The 150 of Us – An Interview with Robin Dunbar, Honorary Member of MTA, about the Social Network Size Number of Humans

https://www.flickr.com/photos/mtasajto/albums/72177720303191661

The title of your inaugural lecture is: “The Neuro-Anatomy of Friendship and Community.” How can we study the anatomy behind the social skills of humans?

Robin Dunbar Photo: Wikimedia Commons

These days we mainly use brain scanning – neuroimaging – techniques for this. These techniques allow us to map which bits of the human brain are active when someone engages in a social interaction or even when someone merely thinks of other people’s state of mind. Furthermore, we can also study the correlation between the size of people’s personal social network – including friends and family for instance – and the size of certain parts of the human brain. For that matter, it has been found that the so-called default mode network in the human brain – an extensive neuronal network, which had been considered the most active part of the brain when you are daydreaming – works most intensively when one thinks of their friends and family. Perhaps it has to do with the fact that when you are daydreaming, you are actually thinking about other people doing things and about what they might be thinking. Anyhow, a large part of the brain is dedicated to the maintenance of one’s social world. In this process, the neocortex – that is the more recently evolved parts of the cerebral cortex – the limbic systems and the cerebellum all have a role to play.

Does this mean that a significant part of our brain is involved in managing our social relationships?

That is correct; a big part of the neocortex is there for this function. In the course of primate evolution, the neocortex has expanded dramatically. In mammals in general, the neocortex normally accounts for 10-40 percent of the total volume of the brain, while in primates it starts at 50 percent, and increases to 80 percent in humans. An essential function of the human brain – and of the primate brain, for that matter – is to manage a large number of social relationships so as to keep social groups coherent, preventing them from breaking up and fragmenting into smaller units. The size of primates’ groups is adjusted so as to enable them to defend themselves against groups of predators or against groups of individuals of their own species, therefore it is vital for them to keep their own group together. Other mammals, like various antelope species, may also form large groups, but these tend to be less stable, regularly admitting new members to the herd and letting others wander away. With primates, this is different though: their groups tend to remain quite stable over time.

What is the cause of this stability?

The fact that there is a permanence in terms of group cohesion can be put down to the “friendships” that evolve from the strong bonds forged among individual members of the group. But taking care of such a large number of connections is a huge cognitive task. Thus, it can be concluded that in the case of primates there was an external ecological constraint which required the maintenance of permanent and relatively large groups, which, in their turn, required the cognitive management of extremely large numbers of social connections. This explains the major growth in the size of the neocortex. This had led me to look into the eventual correlation between primates’ neocortex size and their social group size. What I found was in fact a solid correlation between them. Based on this correlation I made a prediction of the effective group size for humans in relation to their neocortex. The number I got showed an average around 150 – Dunbar’s number, as it came to be known.

Is there any other evidence to the existence of Dunbar’s number?

Yes, quite a lot. Once we got the number from the correlation, we started looking for possible evidence in all sorts of areas to see whether natural human groups did have a size of this magnitude, whether real-life data did support our theory. It turned out to be the case that the typical community size in hunter-gatherer societies, that is, pre-historical small-scale traditional societies before their settling in villages, did correspond to Dunbar’s number, and that we could also discover the same number of 150 in terms of the size of friendship and family networks of people of our time. By now there have been loads of data gathered which suggest that this figure of 150 is a very constant value in human organisations of all kinds. Take, for example, the size of Italian Alpine villages over a 600-year period, that is, from 1250 AD to 1850 AD; you will find that the size of the villages in one particular valley in the Alps remained, on average, about 145 despite the fact that the overall population was increasing through time. So in order to keep the community size around 150 people while there was a population growth in general, there was a tendency to split the villages. This shows that this figure is extremely widespread, it can be found in very many different contexts. Network science has shown that the size of human information networks tends to gravitate toward this figure of 150 because it optimises the flow of information within a network. If you have a network which is either smaller or larger, information flow within there seems not to be as efficient.

In recent years you have had publications about other “Dunbar’s numbers”, claiming that our social contacts form layers – concentric circles, as it were – and the larger a circle is, the bigger the number of acquaintances it contains and the looser our contact with these people tends to become. How can you square this with the classical Dunbar’s number, 150?

It still holds true that the number of people with whom we are able to maintain a meaningful relationship is 150. But this number has been found to be fractally structured, capable of being broken down into circles, which contain, respectively, the closest, the best, the good and the ordinary friends. The circles have a scaling ratio with respect to each other of approximately three, which means that there are about five close friends or family members, 15 good friends, 50 ordinary friends, and so on, but the reason for all this still remains unclear. As you move farther away from the innermost circle and go further out through the layers, there is a decreasing degree of altruism, and a decline in the willingness to help and in the emotional quality of the relationship. Measurements show that we devote 60 percent of our total social effort, that is, the time we spend on social interactions, to our 15 good friends; in other words, our sympathy group. All these layers make up our network of 150 contacts. This structure turns up in various kinds of human social networks as well as in numerous human organisations. If we look at the frequency with which people call one another on the phone, we will be able to see this pattern of layers of social contacts quite nicely. They constitute the basis for the organisation of all modern armies, for instance. The regular units within armies such as companies, platoons, battalions, regiments and so on, are structured along similar principles.

