Why Every Group You Join Generates Its Own Politics

The Software You Can’t Uninstall: Why Every Group You Join Immediately Starts Running Ancient Political Machinery

Picture this: it’s a Tuesday evening, you’re on your phone, and you’ve just sent a message in a group chat for a hobby you genuinely enjoy — maybe a book club, a gaming server, a neighborhood birding group, whatever it is. Someone else responds to a different message in the thread and gets four reactions. Your message gets none. It was a good message. Maybe even a better message.

And now you feel something.

It’s subtle. A faint, deflated quality to the next few minutes, a barely-conscious scan of why theirs landed and yours didn’t, a recalibration of how you’ll pitch your next contribution. You might notice it, you might not. But something in you has registered a social signal and responded to it, completely without your permission. Here’s what’s actually happening in your body during those few minutes: brain imaging research has shown that social exclusion and undervaluation activate the dorsal anterior cingulate cortex and the anterior insula, the same regions that fire during physical pain, including regions associated with its affective components. Not “similar regions.” Overlapping regions. Your brain is processing being ignored in a group chat using some of the same hardware it uses to process a broken bone.

This is not a metaphor. This is the mechanism. And it is the beginning of the answer to a question most people think they already understand: why does every human group, no matter how small, no matter how low-stakes, eventually generate politics?

The Software Was Written for a Different Machine

The answer starts about 300,000 years ago, give or take, in small bands of humans living in conditions where the social group was not an optional feature of life — it was the life-support system. Being genuinely excluded from a group of 50 people who shared resources, protected each other from predators, and cooperated in food acquisition wasn’t a social inconvenience. It was a death sentence. Being low-ranked in that group wasn’t embarrassing. It meant less food, less protection, less access to mates, less of everything that kept you alive and reproductively viable.

Under those conditions, your brain needed to track your social standing the way it tracked hunger or pain — automatically, continuously, with an urgency proportional to the actual stakes. And so it evolved exactly that system.

Psychologist Mark Leary’s sociometer theory is the cleanest description of how this works. The self-esteem system, Leary proposes, didn’t evolve to make you feel good about yourself. It evolved as a monitoring instrument, a gauge that reads your current standing in the social environment and alerts you when that standing is declining. People aren’t motivated to maintain their self-esteem for its own sake; they’re motivated to maintain their relational value and social acceptance, and self-esteem is just the needle on that instrument. When the needle drops, the alarm fires. When the alarm fires, you are motivated, below conscious deliberation, before you’ve had time to reason about whether it’s appropriate, to do something that will raise your standing in others’ eyes.

The system cannot distinguish between a tribe of 50 people on whom your survival depends and a Discord server of 12 people arguing about a hobby. It processes both as “group” and reads status cues from both. The pain calibration was set for the first context. You are now experiencing it in the second.

The Hierarchy Forms Before Anyone Opens Their Mouth

One of the most striking findings in the research on small groups is how fast this machinery gets to work. We tend to imagine that hierarchies develop over time through demonstrated competence, through accumulated history, through the slow accumulation of interpersonal dynamics. The research says otherwise.

Studies on small group dynamics have found that informal hierarchical differentiation develops spontaneously and rapidly, individuals form inferences about others’ competence and power based on mere seconds of observation, and differences in participation that emerge within the first few minutes of interaction can produce hierarchical differentiation that shapes the entire subsequent group experience.

Sociologist Robert F. Bales convened small decision-making groups of undergraduate students — strangers, no formal titles, no assigned roles and found that a status hierarchy had developed within a single meeting. And critically, group members showed high agreement about each individual’s rank, suggesting this isn’t just individual impression formation, it’s a shared social reality that the whole group is constructing simultaneously, from the same subtle cues.

Those cues include things no one is consciously tracking: speaking time, turn-taking patterns, whose suggestions get picked up and developed versus quietly dropped, microexpressions, posture, the direction other people’s eyes move when a question is asked. The hierarchy is being assembled from this data in real time, by every brain in the room, before anyone has formed a deliberate opinion about anyone else.

