first thing you must understand about “bands” is that they are networked with other groups. Bands are foraging units and they are fluid. Bands don’t exist by themselves, or if they do, they don’t last long. I don’t know how you can possibly claim that an ethno-linguistic group of 500,000 people who self identify as “Turkana” as a “band” level society. Even the guys who used to use the terminology (Elman Service, Robert Carneiro, etc) wouldn’t say that because it is patently ridiculous. Like I have said several times, after you have read several ethnographic books and articles on the Turkana, we can have discussion.

  15. admin Post author

    Why do you think cooperation is such a big deal? You can find astonishing amounts of cooperation in the animal kingdom, among mammals, and especially among primates. You seem to be mystified by the whole thing.

Leave a Reply