The Dunbar number regularly gets bandied around in social media circles, and for good reason. However, it is usually misunderstood.

In today’s hyper-connected world, where technology lets us have hundreds (if not thousands) of ‘friends’, people are increasingly interested in understanding what the human limits on maintaining human friendships might be, and why.

Real world relationships have been studied by psychologists, sociologists and host of other ‘-ists’ for decades. However, Dunbar, who works in the area of behavioural brain science, has emerged as one of the most frequently quoted figures, in the blogosphere at least.

Where did Dunbar’s Number come from?

One of Dunbar’s papers, published in 1993, wonderfully titled “The Co-evolution of Neocortex Size, Brain Size and Language in Humans” is cited for something commonly referred to as Dunbar’s number. Shock number one: There isn’t really any such thing as Dunbar’s number, in the sense that people normally refer to it.

The common mythology is that Dunbar said that people can only sustain a network of 150 contacts. Strictly speaking that is not what Dunbar’s paper said. Think of humans as brains on legs for a minute, and put your evolutionary psychology hat on. Dundar argued that, in evolutionary terms, there may be an upper group size that animals can and will live in, determined by cognitive constraints – specifically the processing capacity of the neocortex – and selected for based on various environmental constraints.

Effectively, from an individual animal’s point of view, the neocortex size sets a limit on the number of relationships that can be maintained. That in turn limits the maximum group size for the community that individual lives within, assuming they have the same constraints. At a simple level, if a species is made up of individuals that can only sustain 10 relationships, we might expect to see groups of 10 wandering around. If another member attempts to join the group, the individuals don’t have the capacity to support that extra relationship, and the member wouldn’t be accepted, or the group would fail.

The neocortex size is driven by all sorts of ecological factors that select for group size, but we could potentially use that relationship the other way round to predict group sizes, based on the neocortex. Take one group with a known group size and look at the size of its neocortex. Work out a ratio, then take another species and look at the size of its neocortex and use that ratio to predict how many individuals it would be able to support.

Would the Real Dunbar’s Number Please Step Forward

Dunbar took existing data from a number of primate studies, where typical group sizes can be observed. He then looked at the neocortex size for those primates and projected forwards to the larger human neocortex. His calculations predicted that human group sizes would typically be around 147.8. It should also be noted that Dunbar worked with average group sizes, not maximum sizes.

Now, the maths is much more complex than this summary indicates, but I’ll spare you the detail. Even so, the statisticians out there are probably gagging on their most recent meal at this point. Hang on in there. Some statistical juggling means that confidence limits around this number can be calculated, which ends up giving a range being between 100 to 231, hence my earlier comment about Dunbar’s number not really existing in the way most people expect. Think of it as a range of typical group size, rather than a number limit. Dunbar’s work has been criticised and supported. However, it is interesting to note that his suggested number does seem to match with studies of human group size from other disciplines. I commonly read both military and business books that suggest a number in this range as the typical or maximum size for a group.

Speaking of Language

There is an interesting aside here. Dunbar and others argue that social grooming is important for maintaining relationships, and for sustaining the coalitions that facilitate large group structures. While apes might spend their time picking insects out of each other’s fur, we waggle our tongues and use our voice boxes. Evolutionary psychologists often argue that we evolved language as a very efficient form of social grooming. Well, half of them would. The other half would argue something more along the lines that our brains got so big and heavy we had to do something useful with them, and language turns out to be a rather beneficial thing to have.

Think of it this way: If a large group of apes is only able to be large because they spent their entire time picking nits out of each other’s fur, it won’t be a large group for very long. While they have the advantage of being able to defend themselves and pool resources, they don’t have any time left for finding food. They will be an extinct bunch of apes in very short order.

One of the factors that gives us large group structures is our ability to use language in communication. Language is much more efficient than picking nits out of fur. I can deal with more than one person at once, for a start. That means we can be more efficient about maintaining relationships, using quick bursts of language, rather than all of the time being taken up with social grooming. Personally I find that a great relief. I love communicating with you, but I’d rather not have you dealing with my parasites!

A simpler summary would be that available time, combined with efficiency, determines the number of relationships that can be supported. Language enables us to be more efficient with our time. That in turn enables us to build a larger social world, and still have time to do other things.

Does Social Media Make Us More Social?

One of the (many) aspects of social media of that fascinates me is this: Can it enable us to be more efficient and effective in maintaining relationships? Can computer-based tools enable our brain to cope with more than it would be able to otherwise? If so, that has social ramifications, as well as organisational design ones.

So, if Dunbar’s number (or one of the equivalents from Anthropological studies) is so small, how do we end up with significantly larger groups, like 1,000+ person companies? It comes down to rings of friendships. Think of bands of 30-50, then clans of 100-200, and above that tribal groups of between 500 and 3000. Imagine that I have 30-50 relationships, and those individuals have partially overlapping relationships with others. You can now imagine an inner group and an outer group, with cohesion maintained by those individuals holding relationships across the different bands and within bands.

The way that large groups work is significantly more complex than suggested here so far. Dunbar and others argue for these layers or rings of friendships, with different strengths at each layer. This layered structure enables sustainable group dynamics. The coalitions mentioned earlier are important, since these stronger relationships provide the individual with others to protect them from potential hostility from members of the larger group, by individuals with relationships to both parties.

Bands and clans interact in a way that protects individuals and sustains tribes and population, and reflect different types and strengths of relationship. One of the challenges of today’s social media is that it doesn’t model this subtlety and complexity. Psychologists are still trying to understand the diverse nature of human relationships, and the complex properties that they have.

The Future is Still Social

It may be many years before social media catches up with even today’s understanding, and by then that understanding may have moved on – potentially due to social media itself. Social networking tools let us understand how we maintain relationships, by giving us greater visibility into how people interact, but they also potentially change the way that we do these things, a kind of social version of the schrodinger’s cat problem.

There are still going to be psychological limits on how many relationships we can maintain, whether we fundamentally change them or not. Perhaps technology enables us to have a feeling of maintaining more relationships, or it deepens relationships that have been weakened by our modern life styles. At a more human level, it is raising the question of what we mean by  ‘friend‘. It is certainly making qualitative changes to what we know about those around us, and our ability to discover new people to communicate with.

Maybe you’ve worked out your own personal Dunbar number, or found ways of recreating those different circles of friendship with social media? Do you see tribes and bands in action on-line?