The Dunbar Number And What It Means Practically

The science, more carefully

Dunbar's original research, published in 1992 and expanded in subsequent work, was based on neocortex ratio — the ratio of neocortex volume to the rest of the brain — across primate species. The relationship between this ratio and typical group size in those species is statistically robust. Extrapolating to humans using the typical human neocortex ratio gives approximately 147.8, rounded to 150.

The criticisms of this work are worth knowing. First: the extrapolation from non-human primates to humans assumes that the same cognitive mechanism that determines group size in baboons and chimps applies to humans. This is plausible but not proven. Second: group sizes in human societies are notoriously variable and context-dependent. Hunter-gatherer bands range from 20 to 500 and beyond in different environmental and cultural conditions. The 150 number appears consistently in some contexts but not others. Third: the emergence of language may have fundamentally altered the social cognition calculus for humans in ways the neocortex ratio model doesn't capture.

More recent empirical research — including studies of phone call patterns, Facebook network analysis, and Christmas card distribution — tends to support something in the 100-250 range as a stable social network size, but with substantial individual variation. The specific number 150 has been critiqued as suspiciously round and potentially artifact of the particular data sets Dunbar worked with.

None of this means the core insight is wrong. The insight — that there is a cognitive capacity limit on the number of genuine social relationships a human can maintain simultaneously — is well-supported even if the specific number is uncertain. Call it 100-200 if you want to be precise about the uncertainty.

The nested structure and its implications

Dunbar describes the clustering at different scales as reflecting different functional purposes:

The 5-layer (support clique) is defined by frequency of contact and emotional intimacy. These are the people whose emotional states affect yours directly, whose problems feel like your problems. Research on mortality and health shows that having a dense support clique of 3-5 people is the strongest social predictor of health and longevity — more powerful than exercise, diet, or access to healthcare in many studies. This is the layer that prevents the health impacts of loneliness.

The 15-layer (sympathy group) is defined by the mourning criterion — the people whose death would significantly disrupt your life. These are the people you feel genuine sympathy for rather than just concern. Research on social support in crisis conditions shows this is the network that provides meaningful help during genuine adversity — the people who actually show up.

The 50-layer corresponds roughly to what some anthropologists call the "overnight group" — the people you'd be comfortable camping with, sleeping in proximity to, with lower vigilance. This is the group within which resource sharing and genuine cooperation happens most naturally.

The 150-layer is what Dunbar calls the "tribe" — the unit within which reputational knowledge is complete. Everyone in this group knows everyone else well enough to have a genuine read on them. This is also roughly the unit within which informal social control works — gossip, reputation, the expectation of future interaction — rather than formal enforcement.

Above 150: Dunbar identifies layers at 500 (people you recognize and have met), 1500 (faces you'd recognize on the street), and 5000 (the outer limit of what Dunbar calls the "megaband"). These larger groupings have cultural and economic functions but involve much shallower relationships and require formal institutional structures to manage.

The key mechanism: cognitive overhead

The mechanism is worth understanding. Relationships require what Dunbar calls "social brain" processing — the ongoing modeling of another person's mental states, history, situation, and the state of your relationship with them. This is computationally expensive. You have to track: what does this person believe and value? What are they going through right now? What's the history between us? What do we owe each other? How will they interpret what I'm about to do? What are our mutual friends, obligations, shared contexts?

This cognitive load is continuous even when you're not actively interacting. You think about people in your social network. You plan interactions. You interpret events through the lens of how they'll affect relationships. You feel things about people who aren't present. All of this takes neocortical resources.

There are strategies for extending the effective limit slightly. Dunbar's own research shows that people who spend more time in face-to-face contact with their network, people who use humor more effectively (humor appears to release endorphins and functions as an efficient social bonding mechanism that reduces the cognitive overhead per interaction), and people with stronger general social skills can maintain somewhat larger networks. But these are adjustments at the margins — the fundamental limit is structural.

What this tells us about scale and institutions

The insight that communities above ~150 require formal structures rather than relational coordination has profound implications for institutional design.

