Commentary on Dunbar and Baumard

By Herbert Gintis May 12, 2012 6 Comments

In his engaging Social Evolution Forum contribution, Networking Past and Present, R.I.M. Dunbar argues that in traditional societies, most people share the same network of friends and relatives because they belong to the same community. In contemporary society, by contrast, our social networks have become fragmented, and we live predominantly with casual acquaintances and strangers. In contemporary life, then, people are no longer willing to intervene to correct violations of social morality, and hence are not prepared to protect the wider interests of the community.  In his comment on Dunbar’s analysis, Nicolas Baumard argues that the sort of third-party punishment that Dunbar considers the stabilizing force of traditional societies is very unimportant, on grounds that among hunter-gatherers, punishment is rare if not absent. What leads individuals to cooperate, Baumard argues, is the prospect of losing  their reputation as reliable partners.  Baumard then observes that as we move from hunter-gatherer groups to modern societies, reputation becomes less and less useful because people must cooperate with large numbers of virtual strangers. However, he argues, contemporary institutions, including firms, service organizations and government bureaucracies, take the place of informal reputational systems. In addition, he asserts, citing the work of Elinor Ostrom, groups spontaneously set up such associations and do so without the need for formal state support.

Baumard concludes that while we increasingly surrounded by strangers, our society has also appointed more and more strangers to help us – policemen, firemen, judges, teachers, journalists, epidemiologists, therapists, etc. “In fact,” he observes, “the welfare state now accounts from a third to half of GDP in most developed countries.”

I suggest that Dunbar is correct in stressing the centrality of third-party punishment of social deviants in maintaining social order in all known forms of human society. However, Dunbar views such punishment as the rational behavior of self-regarding individuals, and hence the incentive to punish is severely weakened in modern societies in which we are predominantly surrounded by strangers. In fact, there is a critical type of third-party punishment found in all known human societies that is based on moral values and is carried out by other-regarding individuals and coalitions of such individuals even though it is personally costly to the punishers. I believe Dunbar is simply incorrect in believing that contemporary societies generally have a problem sustaining a social morality, although of course there are communities and even whole nations with low ‘social capital’ that function poorly.

I suggest also that Baumard’s analysis is not well founded at all. The literature to which he refers denying the importance of altruistic punishment is incorrect. I and my coauthors, in our Behavioral and Brain Sciences commentaries, have shown that Guala’s (2012) and Baumard, Andre and Sperber’s (forthcoming) critiques are not persuasive.

In fact, human beings are not purely self-regarding, but rather have moral preferences that often lead them, in situations involving unrelated other human beings to cooperate, and to punish non-cooperators, at personal cost. The evidence for this is the high frequency of cooperation and punishment in social situations in which there is anonymity, or there is no repetition and hence no long-term reputational effect (so-called ‘one-shot’ interactions), or the probability of repetition is low. In these situations, self-interest cannot explain the observed prosocial behavior.

Christopher Boehm’s systematic studies Hierarchy in the Forest (2000) and Moral Origins: the Evolution of Virtue, Altruism, and Shame (2012), carefully document the importance of third-party punishment in extant hunter-gatherer societies. In his newer book, Boehm located 150 simple hunter-gatherer societies. Boehm coded fifty of these societies from around the world.  Despite the fact that these societies have faced highly variable ecological conditions, Boehm finds that their social organization maintains an egalitarian social order by means of the collective punishment of ‘bullies,’ and  they subscribe to a common human social morality, operating through internalized norms, so that individuals act prosocially because they value moral behavior for its own sake and would feel guilty behaving otherwise.

More generally, Edward O. Wilson’s The Social Conquest of Earth tells the story of human eusociality, in which other-regarding preferences and generalized within-group altruism are key to our success as a species in fitness-enhancing cooperation. Wilson summarizes his analysis of human evolutionary success as follows: “All normal people are capable of true altruism. We … attend to the sick and injured, help the poor, comfort the bereaved, and even willingly risk our own lives to save strangers… Authentic altruism… enhances the strength and competitiveness of groups, and it has been favored during human evolution by natural selection at the group level.”

For evidence from behavioral game theory analyzing altruistic reward and punishment, I refer the reader to my book with Samuel Bowles, A Cooperative Species (2011), and Joseph Henrich, Robert Boyd, Samuel Bowles, Colin Camerer, Ernst Fehr and Herbert Gintis, Foundations of Human Sociality: Economic Experiments and Ethnographic Evidence from Fifteen Small-Scale Societies (2004) and the references therein.

Baumard’s assertion that institutions can stabilize social cooperation with completely self-regarding agents who care about their reputations is obviously incorrect. Baumard refers to the welfare state as one such institution, but in democratic countries, voters determine the size of the welfare state, and the observed size cannot be explained by the actions of self-interested voters, who would never vote for a redistribution to the poor.