And once it forms, it sticks. Speaking-time rank orders from the end of a first group session were substantially predictive of rank orders at the beginning of a second session a week later, in groups that had barely spent time together. The brain doesn’t just notice the hierarchy, it encodes it, and then uses that encoding to filter and interpret everything that comes after.

The reason this happens so fast is the same reason the sociometer fires so reliably: in ancestral environments, not knowing your rank or not knowing the rank of a stranger you’d just encountered was a genuinely dangerous information gap. A person whose position in the hierarchy you hadn’t assessed was either a potential threat or a potential ally, and getting that calculation wrong could cost you. Speed and accuracy in reading the social landscape were adaptive. The brain that could rank everyone in the room within minutes had a survival edge over the brain that needed months of careful observation. We are the descendants of the fast rankers.

Two Ways to the Top, Both Running Simultaneously

The speed of hierarchy formation is striking, but equally interesting is the shape of it. Research by Joseph Henrich and Francisco Gil-White identified two entirely distinct routes to high status in human groups, both of which are active in informal settings.

The first is dominance — the older route, shared with most social species. The dominant individual achieves status through intimidation, coercion, and displays of force. It’s associated with what researchers call “hubristic pride” and a psychology of fear: people defer not because they want to be near this person but because they’re calculating the costs of not deferring. The dominant individual gets compliance.

The second is prestige — and this one, Henrich and Gil-White argued, may be unique to humans. The prestigious individual achieves status through demonstrated skill or knowledge that others want access to. It’s associated with authentic pride and a psychology of attraction: people seek out the prestigious individual voluntarily, pay close attention to them, and copy their behavior, because proximity to expertise is genuinely useful. The prestigious individual gets influence.

Both of these routes activate in informal groups with no material stakes. The person who becomes the informal authority in a hobby group, who gets deferred to on questions about gear, technique, or interpretation is running the prestige route. The person who controls the tone of the conversation, who sets the terms of what’s acceptable to say and how challenges are met, may be running something closer to the dominance route. Most informal groups have both operating at once, often in different people, which creates its own structural tension.

What this means practically is that the “power struggle” in a low-stakes group isn’t a single thing. It’s at least two overlapping contests with different rules and people are participating in both simultaneously, often without knowing which game they’re playing.

The Hormone Layer Nobody Talks About

Beneath the psychology is a layer of endocrinology that popular coverage almost never reaches, and it’s where the mechanism gets genuinely strange.

Research on the dual hormone hypothesis has found that individuals with both high testosterone and low cortisol may be especially likely to occupy high-status positions in social hierarchies, while individuals with high testosterone and high cortisol may not. Researchers examined testosterone and cortisol levels in actual executives alongside a concrete measure of attained status: the number of subordinates over whom each executive had authority. Despite the enormous number of nonhormonal factors that influence career outcomes, the hormonal pattern held. In separate studies, one on leadership in a mixed-sex sample, one on competition, testosterone was positively related to dominance, but only in individuals with low cortisol. In individuals with high cortisol, the relationship between testosterone and dominance was blocked or reversed.

What this suggests is counterintuitive: testosterone isn’t simply an aggression hormone that drives people to compete. It functions more like a status-position tracker — a system that amplifies whatever the current social context says you should be doing. Longitudinal data shows that those with higher baseline testosterone display more dominance as their status increases, the hormone responds to the hierarchy as much as it drives it. It’s a feedback loop, not a one-way push.

Cortisol tells the other half of the story. It rises during social threat and is dampened when individuals gain social control. This is the machinery behind that group-chat moment from the opening. The faint deflation, the recalibration, the heightened attention to why your message didn’t land: that’s not oversensitivity or vanity. That’s a stress-response system that evolved to register social threat as seriously as physical threat, doing exactly what it was built to do. The brain isn’t overreacting. It’s running a program that was written for a context in which being undervalued in your group genuinely could mean the difference between survival and death. It just has no way to know that you’re in a birding Discord and not a Pleistocene foraging band.

