The Anatomy of Friendship in a “Digital Age”

Why is the number of friendships that we can actively maintain limited to 150? The evolutionary psychologist and anthropologist Robin Dunbar at the University of Oxford is a pioneer in the study of friendship. Over several decades, he and his colleagues have investigated the nature of friendship and social relationships in non-human primates and humans. His research papers and monographs on social networks, grooming, gossip and friendship have accumulated tens of thousands of academic citations but he may be best known in popular culture for “Dunbar’s number“, the limit to the number of people with whom an individual can maintain stable social relationships. For humans, this number is approximately 150 although there are of course variations between individuals and also across one’s lifetime. The expression “stable social relationships” is what we would call friends and family members with whom we regularly interact. Most of us may know far more people but they likely fall into a category of “acquaintances” instead of “friends”. Acquaintances, for example, are fellow students and colleagues who we occasionally meet at work, but we do not regularly invite them over to share meals or swap anecdotes as we would do with our friends.

Dunbar recently reviewed more than two decades of research on humans and non-human primates in the article “The Anatomy of Friendship” and outlines two fundamental constraints: Time and our brain. In order to maintain friendships, we have to invest time. As most of us intuitively know, friendship is subject to hierarchies. Dunbar and other researchers have been able to study these hierarchies scientifically and found remarkable consistency in the structure of the friendship hierarchy across networks and cultures. This hierarchy can be best visualized as concentric circles of friendship. The innermost core circle consists of 1-2 friends, often the romantic partner and/or the closest family member. The next circle contains approximately 5 very close friends, then progressively wider circles until we reach the maximum of about 150. The wider the circle becomes, the less time we invest in “grooming” or communicating with our friends. The social time we invest also mirrors the emotional closeness we feel. It appears that up to 40% of our social time is invested in the inner circle of our 5 closest friends, 20% to our circle of 15 friends, and progressively less. Our overall social time available to “invest’ in friendships on any given day is limited by our need to sleep and work which then limits the number of friends in each circle as well as the total number of friendships.

The Circles of Friendship – modified from R Dunbar, The Anatomy of Friendship (2018)

The second constraint which limits the number of friendships we can maintain is our cognitive capacity. According to Dunbar, there are at least two fundamental cognitive processes at play in forming friendships. First, there needs to be some basis of trust in a friendship because it represents implicit social contracts, such as a promise of future support if needed and an underlying promise of reciprocity – “If you are here for me now, I will be there for you when you need me.” For a stable friendship between two individuals, both need to be aware of how certain actions could undermine this implicit contract. For example, friends who continuously borrow my books and seem to think that they are allowed to keep them indefinitely will find that there are gradually nudged to the outer circles of friendship and eventually cross into the acquaintance territory. This is not only because I feel I am being taken advantage off and the implicit social contract is being violated but also because they do not appear to put in the mental effort to realize how much I value my books and how their unilateral “borrowing” may affect me. This brings us to “mentalizing”, the second important cognitive component that is critical for stable friendships according to Dunbar. Mentalizing refers to the ability to read or understand someone else’s state of mind. To engage in an active dialogue with friends not only requires being able to read their state of mind but also infer the state of mind of people that they are talking about. These levels of mentalizing (‘I think that you feel that she was correct in …..) appear to hit a limit around four or five. Dunbar cites the example of how at a gathering, up to four people can have an active conversation in which each person is closely following what everyone else is saying but once a fifth person joins (the fifth wheel!), the conversation is likely to split up into two conversations and that the same is true for many TV shows or plays in which scenes will rarely depict more than four characters actively participating in a conversation.