More generally, a political democracy would be infeasible if its members were purely self-regarding. In large democratic elections, the rational self-regarding agent will not vote because the costs of voting are positive and significant, but the probability that one vote will alter the outcome of the election is vanishingly small. Thus the personal gain from voting is vanishingly small. For similar reasons, if one chooses to vote, there is no plausible reason to vote on the basis of the impact of the outcome of the election on one’s self-regarding gains. It follows also that the voter, if rational, self-regarding, and incapable of personally influencing the opinions of more than a few others, will not bother to form opinions on political issues, because these opinions cannot affect the outcome of elections. Yet people do vote, and many do expend time and energy in forming political opinions. This behavior does not conform to Baumard’s story.

It is a short step from the irrefutable logic of self-regarding political behavior that rational self-regarding individuals will not participate in the sort of collective actions that are responsible for the growth in the world of representative and democratic governance, the respect for civil liberties, the rights of minorities and women in public life, and the like. In the self-interest model, only small groups of socially dominant individuals will act politically. Yet modern egalitarian political institutions are the result of such collective actions. This behavior cannot be explained by the self-interest model.

Except for professional politicians and socially influential individuals, contrary to the implications of Baumard’s theory, electoral politics is a vast morality play to which models of the self-regarding actor are a very poor fit.

Defenders of the self-interest model may respond that voters believe their votes make a difference, however untenable this belief might be under logical scrutiny. Indeed, when asked why they vote, voters’ common response is that they are trying to help get one or another party elected to office. When apprised of the illogical character of that response, the common reply is that there are in fact close elections, where the balance is tipped in one direction or another by only a few hundred votes. When confronted with the fact that one vote will not affect even such close elections, the common repost is that “Well, if everyone thought like that, we couldn’t run a democracy.”

Politically active and informed citizens appear to operate on the principle that voting is both a duty and prerogative of citizenship, an altruistic act that is justified by the categorical imperative: act in conformance with the morally correct behavior for individuals in one’s position, without regard to personal costs and benefits. Such mental reasoning, which has been called ‘shared intentionality,’ is implicated in many uniquely human cognitive characteristics, including cumulative culture and language. Shared intentionality rests on a fundamentally prosocial disposition, not self-interest.

The model of human strategic interaction on which my argument is based suggests that the human capacities for thriving in a society of strangers was laid down long ago in the period of our evolutionary emergence, and is part of the passage of hominins from ape-like social organization to full sociality based on an evolved social morality in which individuals behave morally under normal conditions because it is the right thing to do, not because it serves their narrow self-interest.

It is also true that moral behavior is often highly rewarded in human society, which is doubtless why we teach our children to be moral. Indeed, classical philosophers, including Aristotle, never question but that being virtuous was a critical part of human flourishing. Nevertheless, it is a key aspect of moral behavior that one acts appropriately even in situations where the personal costs are high and even extreme. It is this aspect of human morality that accounts for our success as a species.

Herbert Gintis
Santa Fe Institute and Central European University

Published On: May 12, 2012

Herbert Gintis

Herbert Gintis

Herbert Gintis (Ph.D. in Economics, Harvard University, 1969) is External Professor, Santa Fe Institute, and Professor of Economics, Central European University. He and Professor Robert Boyd (Anthropology, UCLA) head a multidisciplinary research project that models such behaviors as empathy, reciprocity, insider/outsider behavior, vengefulness, and other observed human behaviors not well handled by the traditional model of the self-regarding agent. His web site, www-unix.oit.umass.edu/~gintis, contains pertinent information. Professor Gintis published Game Theory Evolving (Princeton: Princeton University Press, 2000), and is coeditor, with Joe Henrich, Robert Boyd, Samuel Bowles, Colin Camerer, and Ernst Fehr, of Foundations of Human Sociality: Economic Experiments and Ethnographic Evidence from Fifteen Small-scale Societies (Oxford: Oxford University Press, 2004), and with Samuel Bowles, Robert Boyd and Ernst Fehr, Moral Sentiments and Material Interests: On the Foundations of Cooperation in Economic Life (Cambridge: MIT Press, 2005). He is currently completing a book with Professor Bowles entitled A Cooperative Species: Human Reciprocity and its Evolution.


  • tmtyler says:

    Voting probably was adaptive for our ancestors. The “evolutionary legacy” hypothesis probably has some truth to it here.

    However, politics is a high-status activity. Political activity by commoners could well be to do with signalling status and affiliations with impressive folk – and it is hard to be politically active and then not vote. Votling in large elections would then be like a pleiotropic side effect.

  • tmtyler says:

    I note that this post doesn’t consider the “anthropic” aspect of voting – as expressed in the article: “Why Voting May Not be Such a Stupid Idea”. You have to account for that to have an accurate view on the topic, really.

  • Peter Turchin says:

    On the issue of whether humans engage in moralistic punishment in pre-state societies, I tend to side with Herb Gintis. Particularly extreme forms of punishment (exile, capital punishment) should be rare, because they are so drastic. Chris Boehm, in an e-mail sent to me on this issue, makes the point that in small-scale societies, studied by anthropologists, even rare observed such events would be very significant in per capita terms.

    A recent study by Sarah Mathew on the Turkana indicates that free-riders (those who desert during cattle raids against other ethnic groups) are punished verbally, monetary, and corporal (she shows a photo of scars left from a severe beating). About half of deserters are punished.

  • 1. On punishment in small-scale societies, I won’t argue further. For those who are interested by the data, I strongly recommend Guala’s article.


