Recent history has witnessed two important dramatic changes that have had a deep bearing on our social lives. One has been the way travel has shrunk the world to create a growing level of economic interdependence: butterflies flapping their wings in Brazil really do have reverberations on the economics and politics of every other continent in a way that has never previously been the case. The other has been the explosive rise of urban concentrations. In 1800, just 2% of the world’s population lived in cities, but by 1900 the figure was 13%, by 1950 49%, and by 2000 60%, with a current forecast of 80% by 2050 (UN 2009). These two trajectories have influenced our social world in ways that could not have been anticipated.

Let’s backtrack for a moment to the kinds of societies that we have lived in for most of our history as a species – and so at the very least for the last few hundred thousand years. The ancestral condition is what biologists usually refer to as a fission-fusion social system, the social system that characterises contemporary hunter-gatherers. These consist of a community that is normally fragmented into a number of separate foraging groups. These foraging groups typically number 30-50 in size and are relatively unstable, losing and gaining individuals and/or families over time. However, when they do gain members, they normally do so from other foraging groups within the same community. Communities typically average around 150, with a range in variation between 100-200. Unlike foraging groups, community membership is quite stable over time (aside, of course, from births and deaths). Communities themselves are clustered into higher order units, forming a series of hierarchically inclusive circles of sociality whose base is formed by clusters of best friends (typically around 5) and whose uppermost level (the “tribe” defined as all the people that speak the same dialect) numbers about 1500. These layers have been shown to scale with a consistent scaling ratio of about 3 (each layer is three times as large as the layer immediately inside it, yielding a series of groupings of size 5, 15, 50, 150, 500 and 1500) (Zhou et al. 2005, Hamilton et al. 2007).

The Neolithic Revolution that occurred around 10,000 years ago introduced agriculture and made possible a major sea-change in social style: the distributed networks of fission-fusion societies were able to converge on a single location and live together in settled villages. This set in train the increasing urbanisation of human societies that would take shape over the ensuing ten millennia. Though there is inevitably some variation in size, early settlements seem typically to have been of natural community size (i.e. about 150). And, indeed, aside from the drift to urban centres, rural villages seem to have remained about this size right through until modern times (Dunbar 2008).

In both dispersed hunter-gatherer and rural village societies, social exchange is common within the community layer, but much rarer between communities. The result is that most people share the same network of friends and relatives because they belong to the same community. The size of this social circle is around 150 individuals. We know more people than this (the number of individuals we can recognise and put names to is around 1500), but the number we can be said to have meaningful relationships with seems to be restricted to the 150 that form the natural community size of small scale societies.

Set against this long period of structural stability, the past half century has witnessed dramatic changes. These have been the direct result of the improvements in the speed and geographical scale of travel. This has enabled a degree of economic mobility unparalleled in the history of our species. People have, of course, always moved and no more so than during the nineteenth century when tens of millions of people drifted to the industrial centres of Europe to meet a voracious demand for labour or were part of the mass emigrations from Europe to the New World and Australasia. The big difference was that, in the past, people who moved lost their ties to their natal communities and created new small scale communities in their new homelands, whether these were the industrial urban centres of Europe or the plains of North America. Since the mid-twentieth century, however, fast cheap travel and efficient digital communication have meant that those who move no longer have to sever contacts with their home community. As a result, we move repeatedly, first to university, then every few years in the pursuit of jobs and promotion. Instead of consisting of 150 people who also share the same 150 contacts, our networks now consist of small subsets of friends that we accumulate with each successive move.  This might not in itself be a problem, but for the fact that these subsets of friends do not overlap and rarely have the opportunity to meet up and get to know each other. The result has been that our social networks have become fragmented and geographically dispersed. Where once the immediate neighbours that we live among were our friends and family, they are now strangers. In our transient modern lifestyle, we live as ‘Ruths among the alien corn’ with casual acquaintances and strangers.

One by-product of this is that our social networks are no longer as densely interconnected as they once were. Some might view this as an advantage: one of the commonest complaints about small communities is their cloying oppressiveness – everyone knows your business. But that is also precisely their strength. We have shown, for example, that more densely interconnected networks of friends are more likely to support each other and behave altruistically towards each other than those who have weakly interconnected networks (Curry & Dunbar 2011).

