Sunday, April 12, 2009

the evo-psych of not being Larry Summers

I’ve just finished reading Robin Dunbar’s 1996 book, Grooming, Gossip, and the Evolution of Language. Dunbar is an Oxford anthropologist and evolutionary biologist, most well known for the ‘Dunbar number’: the hypothesis that the ratio of human neocortex to overall brain size predicts (successfully, it seems) stable human social groups of around 150 individuals. Although the book bears a Harvard University Press imprint, it reads at times like popular science, a bit scanty on citations and too-quick with some important arguments. Nevertheless, it’s quite fascinating to read; Dunbar attempts to explain human language (and high-level cognition) as an evolutionary solution to the problem of social friction in a species whose group size has outgrown the practical limits of bonding behavior in other primates (e.g. reciprocal grooming).

I want to riff on a few different points in the book, starting in this post with a point about gender differences in the use of language, and its relation to the status of women in intellectual professions. In chapter 9, Dunbar describes research conducted with colleagues, from which they conclude that “conversations often function as a kind of vocal lek. Leks are display areas where males gather to advertise their qualities as potential mates to the females.” (176) This idea is based upon data gathered from studies of actual human conversations. In one-on-one male/female dyadic conversations, men and women talk for more or less equal periods of time. However, in conversational groups of any larger size, men tend to dominate air time. This is true even when one particular man and one particular woman break off to form a dyadic sub-conversation on the periphery of the main group; the man will tend to speak much more than he would were the pair isolated.

Dunbar and his colleagues think that the lek hypothesis can explain this difference between male/female conversational behavior in dyads and in larger groups. In lek species, males engage in competitive displays to persuade observant females of their reproductive superiority. The females are welcome to sit back and merely observe; indeed, they must do so if they want to pay attention and accurately size up the males. Dunbar and colleagues hypothesize that a similar thing happens in human conversational groups; the men (unwittingly, most of the time) are driven to speak up, to out-compete one another for the attention of the women. Such competition is unnecessary when there are no other men around, hence in isolated dyads men and women share speaking time more evenly.

One more bit of data ties this into the status of women in intellectual professions. Dunbar and colleagues analyzed not only the total quantity of speaking time, but the frequency of particular conversational content. In general, all humans of any gender tend to talk mostly about social matters; about 65% of their talk-time centers around ‘who-is-doing-what-with-whom’. (That’s one bit of critical data behind Dunbar’s central hypothesis that language evolved primarily to facilitate gossip.) Dunbar observes that in single-gender conversational groups, certain relatively intellectual topics like “work and academic matters or religion and ethics” take up only 0-5% of total conversational time. However, in mixed-gender conversational groups, the proportion of male talk-time devoted to these topics rises significantly, with much less increase in corresponding female patterns.

Dunbar offers a somewhat sketchy evolutionary hypothesis to explain that last bit. In mixed-gender groups, discussion of relatively intellectual topics increases among men because they are a way of advertising intelligence, a type of reproductive advantage. Women have no need to advertise (they are the ones who get to be choosy about mates) and so no need to out-compete each other (or men) in conversational prowess. Hence, women are less likely to increase their conversational attention to abstract subjects. (It’s important to note that all of the underlying data comes from observation of casual conversations – like those in the lunch room or hallway – not from formal contexts as in the seminar or board room, where participants gather with a specific purpose in mind.) Following on all of this, Dunbar concludes the section with the following remarks, whose relevance to his preceding points about gender difference is left unclear:

In the intellectual world of the university, demonstrating your intellectual skills by showing off your knowledge of Kant or the Romantic poets, or by being able to explain yesterday’s lecture on the second law of thermodynamics, may be quite acceptable hallmarks of competence and status. They mark you out as a cut above the rest, the obvious choice in the mating stakes. In that kind of environment, intellectual prowess is as appropriate a criterion of future status or earning power as being the best card player in a bridge club or the best musician in a music club. Knowledge, as it is so often said, is power. (177)

Given all the attention he’s just paid to the relative preponderance of intellectual topics in male conversation, and their role as a competitive advertising mechanism, is Dunbar here attempting to explain the greater success of men in intellectual, academic fields? He doesn’t say – presumably he’s smart enough not to explicitly stick his leg into that political bear trap. But it certainly sounds as if he’s suggesting as much.