Has the digital age changed the number of friends we can have? The prior research by Dunbar and his colleagues relied on traditional means of communication between friends such as face-to-face interactions and phone calls but do these findings still apply today when social media such as Facebook and Twitter allow us to have several hundred or even thousands of “friends” and “followers”? The surprising finding is that online social networks are quite similar to traditional networks! In a study of Facebook and Twitter social media networks, Dunbar and his colleagues found that social media networks exhibit a hierarchy of friendship and numbers of friends that were extremely similar to “offline” networks. Even though it is possible to have more than a thousand “friends” on Facebook, it turns out that most of the bidirectional interactions with individuals are again concentrated in very narrow circles of approximately 5, 15 and 50 individuals. Social media make it much easier to broadcast information to a broad group of individuals but this sharing of information is very different from the “grooming” of friendships which appears to be based on reciprocity in terms of building trust and mentalizing.

There is a tendency to believe that the Internet has revolutionized all forms of human communication, a belief which falls under the rubric of “internet-centrism” (See the article “Is Internet-Centrism a Religion“) according to the social researcher Evgeny Morozov. Dunbar’s research is an important reminder that core biological and psychological principles such as the anatomy of friendship in humans have evolved over hundreds of thousands of years and will not be fundamentally upstaged by technological improvements in communication. Friendship and its traditional limits are here to stay.

Reference

Dunbar R.I.M. (2018). “The Anatomy of Friendship” Trends in Cognitive Science 22(1), 32-51

 

Note: An earlier version of this article was first published on the 3Quarksdaily blog.

Shame on You, Shame on Me: Shame as an Evolutionary Adaptation

Can shame be good for you? We often think of shame as a shackling emotion which thwarts our individuality and creativity. A sense of shame could prevent us from choosing a partner we truly love, speaking out against societal traditions which propagate injustice or pursuing a profession that is deemed unworthy by our peers. But if shame is so detrimental, why did we evolve with this emotion? A team of researchers led by Daniel Sznycer from the Center for Evolutionary Psychology at the University of California, Santa Barbara recently published a study in the Proceedings of the National Academy of Sciences which suggests that shame is an important evolutionary adaptation. According to their research which was conducted in the United States, Israel and India, the sense of shame helps humans avoid engaging in acts that could lead to them being devalued and ostracized by their community.

A Belgian Iron ‘scolds bridle’ or ‘branks’ mask, with bell, used to publicly humiliate and punish, mainly women, for speaking out against authority, nagging, brawling with neighbors, blaspheming or lying via Wellcome Images

For their first experiment, the researchers enrolled participants in the USA (118 participants completed the study; mean age of 36; 53% were female) and India (155 participants completed the study, mean age of 31, 38% were female) using the online Amazon Mechanical Turk crowdsourcing platform as well as 165 participants from a university in Israel (mean age of 23; 81% female). The participants were randomly assigned to two groups and presented with 29 scenarios: The “shame group” participants were asked to rate how much shame they would experience if they lived through any given scenario and whereas the “audience group” participants were asked how negatively they would rate a third-party person of the same age and gender as the participants in an analogous scenario.

Here is a specific scenario to illustrate the study design:

Male participants in the “shame group” were asked to rate “At the wedding of an acquaintance, you are discovered cheating on your wife with a food server” on a scale ranging from 1 (no shame at all) to 7 (a lot of shame).

Female participants in the “shame group” were asked to rate “At the wedding of an acquaintance, you are discovered cheating on your husband with a food server” on a scale ranging from 1 (no shame at all) to 7 (a lot of shame).

Male participants in the “audience group”, on the other hand, were asked to rate “At the wedding of an acquaintance, he is discovered cheating on his wife with a food server” on a scale ranging from 1 (I wouldn’t view him negatively at all) to 7 (I’d view him very negatively).

Female participants in the “audience group” rated “At the wedding of an acquaintance, she is discovered cheating on her husband with a food server” on a scale ranging from 1 (I wouldn’t view her negatively at all) to 7 (I’d view her very negatively).

To give you a sense of the breadth of scenarios that the researchers used, here are some more examples:

You stole goods from a shop owned by your neighbor.

You cannot support your children economically.

You get into a fight in front of everybody and your opponent completely dominates you with punch after punch until you’re knocked out.

You receive welfare money from the government because you cannot financially support your family.

