What Dunbar's number actually says

At some point you absorbed the fact that humans can only maintain about 150 relationships. It gets quoted in management books, on podcasts, in arguments about why your group chat stopped working. It is one of those numbers that feels like it settles something. It is worth knowing where it came from, because the original claim is both narrower and much less certain than the version that travels.

Where the number actually came from

In the early 1990s the anthropologist Robin Dunbar noticed a relationship in primate data: across non-human primate species, the ratio of neocortex to the rest of the brain correlated with the typical size of the social group that species lived in. Bigger relative neocortex, bigger group. The suggested explanation was that keeping track of who is allied with whom, who owes whom, and who is currently annoyed with whom is cognitively expensive, and that brains and group sizes had grown together under that pressure.

The famous number is what you get when you take that correlation, plug in the human neocortex ratio, and read off the answer. It is an extrapolation from other species to ours. Dunbar then cross-checked the result against whatever human group sizes he could find — village sizes in small-scale societies, historical military unit sizes, the size of certain communities in the historical record — and found figures in a broadly similar range.

Three things are worth holding onto here, and all three get lost in retelling. First, the underlying data are about non-human primates; humans were the thing being predicted, not the thing being measured. Second, the human cross-checks are group sizes — villages, congregations, units — not counts of anybody's friendships. Third, and most importantly, the calculation did not produce 150. It produced a range, and 150 is roughly its midpoint. The range was wide enough that the tidy number is a convenience, not a constant.

The serious objections

This has been argued about properly, not just on the internet. In 2021 a team at Stockholm University led by Patrik Lindenfors reanalysed the primate data with updated methods and concluded that the confidence intervals around any human estimate are so wide that the exercise cannot support a usable number at all. Depending on the modelling choices, the plausible range ran from a couple of dozen people to several hundred. Their argument was not that humans have unlimited social capacity; it was that this particular method cannot tell you where the limit sits.

A second objection is about what the human evidence measures. Ethnographic village sizes are averages of settlements, and settlement size is set by farming, water, defence, and how far people are willing to walk, at least as much as by anybody's cognitive ceiling. A number derived from how many people live within a day of each other is not a measurement of how many relationships a single person can hold.

A third objection is that human sociality is not one thing. A person raised in a dense extended family, a person with a hundred colleagues, and a person who moves cities every three years are running very different social lives. Treating a species-level average as a personal budget is the kind of move that sounds rigorous and is not.

The layers get quoted even more loosely

You will also have seen the tidy nested version: roughly five intimates, fifteen close friends, fifty friends, 150 acquaintances, each ring about three times the last. This comes from the same line of research and inherits all of its uncertainty, plus some of its own. The layers are estimates fitted to patterns in social network data, not measurements taken from individuals, and they have been contested on the same grounds as the headline number.

What the layered picture does capture, and this part is much more solidly supported by ordinary observation as well as by network research generally, is shape rather than size. Almost everyone's social life is steeply uneven: a very small number of people receive most of your attention and disclosure, a somewhat larger group gets a moderate amount, and a long tail gets very little. That pattern shows up more or less wherever anyone looks. The specific counts attached to it are much softer than they sound.

What survives all the objections

Strip away the number and something genuinely useful remains, which is why the idea refuses to die.

Attention is finite and it is the actual currency. Whatever the ceiling is, there is one, and it is not set by how much you care. Staying close to someone costs conversations, memory, and time in the same room. Those are rivalrous resources: an hour spent with one person is not available for another. Nothing about modern life has changed that, even though modern life has enormously increased the number of people you can technically be in contact with.

Contact is not the same as relationship. The reason a follower count or a contact list feels hollow is that it counts the cheap thing. Being reachable by a thousand people costs nothing. Being known by twelve costs a great deal, and the second one is what people actually mean when they say they want friends.

The unevenness is not a defect. If your capacity is limited, then being much closer to a few people than to most is not a moral failure — it is the only shape a finite thing can take. The alternative, spreading yourself evenly across everyone you have ever liked, produces uniformly thin contact and nothing else.

How to use it without misusing it

The right way to hold Dunbar's number is as a metaphor with a decent pedigree, not as a specification. Use it the way you would use the observation that you cannot read every book: true, useful, orienting, and completely useless as an exact figure.

Concretely, that means it is a reasonable answer to the question "why do I feel like I am falling behind on everyone?" — because you probably are, structurally, and no amount of organisation will fix a capacity problem. It is not a reasonable basis for pruning your contacts down to a target, deciding a team of 160 is doomed, or telling yourself that a friendship should be dropped because you are over quota. Nobody has a quota. The research never established one, and the researchers who reanalysed it said so plainly.

What you actually have is a limited amount of attention and a lot of people worth spending it on. The number was never going to solve that. Deciding, honestly and without guilt, who you want to be genuinely close to right now, and letting everyone else be a warm acquaintance rather than a neglected friend, gets much closer to the real problem than any figure could.