The other side of the coin, of course, is the fact that besides being more altruistic, members of densely interconnected networks are also more likely to be critical of each other’s behaviour – in order words, they act as informal policemen of the community’s social norms. It is perhaps inevitable that, in the kinds of weakly interconnected communities in which we now live, people are no longer so willing to intervene either in minor infringements of social mores or in the abuse and mistreatment of others. We are no longer prepared to protect the wider interests of the community. There are probably two separate, but mutually reinforcing, reasons for this. One is the simple fact that we no longer owe obligations to those we pass among. The second is, almost certainly, the fact that if we do intervene, it is quite unlikely that anyone else will support us. We risk exposing ourselves to concerted attack. And these days, thanks to a misplaced obsession with private rights on the part of the legal profession, we may even face the added risk of prosecution. It may be no surprise that as our world has become more urbanised, so there has been a parallel decline in our levels of social engagement with the wider community and a corresponding decline in social satisfaction (Putnam 2001, Wilkinson & Pickett 2009). We live in an increasingly disengaged, dissatisfied and self-obsessed world.

The ready availability of the online environment has merely served to exacerbate this problem. In a recent stratified national survey, no less than a third of respondents had witnessed or been involved in cyber-bullying, while 13% admitted that they had actively encouraged it (Dunbar 2012). There was a striking age difference in this respect. Nearly half (47%) of 18-24 year olds had observed online bullying, but only 16% of those aged 55 and over had done so. It is not clear whether this reflects a natural ageing process (we are more likely to speak before thinking when young, and only learn to tune our behaviour more sensitively as we age) or is a reflection of the changing pattern of social network structure (the older pre-Facebook generations rely on more direct face-to-face interaction to manage their relationships, and so have more closely structured networks).

There have, of course, been repeated claims that the advent of digital media has offered us the means of solving this problem. The most common claim is that the world of Facebook and other social networking sites (SNSs) allow us to maintain larger networks of friends. The online world has enabled us to break through the constraints imposed by Dunbar’s Number (the limit of 150 on the number of ‘friends’ we can have).  Unfortunately, this claim has been based on a misunderstanding of the issues at stake, partly thanks to Facebook’s rather loose use of the term ‘friend’ to cover all kinds of relationships.

The reality is that Facebook does not allow people to have more friends. In fact, the average number of friends on Facebook pages is around 150 (Marlow 2011, Facebook 2011). Moreover, those who use social digital media (SNSs, etc) more do not seem to have larger offline social networks (Pollet et al. 2011a). Rather, what we do when inviting large numbers of people to sign up as ‘friends’ on our Facebook page is to create the layer of acquaintances between 150 and 500 that we already have in natural offline social networks. This, of course, does not make these friends in any meaningful sense: their status with respect to us does not change by virtue of the fact that they are formally recorded as ‘friends’ on a social networking site. Just because we have a larger everyday offline social networks, it doesn’t mean to say that we are more social or sociable. Our research has shown that those who have large numbers of genuine friends (i.e. a larger than average face-to-face network size) typically sacrifice relationship quality to be able to do so (Pollet et al. 2011b). It is as though we have a limited amount of social capital and we can choose to spread that thinly (and have many weak friends) or thickly (and have a few strong friends) (see also Sutcliffe et al. 2012).

So, the question that we are left with is how to create a more engaged community despite the natural constraints that our psychology and the modern world impose on us. There are probably two options. One is to manipulate our psychology so as to improve our capacity to manage more individuals. However, if the constraint on the number of relationships we can manage is a hardware constraint and there is a monotonic relationship between network size and brain (or even frontal lobe) size as implied by the findings of Lewis et al. (2011) and Powell et al. (2012), then a significant increase in network size is not likely to be possible. The problem is that we already face a severe constraint on brain size at birth: the birthing process is already difficult for us because the human neonate’s head only just fits through the pelvic birth canal. In any case, we have tried to resolve this problem already by giving birth to premature babies which complete brain growth outside the womb.  A further reduction in the length of gestation is not a serious option since our babies are already born as close to the margin of survival as we can manage. In any case, even if we could compensate for an increase in natal brain size, this could only be achieved by a proportional increase in women’s hip size (with knock-on consequences for their ability to walk). A more feasible possibility might be to extend the period of growth outside the womb. However, all of these solutions require genetic changes, and selection is too slow to solve the problem within the time scale we need it to.