How can you explain this?

It seems that the military planners and organisers discovered these Dunbar’s numbers themselves a long time ago by trial and error. You can see that over the course of about three-hundred years of military history, they had been trying different unit sizes, and about a hundred years ago they eventually settled on these numbers, this layered structure of human social relations. They just experienced on the battlefield that these numbers worked and that if they grouped too big a number or too small a number of troops into a unit, the cooperation was not efficient and everybody got killed. So they had to arrive at the optimised unit size through tough learning. But of course they didn’t know why these numbers worked, they just found out that they did work.

In online social media networks, some people have more than 150 “friends”. Does this suggest that the numbers are not realistic or that we can manage many more friendships through the internet?

We have actually been looking into this phenomenon closely because this allows us to see whether the constraint of 150 has something to do with the nature of the human mind or with external constraints, above all with the limited amount of time available. Maintaining friendships and close human relationships takes time and energy. People have limited time for social interactions, which prevents them from keeping contact with an unlimited number of people. If the time constraint is really what matters, then we could argue that it is possible on Facebook, for instance, to keep up a meaningful connection with more than 150 people, as this does not take so much of our time. On Facebook, it takes the same amount of effort to speak to several hundred people as it takes to speak to just one. In real life, though, it would be extremely difficult to get 100 acquaintances of ours together for a beer. Our research has shown that there are indeed people with a large number of acquaintances, but the average number again turns out to be at around 150. Facebook has already transmitted to us four huge samples of data, the biggest sample being a collection of interaction data for 61 million Facebook profiles, and the data showed that the common user in the online space would have a meaningful relationship with exactly 149 other persons. Needless to say, as in the real world, so on Facebook, there will be a great deal of variability in people’s behaviours. Our findings do not claim that each and every one of us has exactly 150 acquaintances. Some of us have very many, others will do with only a handful, but the average is around 150.

How do you think people having well above the average number of friends are able to cope with this quantity?

We assume that they extend their 150-strong personal social network to include their online friends as well. These online friends are people they know only by way of the internet, or they are just colleagues with whom they are having contact on work matters, being on good terms, but they would not invite them home to a party. And the reason they would not invite them is that they fall outside the layer of 150, that is, of the people who qualify as meaningful contacts. These online friends would in the everyday world correspond to the barista you’ll buy your latte from at the station every morning, with whom you’ll only have a few words of conversation, you’ll discuss the weather if you are in Britain, he knows you and you know who he is, so he is a familiar person, but in fact you hardly know anything more about him. The algorithms of Facebook encourage us to reach out to more and more people, which in turn helps us extend our own social network. But if you look at the frequency of your communication with your real friends on Facebook, in particular in terms of named postings and named replies, it is apparent that beyond the layer of 150, you will barely care about anyone. Or, if you do, this is not real communication between you and them, it is rather being a voyeur on others’ social life.

Two years ago, during the COVID lockdowns, people all over the world were forced to separate from one another and could not meet face-to-face. There was only online contact left, such as Zoom or Facebook. Is it possible to maintain close friendships via online channels without personal meetings?

It is possible to maintain contacts through the internet, but it is not easy. Research over the past two years found – as we ourselves could confirm based on our everyday experience – that online interaction, even with live video on, is not as satisfying as a face-to-face meeting. There is something special about being able to sit next to each other, even to touch each other, to talk to each other at any moment we want to. These elements are all very important in maintaining relationships. Zoom, Skype and similar media which enable video calls seem to work better between people who have a well-established relationship, especially if the conversation is taking place in a very small group. But if there are many participants to the online call, who do not know one another very well, the conversation may easily become sluggish. In the business world, digital media is not a very satisfactory way to introduce a new colleague either, it will just not have the same effect as being able to sit in the same room together when the introduction takes place. Many people found during lockdowns that if they did not get to see their friends as frequently as the layer those friends belonged to within their personal social network would require, they risked sliding into the next layer out, that is, becoming less and less close friends. This process may actually be very quick, taking only a few months. That caused people to re-evaluate their friendships in the course of lockdown, and they certainly lost some friends as a result of that (and perhaps found some new ones online) as they had come to feel over time that they did not any longer have a friendship with a particular person. Many people (especially women) found that not being able to be with their friends was very stressful, and the effects of loneliness became quite distressing for them and made them more anxious.