The practical implication of the hormonal data is uncomfortable: the person who informally rises in a group may be partly determined by endocrinology, not deliberate strategy. The group isn’t electing its informal leader through a rational assessment of everyone’srelative merits. It’s responding to a set of physiological signals that are shaping behavior below anyone’s conscious awareness, including the behavior of the person being elevated.

Why Nothing Real Has to Be at Stake

This is the part that most explanations of group dynamics skip past, and it’s arguably the most important piece: why do these dynamics play out with exactly the same intensity in groups where nothing material is actually contested?

The answer lies in a cognitive bias called zero-sum perception. People often hold zero-sum beliefs, subjective beliefs that, independent of the actual distribution of resources, one party’s gains are inevitably accrued at the other party’s expense, about interpersonal, intergroup, and international relations. The crucial qualifier is “independent of the actual distribution.” The belief that status is a fixed pie doesn’t require status to actually be fixed. It fires regardless.

The evolutionary logic is straightforward: a zero-sum heuristic evolved as a cognitive adaptation to enable successful intra-group competition for genuinely limited ancestral resources. Food, mates, shelter, protection — these really were fixed-sum in the environments that shaped our psychology. If someone else ate that meat, you didn’t eat it. If someone else had the alliance with the most capable hunter, you didn’t have it. The zero-sum calculation was often accurate, and treating non-zero-sum situations as zero-sum had lower costs than missing a real competition.

Now the heuristic fires in contexts where status and attention are not genuinely fixed-sum but feel exactly as if they are. When someone else gets praised in the group chat, the ancestral accounting system registers a loss even though no actual resource has moved. When someone else’s suggestion gets adopted by the team, the status ledger updates as if a territory has changed hands. The cognitive machinery cannot tell the difference between a situation where the zero-sum perception is accurate and one where it isn’t.

Three psychological channels reliably trigger zero-sum beliefs: forces that elicit threat, generate real or imagined resource scarcity, and inhibit deliberation. Informal group dynamics have a remarkable tendency to activate all three at once. The social threat is real, the sociometer is firing. The resource — attention, deference, influence feels scarce even when it isn’t. And the inhibition of deliberation is built into the speed of the process: hierarchies form in seconds, before reflective reasoning has a chance to intervene.

The Storming Phase Was Never a Bug

Bruce Tuckman’s 1965 model of group development — forming, storming, norming, performing is probably the most widely cited framework in team management, and it is almost universally misunderstood. In most summaries, “storming” is treated as a dysfunction that good teams should minimize and bad teams get stuck in. That framing has it exactly backwards.

Storming is the status-resolution system doing exactly what it evolved to do. Tuckman described it as a stage where group members become hostile toward one another and toward the task as a means of expressing individuality and resisting the formation of group structure. That framing, “resisting the formation of group structure” sounds like obstruction. The more accurate description is that multiple individuals are simultaneously running status-acquisition programs, their bids are colliding, and the conflict is the resolution mechanism, not the problem. The hierarchy has to be contested before it can be settled.

What drives this is the same thing that makes hierarchies form so fast: rank ambiguity is aversive. The status-processing system is designed to close information gaps about social position, and an unresolved hierarchy is exactly that kind of gap. The discomfort of not knowing where you stand relative to everyone else in a new group is not a personality quirk of anxious or competitive people, it is a broadly shared feature of human social cognition. The storming phase is what happens when everyone in a group is simultaneously trying to resolve that ambiguity, and their respective resolutions are incompatible.

What most summaries of Tuckman’s model also omit is the regression problem. Teams don’t move through these stages once in a linear progression. They regress. A team that has reached norming or performing can be thrown back into storming by a single new member joining, or by a significant change in context or task. Every new member resets the hierarchy’s unresolved ambiguity, their rank relative to everyone else’s is now an open question, and the status-resolution machinery fires again. This is why adding one person to a settled team can destabilize dynamics that seemed completely solid. It’s not that the new person is disruptive. It’s that their arrival is an input the hierarchy-processing system has to incorporate, and incorporating it requires re-running the whole contest at a lower intensity.