Now, here’s the thing. I think we can react to this with the sort of bristly outrage that usually greets evo-psych-fueled gender difference research in feminist circles. The research is reductivist and essentializing; it provides ammunition to those who would naturalize (and implicitly justify) oppressive stereotypes. Et cetera. But I don’t think that’s the correct reaction. I actually think this sort of research should be welcomed by feminists, provided that it rests on solid methodological grounds (about which I won’t judge here).

Notice that nothing in Dunbar’s story implies or entails that men are actually better at intellectual pursuits than are women. The account just explains why men are more prone to talk about them even in informal contexts, and why men are more prone to dominate conversations regardless of topic. So it is entirely consistent with the Dunbar story that we might insist there is no (presently relevant) evidence for suspecting latent intellectual superiority in males. Instead, the account offers us two different tools for correcting gender imbalance in the intellectual professions.

First, the account provides a potential debunking explanation for gender imbalance. We needn’t invoke (as Larry Summers once did) a natural superiority among men in certain intellectual areas. Rather, we need merely presume certain sociological facts about the intellectual professions, and add these to what Dunbar and colleagues tell us. Who is most likely to get ahead? Those who speak up in front of the professional power-brokers. Who is most likely to speak up? Men. Does tendency to speak up have anything necessarily to do with actual competence? No. So we may very well understand the relative success of men in intellectual fields as an incidental consequence of behaviors that are simply arbitrary as to ability.

I would go one step even further. If Dunbar et al. are right, then all of us – men and women alike – are hardwired to expect differential linguistic behaviors between genders. We expect men to be more dominating and intellectual (even in casual conversation), and we expect women to be more quiet, or attentive only to gossipy matters. We’re generally not aware of having these expectations, still less of why we have them. But have them we do – and as with all expectations, we are unnerved when they are not met. We are biologically conditioned to react appreciatively to dominating, intellectual discursive behavior from men, and to react with at least some measure of confusion to the same from women. Hence, in addition to the success-generating basic tendencies mentioned in the last paragraph, we see another explanatory factor for relative male domination of intellectual fields: we are disposed to reward intellectual success-seeking men, and to punish similar women. Indeed, this might explain why, until very recently, nearly every human culture excluded women from political and other intellectual discussion, on grounds of propriety.

Building upon these explanatory stories, Dunbar’s account offers us a practical tool for remediating gender imbalance. Consider your academic department ‘s or workplace’s resident alpha male blowhard. He seizes control of every conversation. He clearly experiences intellectual discourse as a form of combat. Merely pointing out to him that his behavior edges out other conversational participants probably has zero effect. He cannot understand why they don’t just talk more, if they want to!

But imagine now another strategy. Show him Dunbar’s research. Explain to him that his behavior is at least partly explicable as the bidding of his lek-frenzied ape brain. Ask: does he want to be the sort of creature whose manner of participation in an abstract, rarified environment conforms better to that of a horny chimpanzee? He is a conscious, reflective being; if he wants to, he can modify the output of his evolutionary programming through a regime of attentive habituation. And with this new insight, he might also try to correct inferences he makes about the competence of others based upon their inability or unwillingness to act just like him. Is such an argument likely to work on most blowhards? No - people’s self-concepts are extremely resilient. But getting the story out might have appreciable effects upon more moderate men, and more generally on everyone in authority, whose assessments of individuals’ conversational practices matters to their advancement in the profession.

If nothing else, here’s a take-away point. Evolutionary psychology, even gender-difference evolutionary psychology, needn’t always be a vehicle for unreflective stereotype reification. We – our brains, our behaviors, or social institutions - are products of naturally selective forces. Some of the outcomes of these evolutionary processes worked perfectly well for our primate forebears, but they have come out of sync with modern, reflective, socially-conscious humans. Evolutionary psychology can sometimes allow us to trace (speculatively, at least) the causal origins of these undesirable behaviors, and so might sometimes point us to the causal-psychological joints where they may be most efficiently dismantled.

image (c) Charles Burns at


  1. Regina,
    This peice is interesting. I got here via Feminist Philosophers. I have been thinking alot these days about ways to do socially relevant philosophy (there will be a Synthese issue coming out next year) and evpsy is a challenging case. If you are interested in talking that would be super. You can email me at Here is my webpage in case you want to know a little bit more about this out of the blue person:


  2. Hi Carla,

    thanks for writing!

    I'm leaving today for a weekend conference, but I'll definitely send you an email when I'm back next week.