Same village alarms. Bigger room.
Why does a platform of strangers still trip the same alarms a village used to?
A comment lands from a name we have never shaken hands with. The chest still changes. A count moves. The heat is the heat a room used to carry when everyone in it could see our face. The platform is full of strangers. The alarms are village-sized.
Bands became tribes, then institutions, then platforms. The wiring did not get a matching upgrade. The useful work is naming which old instinct the big room is using.
The budget the brain still spends
Robin Dunbar’s number is a proposed ceiling on how many stable, meaningful relationships one person can keep — about 150. It comes from a primate pattern called the social brain hypothesis: species with relatively more neocortex tend to live in larger groups. Dunbar ran that scaling onto humans and landed near a hundred and fifty. The precise integer has been argued over. What has traveled better than any single count is the constraint underneath it.
We do not relate to a hundred and fifty people in the same way. Roughly five sit in the innermost circle. Around them is a sympathy group of about ten to fifteen. Then a band of around fifty. Then the wider set of contacts that still count as real. Beyond that we can recognize hundreds of faces. Recognition is not the same as a bond we can keep warm. Time is the other limit. Primates spend a large share of the day grooming. Humans, in this account, spend a similar share of waking hours in conversation, the cheaper substitute. Language amplifies that work. It does not erase the budget. Each real tie still costs hours.
Digital networks let us collect thousands of connections. Studies of those networks keep finding that the inner circles — the ones that take reciprocal time — still cluster around those older sizes. Platforms make the outer layers louder. They do not issue a second neocortex. Past about 150 active ties, interaction frequency drops. A platform of strangers is a bigger room with the same lamps. The body still spends as if the people on screen were a band.
We and not-we, at industrial speed
Once a circle forms, the mind starts sorting. In-group and out-group are the ordinary names: we, and not-we. Henri Tajfel’s minimal group experiments in the early 1970s showed how little it takes. Arbitrary assignment — a preference for one painter over another, a guess about dots — was enough. People then favored their own when they divided rewards, with no shared history, no meeting, and nothing real at stake. Categorization alone triggered the tilt.
That finding is easy to over-read. It does not mean we are built to hate. Reviews of the prejudice literature have often found that in-group warmth does more of the work than active out-group hatred. We like our own. We trust our own. Everyone outside gets less of that warmth. On a platform, “outside” can be almost everyone, tagged in a hurry by a handle, a flag, a joke, a side.
The sort is old. Small bands needed a cheap way to know who to share with and who might raid. Evolutionary models treat that in-group tilt as an aid to cooperation inside a group that had to survive together. Platforms did not invent it. They gave it a room the size of a continent and a speed the village never had. Perceived out-group homogeneity — the sense that they are all alike — travels especially well in a stream of strangers, because types are faster to read than persons. A named us, under pressure, will treat the nearest them as the whole moral landscape. A platform can mint a named us before breakfast.
The new elder in the room
A village had elders, gossip, and a table. A platform has ranking.
Clay Shirky called the shift algorithmic authority: the choice to treat an unmanaged computational process — a ranking, a feed, a crowd filter — as a legitimate arbiter of what matters. The authority does not come from a person we can question. It comes from scale and convenience. The system extracts a signal from a flood of untrustworthy sources, and we let the extraction stand in for a judgment. On social platforms, feed algorithms order the stream by engagement and recency. What rises is treated as important. We rarely go far past what is put in front of us. Recommendation loops then narrow the diet. The algorithm becomes a kind of elder. We did not elect it. We defer to it because it is always on and always sorting.
Normative social influence is the older pressure underneath. We adjust what we say, and sometimes what we show, to stay acceptable to a group — to gain approval, or to avoid rejection. Morton Deutsch and Harold Gerard, in 1955, separated this from informational influence: the first is about belonging; the second is about using the group as a clue to reality. Solomon Asch’s line studies made the belonging engine plain: in a unanimous wrong majority, people still yielded in public on a substantial share of easy trials. Unanimity mattered more than headcount. One dissenter broke the spell.
A feed is a strange group. We often cannot see who is in it. We can still feel the imagined audience. Echo chambers, in this research, work partly because users conform to like-minded opinion to keep belonging. Algorithms that keep showing agreeing views make the imagined majority feel unanimous. The village alarm — do not be the odd one — rings in a room where the “group” is a ranking of strangers.
Bands, then tribes, then this
The scale-up is real. Bands were small enough to know. Tribes added names and stories. Institutions — laws, ranks, offices — let strangers coordinate without each person tracking each other person. Larger polities went past Dunbar’s ceiling with hierarchy and procedure. Trust was outsourced to roles. Platforms copied the scale of institutions and the heat of the band. They simulate community at a size no neocortex can actually keep.
None of this says platforms are unreal, or that strangers cannot matter. It says the alarms are running a program written for a different architecture. When a stranger’s sentence lands like a verdict from the table, that heat is information about old equipment. It is not a census of the species.
Which instinct is this room using
We do not need to leave every platform. We need a question that fits next to the thumb.
Ask which old instinct this big room is using.
Is this Dunbar’s inner circle, or a loud outer layer? Is this in-group warmth, or a ranking dressed as a people? Is this a real unanimous table, or a feed that has hidden the dissenters?
If the answer is an old instinct in a bigger room, several things follow. The heat can be real and still be local to the machinery. The sense that “everybody” thinks this is often a count of what the ranking chose to show.
Naming the instinct is not a sneer at belonging. Belonging is a real need. It is how we stay in the big room without letting village alarms finish every argument.