The alternative is to find ways to exploit the bonding mechanisms that we use to maintain small scale societies and personal social networks and extend them to super-large networks and communities. In fact, we have already tried this during the course of recent human history. Two obvious examples are the military and world religions. However, both resort to the same strategy of using hierarchical structures and strict discipline to ensure cohesion on the very large scale (Dunbar 2011). Even so, the military option works only in very specific circumstances (battlefields where men’s lives are at risk) and really works only when draconian discipline (and/or a great deal of repetitive training in the form of drill) is imposed. Without that, humans seem especially prone to undermining any forms of discipline imposed from above. This is no doubt why the world religions are so bedevilled by constant fragmentation into sects and cults. In any case, neither of these options seems especially attractive. Social cohesion on the very large scale will always be more effective when the commitment comes (bottom up) from the individual. To some extent, religions do this by creating a sense of a “grand project” that attracts the commitment of the members, but they do so at the expense of having to impose a theology and this invariably creates an “us versus them” mindset – something that is not especially desirable if the aim is wider social integration. The problem, then, is how to conceive of a grand enough project to induce that sense of commitment. If we could figure that one out, we might have a serious chance of engineering planet-wide cooperation and goodwill.

R.I.M. Dunbar
Department of Experimental Psychology
University of Oxford
Oxford OX1 3UD
[email: robin.dunbar@psy.ox.ac.uk]


Curry, O. & Dunbar, R.I.M. (2011). Altruism in networks: the effect of connections. Biology Letters 7: 651-653.

Dunbar, R.I.M. (2008). Mind the gap: or why humans aren’t just great apes. Proceedings of the British Academy 154: 403-423.

Dunbar, R.I.M. (2011). Constraints on the evolution of social institutions and their implications for information flow. Journal of Institutional Economics 7: 345-371.

Dunbar, R.I.M. (2012). Speak Up, Speak Out. Report for the Holocaust Memorial Day Trust, London.

Hamilton, M.J., Milne, B.T., Walker, R.S., Burger, O. &Brown, J.H.(2007). The complex structure of hunter-gatherer social networks. Proceedings of the Royal Society, London, 274B: 2195-2202.

Facebook (2011).http://www.facebook.com/press/info.php?statistics

Lewis, P.A., Rezaie, R., Browne, R., Roberts, N. & Dunbar, R.I.M. (2011). Ventromedial prefrontal volume predicts understanding of others and social network size. NeuroImage 57: 1624-1629.

Marlow, C. (2011). Maintained relationships on Facebook.http://overstated.net/

Pollet, T.V., Roberts, S.G.B. & Dunbar, R.I.M. (2011a).  Use of social network sites and instant messaging does not lead to increased offline social network size, or to emotionally closer relationships with offline network members. Cyberpsychology, Behavior, and Social Networking 14, 253-8.

Pollet, T. V., Roberts, S.G.B.& Dunbar, R.I.M. (2011b). Extraverts have larger social network layers but do not feel emotionally closer to individuals at any layer. Journal of Individual Differences 32: 161-169.

Powell, J., Lewis, P., Roberts, N., García-Fiñana, M. & Dunbar, R.I.M. (2012). Orbital prefrontal cortex volume predicts social network size: an imaging study of individual differences in humans. Proc. R. Soc. Lond. (in press).

Putman, R.I.M.(2001). Bowling Alone: The Collapse and Revival of American Community. New York: Simon & Schuster.

Sutcliffe, A.J., Dunbar, R.I.M., Binder, J. & Arrow, H. (2012). Relationships and the social brain: integrating psychological and evolutionary perspectives. British Journal of Psychology (in press).

UN (2009).World Urbanization Prospects: the 2009 Revision. New York: Department of Economic and Social Affairs,United Nations

Wilkinson, R.& Pickett, K.(2009). The Spirit Level: Why More Equal Societies Almost Always Do Better.Harmondsworth: Allen Lane.

Zhou, W-X., Sornette, D., Hill, R.A. & Dunbar, R.I.M. (2005). Discrete hierarchical organization of social group sizes.  Proceedings of the Royal Society, London, 272B: 439-444.