The Ancient Weapon: Why Gossip Is Politics

So far this has been a story about hierarchy formation — how groups rank themselves automatically, quickly, and below conscious awareness. But there’s a parallel story about hierarchy resistance that is equally old, equally automatic, and equally invisible to participants who think they’re just venting.

Anthropologist Christopher Boehm’s research on hunter-gatherer societies documented what he called a “reverse dominance hierarchy” — a coalition of subordinates that actively and consciously suppressed would-be dominants. In mobile foraging societies, egalitarianism was not the natural resting state of the group. It was an actively enforced political achievement, maintained through a specific toolkit: mockery, ridicule, disobedience, ostracism, shunning, and in extreme cases, exile or execution. The group collectively and deliberately prevented any individual from accumulating the kind of power that would allow them to dominate the others. Remove the enforcement, and dominance hierarchies re-emerged.

The modeling research on how this works at a mechanistic level is striking. Simulations of group dynamics have found that a dominance hierarchy of individuals with exclusively self-centered characteristics — the desire to dominate, and resentment at being dominated transitions spontaneously to egalitarianism as their capacity for language develops. Language, specifically gossip, allows resentment against being dominated to promote the formation of anti-dominance coalitions that destabilize the alpha position. The gossip is the coalition. It doesn’t just describe the problem; it assembles the response.

This is directly continuous with what Frans de Waal observed in his long-term study of chimpanzee politics at Arnhem Zoo, documented in Chimpanzee Politics, where the alpha male position was rarely held by the physically strongest animal. The alpha was the animal best at managing alliances: forming coalitions with key supporters, performing reconciliation after conflicts, strategically distributing grooming and support. When an alpha’s coalition fragmented, he fell, regardless of individual strength. Humans and chimpanzees share roughly 98 to 99 percent of their DNA, and the coalition logic appears to be part of that shared inheritance.

What this means for modern informal groups is direct and slightly vertiginous. The person talking behind the domineering moderator’s back in a side channel, the employees coordinating quietly in the office kitchen about the manager who takes credit for their work, the book club members exchanging raised eyebrows over text — they are not merely gossiping. They are running an anti-dominance protocol that is hundreds of thousands of years old, field-tested across countless generations of primate social life, and remarkably effective. The gossip is the coalition. The side channel is the politics. And the target of that coalition, if they’re socially perceptive, will feel it — not as an abstraction but as the cortisol-mediated stress response of someone whose alliance network is visibly fraying.

The 1954 Robbers Cave experiment demonstrated how rapidly tribalism and social structures can assemble, though not quite as seamlessly from “nothing” as originally reported. Social psychologist Muzafer Sherif and his colleagues observed two groups of eleven-year-old boys at an Oklahoma summer camp without their knowledge. Within days, each group of strangers spontaneously produced a sharp internal hierarchy, complete with clear leaders, group norms, and an in-group name. Lower-status boys quickly internalized their rank and self-policed to maintain it. However, the subsequent explosive hostility toward the out-group was not purely organic. While the boys naturalized their tribal boundaries, adult researchers heavily stage-managed the friction from behind the scenes. Staff deliberately engineered zero-sum competitions, sabotaged camp equipment, and destroyed the boys’ belongings to artificially provoke and accelerate the hostility.

While the Robbers Cave findings are often cited as a clean demonstration that hierarchy is what groups do by default, the reality is more complicated. The rapid emergence of internal group structures among the boys does show that humans naturally organize themselves when given a shared context and proximity. However, the study also proves that external design matters immensely. The explosive out-group hostility did not simply assemble itself from nothing; it required deliberate manipulation and engineered zero-sum incentives by adult coordinators. True hierarchy and conflict, then, are not purely spontaneous defaults, they are highly responsive to how the surrounding environment and resources are structured.