    In particular, here is his answer to Gintis’ commentary:

    “Gintis & Fehr write that “anthropologists have confirmed
    that strong reciprocity is indeed routinely harnessed in the
    support of cooperation in small-scale societies.” Without
    further argument or justification, this is just the same
    claim they have repeatedly made in previous publications,
    and which my target article challenges. Surprisingly,
    Gintis & Fehr cite in support the same ethnographic litera-
    ture that I claimed they have misreported in previous
    work. The only new entry is Henrich et al. (2010a),
    which is not a field study but reports the results of cross-
    cultural experiments – perpetuating one of the misleading
    confusions between field and experimental data that I try
    to dispel in my article.

    Of the old literature, Gintis & Fehr keep citing the
    work of Boehm (1984, 2000) and Wiessner (2005, 2009).
    In my article I argue that the evidence reported in these
    studies does not support the costly punishment story.
    (One of the articles by Wiessner [2009], by the way, says
    so explicitly.) The commentaries published in this issue
    of BBS support my interpretation: Wiessner agrees that
    “experimental and ethnographic evidence do not concur”
    (see her Abstract), and Boehm similarly claims that “the
    costs do not necessarily fit with assumptions made in
    models that consider punishment to be altruistic”
    (Abstract). (…)

    Finally (and ironically) Gintis & Fehr refer to Henrich
    & Chudek’s commentary as a source of evidence in favour
    of the importance of costly punishment in small societies.
    But as we have seen (sect. R2), Henrich & Chudek sub-
    scribe to a much broader interpretation of the strong reci-
    procity programme, in which costly punishment does not
    play a prominent role. In fact, in their commentary
    Henrik & Chudek explicitly say that “models relying on
    DCP [diffuse costly punishment] . . . are not consistent
    with how norms are actually stabilized in small-scale

    (About Peter Turchin’s mention of Mathew and Boyd’s article on the Turkana, I can only points to my response with Pierre Liénard (who have worked there for more than 15 years): http://www.sscnet.ucla.edu/anthro/faculty/boyd/Baumard-Lienard-PNAS-2011-Comment.pdf)

    2. On public policies: of course, the choice is not between the idea that humans are selfish and the idea that humans are morals. Everyone — including me or Ostrom — think that humans are morals. The choice is between the idea that humans have an altruistic morality (Gintis uses “strong reciprocity”) and are ready to sacrifice for the society (for instance by punishing others), and the idea that humans have a mutualistic or a contractualistic morality (Guala says “weak reciprocity), and are ready to do their fair share, in particular in voting and financing common good (but not in sacrificing their welfare)

    In misrepresenting other’s views, Gintis is not doing a service to the community. By portraying his adversaries as fools who believe that humans are selfish, Gintis just obscures the debate and prevents people to understand that the debate is not between those who believe that people are sociopaths and those happy few who have come to realize that humans are truly morals, but between policies based on coordinated institutions and policies based on costly punishment.

    Again, I’m going to quote Guala’s paper, because I think he clearly describes what is a stake:

    “Strong reciprocity theorists view punishment
    as local, costly, and uncoordinated. The empirical literature
    instead reports mainly the emergence of local, cheap, and
    coordinated punishment institutions. Both solutions to the
    dilemma of cooperation differ in part from the traditional
    imposition of external sanctions administered by the state,
    and both can be seen as raising second-order social
    dilemma problems. However, they also have rather differ-
    ent properties and should not be treated as if they were
    identical: The devil, in institutional design as in almost
    everything else, is very much in the details.”

  • tmtyler says:

    Voting can be usefully compared to playing bingo. People engage in curious and apparently-illogical behaviours in environments that are far from the ones their ancestors evolved to deal with.

    Voting isn’t *particularly* critical to effective large-scale political organisation – as China demonstrates. It does provide voters with an illusion of control and involvement – helping *slightly* to prevent revolutions – but that’s about it. Also, if the number of voters went down, the benefits of voting would go up. Democracy would still be technically possible – even if there was a lot more political apathy.

  • Peter Turchin says:

    This has become a very interesting discussion. However, I’d like to ask all participants to stay civil to each other! Let us resolve (or not) the issues by appealing to logic and data, rather than by ad hominem attacks.

    In general, the positions of Nicolas and Herb are not quite as diametrically opposed as may appear to a reader not steeped in this theory. Neither is advocating an extreme version of the argument. Nicolas doesn’t argue that humans are purely self-regarding rational agents, but Herb also does not advocate the position that “Strong reciprocity theorists view punishment as local, costly, and uncoordinated” (Guala 2012). As much as I dislike ‘strong reciprocity’ as the term (see my post on it), let’s not oversimplify the message of Gintis and Bowles. In particular, Herb said in his Focus Article: “These capacities allow us to formulate general norms of social conduct, to erect social institutions regulating this conduct, to communicate these rules and they entail in particular situations, to alert others to their violation and to organize coalitions to punish the violators.” Certainly Herb and Sam know that coordinated punishment via coalition formation makes it (punishment) much cheaper. If memory serves, they even discuss an appropriate model in their book.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.