Published On: May 7, 2012

Notice: Trying to get property of non-object in /home/bn8m2v9io5fc/public_html/wp-content/themes/salient-child/single.php on line 293

Notice: Trying to get property of non-object in /home/bn8m2v9io5fc/public_html/wp-content/themes/salient-child/single.php on line 294

Notice: Trying to get property of non-object in /home/bn8m2v9io5fc/public_html/wp-content/themes/salient-child/single.php on line 305

Notice: Trying to get property of non-object in /home/bn8m2v9io5fc/public_html/wp-content/themes/salient-child/single.php on line 306

Notice: Trying to get property of non-object in /home/bn8m2v9io5fc/public_html/wp-content/themes/salient-child/single.php on line 310


  • Peter Turchin says:

    A very interesting argument. However, I feel that the conclusion is overly pessimistic. Humans can be integrated not only by face-to-face interactions or by hierarchical chains of command. We also have an evolved ability to relate to abstract groups of unlimited size. It’s as though we have circuits in our brains that make us believe that imaginary communities are more real than they are – and by believing it, we do make them real. Furthermore, the social scale of such imaginary communities has been increasing over the last hundred thousand years.

    Sometime in prehistory we evolved the ability to demarcate groups using symbolic markers. Initially, such groups were limited to ethnolinguistic communities – people who spoke like us, looked like us, worshipped the same tribal spirits. This evolutionary breakthrough allowed humans to cooperate in groups larger than c.150 (the Dunbar number, groups integrated by face-to-face interactions). The next breakthrough came during the Axial Age (800-200 BCE) with the rise of world religions that integrated multiethnic communities in the new large-scale societies (traditional empires). During the last couple of hundred years we have been living through the last major transition, as humans are learning to cooperate at the level of the whole humanity. This transition is not complete, but things are moving in the right direction. Until now it has always been ‘us’ versus ‘them’ but perhaps we will be able to transcend this logic. At least, I am optimistic that we will.

  • victor demunck says:

    as cliche-ish as it may sound, the only answer i can come up with, and have thought about, is generalized love, which is sort of the opposite of romantic love, yet stems from the same evolutionary drive. it would require,i’m just conjecturing, a re-framing of our educational system in which we are seen both as statistical tokens (if we are thinking X then others are likely thinking X too and if we do X then others see their thoughts as validated and will do X as well) and as having a joint commitment (as in love relations such as pa-child or lovers) to a plural subject (e.g., humanity) based on a few moral principles enforced by implicit coercive pressures that are heterarchical and acephalic.
    It might be done. A lot would have to go into it and there would have to be ways to control free riders (the coercive pressure of the masses, mostly implicit, but real when threatened). The problem with large size is, as dunbar and putnam note is anonymity, but even in the moscow subway system there is an underlying order in human movements and the masses exert an implicit coercive pressure, as well as in other public arenas where heterogeneous anonymous individuals for groups without there being any obvious controlling agents.

    People are really good at reifying and rather than it being “it” being a problem in human reasoning perhaps it can be turned into an asset?

  • roger says:

    As I read this article I felt myself tipping from a Pre to Post Dunbar world.

    All around us there people who are trying to make sense of the world based on how it has been and have set their expectations at level that is utterly unrealistic. They do not realise that all the expectations and preconceptions they have about just about everything no longer apply.

    Many no longer feel any responsibility for others. They are certain they are the most important person in the world – they demand respect and must be understood and acknowledged. This doesn’t happen and so they get angry. They are trapped in a vicious circle and don’t know it. They have no access to the tools they need to escape. The idea of understanding someone else first in order to be understood is alien, as is paying respect in order to be respected.

    As the world changes it seems likely more paradoxes will occur, and this may create an additional burden. Many people think that the psychologist Valeria Ugazio has made the first genuine advance in the field for some time. In her latest work “Semantic Polarities” she has identified our inability to cope with paradox as being a major cause of illness. Put very crudely, we are talking about people who cannot choose between altruism and selfishness, dependence and independence, winning and losing, belonging and not belonging.

    Now may be the only time in history that we can make a confident prediction about the future – it will be completely different from the past. Understanding why may help free those currently trapped and re-engage with society, and this a compelling explanation.

