Published on Let's Talk Development

The origins of social boundaries

This page in:

This is the first in an occasional series of blogs on social boundaries and identity. I’m interested in the topic for obvious reasons. Social boundaries and identities, at least in some forms (and that is the rub!) have been argued to affect generalized trust and/or prejudice, governance and cooperation, and development outcomes. They may also be relevant to certain recent political developments. Here at the World Bank, the Mind, Behavior, and Development Unit (eMBeD) is involved in projects that aim to support social cohesion.

Let’s start at the beginning – where do social boundaries come from? In other words, are they a natural or inevitable part of human cognition? Does the widespread creation of in-groups and out-groups arise from evolutionary necessity? If so, there may be limits as to how much the #MeToo movement can accomplish, the extent to which social injustices in countries like South Africa and India can be rectified, and the ability of bodies like the United Nations to one day achieve a world without warfare. Or are the categories of in-group and out-group learned and socially constructed, in which case there may be more room for change, interventions, and optimism?

One way to examine the naturalness of social boundaries is to examine their use by young children, who have undergone less socialization than adults. I recently came across two papers that do this in creative ways.

Noyes and Dunham (2017) showed children ages 4-9, recruited from science museums in Connecticut, pictures of two imaginary social groups, the Flurps and Zazzes. Experimenters told the kids the names of the groups; and the groups were separated, so that group members were seen to be playing only with members of their own group. In the baseline condition, there were no other physical differences between the groups. The key question the experimenters asked was: when one person wants to join the other group, and when the other group welcomes him or her, is that person indeed a member of the other group? For example, “Mark says he wants to be a Flurp. Mark tells the Flurps he wants to be a Flurp. The Flurps tell Mark he is a Flurp. Is Mark a Flurp?” The main finding was that for younger as well as older kids, these kinds of mutual intentions were sufficient for Mark to become a Flurp, but individual-only and group-only intentions were not: If Mark wanted to join but the Flurps would not have him, or if the Flurps wanted him and Mark did not want to join, Mark was not considered a Flurp.

Follow-up experiments showed: a) mutual intentions of this sort are sufficient for group membership even when the Flurps and Zazzes were wearing clothing of different colors; however, b) mutual intentions did not create group membership when the Flurps and Zazzes consisted individuals of visibly different genders; c) common knowledge of mutual intentions was more powerful than private mutual intentions. Overall, the paper shows that, among young children, mutual intentions are powerful sources of group membership for novel social groups, but that there is at least one social category, gender, for which mutual intentions do not suffice. Children at young ages, the authors argue, employ essentialist strategies when thinking about gender.

Diesendruck et al (2013) examined whether there are differences in the types of social categories that U.S. and Israeli children tend to essentialize. Children were shown pictures of two individuals visibly different along only one social category (e.g., black man and white man, Jewish man and Arab man, white woman and white man, female doctor and female policewoman, cat and dog, saw and manner). They were then shown a third individual who resembled one of the two previous pictures, with respect to the relevant social category, but not the other. Almost all the children correctly identified the social match. The children were then told that there exists some place in the world where people think oppositely from us and say that these two are of the same kind (e.g., a black man and a white man are considered to be of the same kind). “Do you think it is wrong how they do it, and that we should correct them and teach them to do it like us? Or is it OK the way they do it, just different from us?”

As in the previous study, the authors found that children in both societies view gender in essentialist terms, less so occupation. The authors contend that categories with arguably high evolutionary functionality, like gender, are primary candidates for essentialism. Notably, children from both countries did not hold a highly essentialist view of race; older children in Israel were less likely to view race in essentialist terms than younger children; but in the U.S., age was positively correlated with race essentialism. With respect to ethnicity (e.g., Jew or Arab), Israeli children viewed differences as objectively correct by kindergarten, and strongly rejected a “culturally relative” view of ethnicity.

Together, the two papers suggest that gender essentialism may run deep. On the other hand, for other social boundaries, like ethnicity and race, there is evidence that culture and social learning loom large, in which case there may be room for change. The latter finding is good news, at least for people who, like me, support a kind of cosmopolitanism in which we gradually expand our moral circle. It may be the case, as Steven Pinker has argued, that we have a “module” in our brains, not necessarily to favor our own race or ethnicity, but our coalition, and that our coalitions are, at least in part, of our own making:

The psychologists Robert Kurzban, John Tooby, and Leda Cosmides point out that in human evolutionary history members of different races were separated by oceans, deserts, and mountain ranges (which is why racial differences evolved in the first place) and seldom met each other face to face. One’s adversaries were villages, clans, and tribes of the same race. What looms large in people’s minds is not race but coalition; it just so happens that nowadays many coalitions (neighborhoods, gangs, countries) coincide with races.


Varun Gauri

Senior Economist, Development Research Group, World Bank

Join the Conversation

The content of this field is kept private and will not be shown publicly
Remaining characters: 1000