You are not generous with others.

For each of the 29 scenarios, the researchers created gender-specific “shame” and “audience” versions. The “audience group” reveals how we rate the bad behavior of others (devaluation) whereas the “shame group” provides information into how much shame we feel if we engage in that same behavior. By ensuring that participants only participated in one of the two groups, the researchers were able to get two independent scores – shame versus devaluation – for each scenario.

The key finding of this experiment was that the third-party devaluation scores were highly correlated with the shame scores in all three countries. For example, here are the mean “shame scores” for the wedding infidelity scenario indicating that people in all three countries would have experienced a lot of shame:

USA: 6.5

India: 5.7

Israel: 6.7

The devaluation scores from the third-party “audience group” suggested that people viewed the behavior very negatively:

USA: 6.4

India: 5.1

Israel: 6.6

For nearly all the scenarios, the researchers found a surprisingly strong correlation between devaluation and shame and they also found that the correlation was similarly strong in each of the surveyed countries.

The researchers then asked the question whether this correlation between personal shame and third-party negative valuation was unique to the shame emotion or whether other negative emotions such as anxiety or sadness would also correlate equally well with devaluation. This experiment was only conducted with the participants in the USA and India. The researchers found that even though the fictitious scenarios elicited some degree of anxiety and sadness in the participants, the levels of anxiety or sadness were not significantly correlated with the extent of devaluation. The researchers interpreted these results as suggesting that there is something special about shame because it tracks so closely with how bad behavior is perceived by others whereas sadness or anxiety do not.

How do these findings inform our view on the evolutionary role of shame? The researchers suggest that instead of designating shame as an “ugly” emotion, it is instead an excellent predictor of how our peers would view our behaviors and thus deter us from making bad choices that could undermine our relationships with members of our community. The strong statistical correlations between shame and negative valuation of the behaviors as well as the universality of this link in the three countries indeed support the conclusions of the researchers. However, there are also so important limitations of these studies. As with many evolutionary psychology studies, it is not easy to ascribe a direct cause-effect relationship based on a correlation. Does devaluation lead to evolving a shame mechanism or is it perhaps the other way around? Does a sense of shame lead to a societal devaluation of certain behaviors such as dishonesty? It is also possible that the participants in the audience group responded with the concept of “shame” in the back of their mind even though they were not asked to directly comment on how shameful the act was. Perhaps their third-party assessments of how bad the behavior was were clouded by their own perceptions of how shameful the behavior would be if they themselves had engaged in it.

Another limitation of the study is that the participants represented a young subgroup of society. The mean ages of 23 (Israel), 31 (India) and 36 (USA) as well as the use of an online Amazon Mechanical Turk questionnaire means that the study results predominantly reflect the views of Millennials. The similarities of the shame and devaluation scores in three distinct cultures are among the most remarkable findings of these studies. However, perhaps they are more reflective of a global convergence of values among the Millennial generation than an underlying evolutionary conservation of an adaptive mechanism.

These limitations should not detract from the provocative questions raised by the studies. They force us to rethink how we view shame. Like all adaptive defense mechanisms, shame could go awry. Our immune function, for example, is an essential defense mechanism but an unfettered immune response can destroy the very body it is trying to protect. Perhaps shame acts in a similar fashion. A certain level of shame could help us function in society by promoting certain moral values such as justice, honesty or generosity. But an excess of shame may become a maladaptive prison which compromises our individuality.

References:

Daniel Sznycer, John Tooby, Leda Cosmides, Roni Porat, Shaul Shalvi, and Eran Halperin. (2016). “Shame closely tracks the threat of devaluation by others, even across cultures” Proceedings of the National Academy of Sciences

Sznycer D, Tooby J, Cosmides L, Porat R, Shalvi S, & Halperin E (2016). Shame closely tracks the threat of devaluation by others, even across cultures. Proceedings of the National Academy of Sciences of the United States of America PMID: 26903649

Note: An earlier version of this article was first published on the 3Quarksdaily blog.