Small organizations (below 150) function on trust, shared knowledge, and informal coordination. Everyone knows what everyone else is doing. Information moves through conversation. Problems get solved through relationship. Conflict gets managed through the parties working it out directly, often with help from mutual contacts. There's no HR department because everyone knows everyone and informal accountability works.

Large organizations (above 150) cannot function this way. The information overhead becomes unmanageable. People make decisions that affect others without knowing enough about those others to make good decisions. Coordination fails. Conflict doesn't get managed informally because the parties often don't share enough mutual contacts for informal accountability to work. Formal structures — hierarchy, written rules, specialized roles — become necessary.

The problem is that most organizations grow through this transition without consciously managing it. They start with the culture and norms of a small-team organization and add people until the organization is large without ever building the formal structures needed for large organizations. Then they wonder why things that used to work effortlessly now require enormous effort, why the culture is "not what it used to be," why decisions are slower and more contentious.

Gore-Tex's W.L. Gore Associates is the most famous example of a company that managed this deliberately. When factories approach 150 people, they build a new factory. This is expensive — redundant equipment, overhead costs, the inefficiencies of smaller production units. The founder, Bill Gore, concluded it was worth it because the difference in collaboration, communication, and work quality in 150-person units compared to 500-person units was large enough to more than offset the higher costs. The company has run this way for decades and remains consistently one of the highest-performing manufacturing organizations in its industries.

Application to intentional community and social organizing

For people building intentional communities — co-housing, ecovillages, worker cooperatives, religious communities, neighborhood associations — the Dunbar framework suggests specific design principles:

Keep the core community at 150 or below if you want relational rather than institutional coordination. If you need to grow beyond this, create sub-units (neighborhoods within the village, work teams within the cooperative, prayer circles within the congregation) that each function at the 15-50 scale for everyday life, with coordination mechanisms between them.

Invest heavily in the 5 and 15 layers. The most common loneliness pattern in large communities is having a large acquaintance network (150+) with depleted inner circles. People are surrounded by familiar faces and feel alone. Building structures that help people form and maintain close bonds within the community is more important than building structures that maximize the total number of relationships.

Expect and plan for sub-group formation. As communities grow, people will naturally sort into sub-groups. This is not a failure — it's the Dunbar mechanism doing exactly what it should. Fighting it is fighting human nature. Working with it means designing the community so that sub-groups have functional ways to relate to each other and the larger whole, rather than becoming competing factions.

Manage information as the network grows. In a 30-person community, everyone knows what's happening through conversation and shared life. In a 150-person community, this breaks down. Information has to be actively managed — not just communicated, but curated, distributed, and made accessible to people who weren't present. This is infrastructure work that small-community thinking tends to neglect.

The individual-level application

The Dunbar framework is also useful at the level of personal social strategy, which is different from community design but related.

Your 150 is not infinitely renewable. Every time you add a relationship, you either make room for it by letting another relationship drift toward the outer rings, or you're overcapacity and the quality of everything declines. This is not a moral failing — it's cognitive reality. Knowing this should make you more intentional about who you're maintaining close relationships with and why, and more compassionate toward yourself when long-held relationships drift despite good intentions.

Geographic mobility is one of the most potent forces for degrading the inner circles. When you move, the people in your 5 and 15 don't automatically relocate with you. The cognitive overhead of maintaining close relationships across distance is high. Without active investment, people drift from the 15 to the 50 to the 150 to the 500. You arrive in the new place with a large acquaintance network in your home community and no one in your inner circles in your new location. Building the inner circles back up takes years and requires real investment.

This is one of the underseen costs of mobility-optimized economies — the regular relocation that many careers require comes with a social cost that isn't usually accounted for, and that cost compounds over time. Understanding this doesn't make relocation wrong, but it should make the cost legible so that the investment in rebuilding social depth in the new location can be made deliberately.

The most practical insight: the inner circles are built through shared experience and mutual vulnerability more than through shared time alone. Expedite the rebuilding of inner circles in new places by doing things with people that create real stakes — shared projects, genuine vulnerability, navigating difficulty together — rather than just accumulating pleasant social encounters.

---

Related concepts: social brain hypothesis, neocortex ratio, community design, social capital, loneliness, inner circles, organizational scale, geographic mobility