  • Allan Hayes says:

    This is rather late as a response to something written in 2012, but dealing with the problem is becoming more important, so here goes:
    I live in one of the UK’s most diverse cities, Leicester. I am deeply involved with trying to bring people from different backgrounds together. I used to speak as a Humanist, but latterly I have noticed an increasing tendency for people start with ‘first I am a human being, THEN I am …”. I think that people want to be part of a story, and I have taken to trying to simply communicate some sense of what humanity is, evolution and religions included, and what we might help it to become.
    We have mandatory religious education in UK schools – I want to change this to education in Humanity.
    More widely, I am trying to encourage open dialogue. This year I have led a team organising the first TEDx event in Leicester,(www.tedxleicester.com). I hope to build on this for 2016 – maybe an event with the theme “One Humanity”.

  • Helga Vieirch says:

    I have wanted to write a comment here for a long time, but i had to go over my data again before I did, and even then, I hoped I had misunderstood something and thought about it for a long time. I cannot see any way around this anymore. I have to let you know that when you talk about hunter-gather bands and say “The result is that most people share the same network of friends and relatives because they belong to the same community. The size of this social circle is around 150 individuals. We know more people than this (the number of individuals we can recognise and put names to is around 1500), but the number we can be said to have meaningful relationships with seems to be restricted to the 150 that form the natural community size of small scale societies.” You are misrepresenting the reality. I have done intensive fieldwork among African hunter-gatherers, and also with small scale subsistence farmers and pastoralists in West Africa. Your figure of about 150 closet networked family and friends is pretty accurate, but almost all individuals have overlapping BUT NOT IDENTICAL networks. This means that if you look at the people well known and contacted over the course of a year, even a husband and wife will only share perhaps a third of their contacts – who are reported as close relatives and friends of the family. They will still have many people – childhood friends, various cousins and other relatives, and even people they got to know through these contacts, which they will not share with their spouse. Now among hunter-gatherers, when you consider the way camping parties (that unit often called a band, consisting of three to five households) are made up, they are, as you say, fluid in membership over time. But since close relatives generally do not marry, husband and wife often have different childhood friends, different circles of siblings and cousins, and each person in the other’s network might be living in one of a hundred other camping parties at any one time. In this way the effective network represented by any ONE household affords contacts with at least one person in bands throughout the entire landscape occupied by that language group (which can be up to several thousand people). Furthermore, of course, each household in that band will have a similar network, and each band in the whole likewise, so effectively everyone knows someone who knows someone who knows someone … and they all talk and gossip all the time, and there is lots of visiting between these camp sites, so everyone will at least have heard – even if only at third or forth hand, all about the life and foibles of just about everyone else.

    Please try to keep the size of cooperative residential groups separate from the size of individual networks. I could go on about the way this worked to create multiple connections between villages in West Africa, but I really do think that the idea that there was some great “sea-change” in social style… bears much scrutiny. Villages do not much alter the size of individual networks, nor interfere much with the range of contacts outside the local sphere. People maintain contacts with friends and family over many different villages. A woman might be born in village A, married and live in village B, while her own mother and mother’s relatives will live in village C, and her sister might live in village D. Meanwhile her best friend, a distant cousin form her natal village, will have married a guy in village E. So over the course of a year, she will manage to visit all these people in all these locations, or they will at some point visit her in her home village. In west Africa, there are weekly markets run by the women, and these rotate among all the villages in a region, permitting friends to get together at regular intervals. And of course, in case it was not obvious, each of the women in the above network also had an extensive network of her own which was not identical to that of the original woman of my example, born in village A. Men are even more mobile as they go to find work in larger towns and make friends there who might be from another part of the country entirely, but who often keep in touch. All of these people use their networks to find out wharf is happening in the families and lineages and villages of everyone else, and to make arrangements for their children’s marriages, for finding traditional healers and midwives, and for locating missing goats or pigs. When a man was hit with an axe during a brawl three villages away, I got to hear of it via a chain of messages sent from household to household between villages within a few hours, and managed to drive out to collect the man and take him to the nearest hospital, all with the axe still stuck in his head. Today they would no doubt just use cell phones, but they would be going through the same grapevine!
    Facebook is just the same pattern by another medium.
    It is what humans do.

Leave a Reply to roger Cancel Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.