The Consent Problem

Here is where most articles on this topic end with a list of coping strategies — how to navigate office politics, how to manage status anxiety, how to build alliances wisely. That framing assumes you are outside the system, looking in, making choices about how to engage with it. The research suggests a different picture.

Consider the phenomenon researchers call countersignaling. The logic runs like this: the lowest-quality individuals in a status hierarchy cannot afford to signal their status loudly, because the signal won’t be believed. Mid-tier individuals must signal conspicuously — the expensive car, the name-dropping, the visible effort to demonstrate credentials because conspicuous signaling is their primary tool for establishing credibility. But the highest-status individuals can afford to do something different: they can minimize display effort entirely, communicating through their relaxed indifference to status markers that their position is so secure it requires no defense. This non-signal is read by informed receivers as the highest-status indicator of all. Trying hard, in this framework, is itself a signal of middling status. Not trying or performing not-trying is the apex move.

The unsettling implication for anyone reading an article about status dynamics and feeling pleasantly immune to them is this: that feeling of immunity is itself a status performance. Claiming not to care about status is one of the most reliable high-status moves available. The person who says “I’ve never really been competitive, status stuff just doesn’t interest me” is, in that very sentence, countersignaling and broadcasting that their position is secure enough that they don’t need to compete visibly. The system has absorbed even the gesture of opting out and converted it into a move within the game.

And the research on status signals in cooperative contexts adds another layer. Studies have found that overt status symbols can actually repel potential friends — people intuit that conspicuous status displays are socially unattractive in cooperative settings, and they refrain from overt signaling when they expect to collaborate with others. But this doesn’t mean the status competition stops. It means it goes underground. Who gets the last word. Whose suggestions get adopted. Who sets the conversational tone. Who gets talked over, and who never does. The competition becomes less visible precisely because it’s more sophisticated, it migrates into channels that are harder to name and therefore harder to resist.

Sociometer theory’s deepest implication is that this system runs whether you endorse it or not. The sociometer is a measurer of social phenomena designed to detect detrimental or negative changes in an individual’s group status. It doesn’t ask for your endorsement before it fires. It doesn’t check whether you’ve decided, on reflection, that status in this particular group doesn’t matter to you. It reads the cues, updates the gauge, and generates the motivational response. The hardware predates your opinions about it by several hundred thousand years.

The Robbers Cave boys didn’t decide to form a hierarchy. Bales’s undergraduates didn’t vote to establish a pecking order within their first hour together. The chimpanzees at Arnhem Zoo weren’t pursuing a conscious political philosophy. And you, in that moment with the group chat and the unreacted message, weren’t choosing to care. The status-processing system made that choice on your behalf, using criteria it inherited from an environment that no longer exists, calibrated to stakes that no longer apply, running on hardware that has no patch and no off switch.

This is not a counsel of despair. Knowing the machinery exists is genuinely useful, not because awareness lets you step outside the system, but because it changes what you’re looking at when group dynamics get strange. The colleague who seems inexplicably threatened by a new team member, the friend group that fractures after one person starts getting more attention, the online community that turns on its most vocal moderator through a coordinated campaign of subtle undermining, these stop looking like personality failures or bad luck and start looking like what they actually are: ancient social software running on modern hardware, doing exactly what it was designed to do, in contexts it was never designed for.

The wonder of it, if you can hold it from that angle, is how precisely it still works. A system shaped by selection pressures in small foraging bands tens of thousands of years ago still produces, reliably and rapidly, in a one-hour meeting of strangers, in a summer camp full of boys, in a Discord server about a television show, the same basic output: a ranked social structure, a set of norms enforced by the group, alliances and counter-alliances, and the continuous low-level hum of status accounting that everyone is doing and almost no one is admitting to.

You didn’t install this software. You can’t uninstall it. But you are, right now, running it, and so is everyone else in every group you’ve ever been part of. The question isn’t whether the hierarchy is there. The question is whether you’re concious and curious enough about the machinery to notice it working.