communicoupling
When Mark Granovetter asked Boston professionals how they had found their jobs, the modal answer was someone they saw rarely — an old colleague, a friend-of-a-friend, a contact from a former life. Close friends and job advertisements trailed far behind. Why would the people who care about you least be worth the most, informationally? His answer became one of the most-cited claims in social science: because strong ties close into triangles, any bridge between social clusters is almost necessarily a weak tie — so the weak ties carry the non-redundant information, and their removal, not the strong ties', is what fragments society (Mark Granovetter, The Strength of Weak Ties, 1973; Getting a Job, 1974).
Your close friends know each other — that is part of what makes them close. And a circle in which everyone knows everyone is a circle in which everyone knows the same things. The job opening your best friend hears about, you have already heard about. The acquaintance in another industry, seen twice a year, moves through rooms you never enter; whatever reaches you from them is, almost by construction, news.
Below, a clustered network: five villages of thick warm friendship, a tunable scatter of thin cool threads, and a job lead seeded at one node. Strong ties transmit faster per contact — the model concedes them that advantage in full. Watch what carries the lead anyway.
The forbidden triad
Granovetter measured the strength of a tie as a bundle: the time spent together, the emotional intensity, the intimacy of what is confided, the reciprocal services exchanged. Strong ties then obey a structural law. If A is strongly tied to both B and C, the two of them share so much time and exposure through A that a B–C tie almost always forms — triadic closure. The triangle with two strong sides and no third side is the forbidden triad: so rare in sociometric data that the theory treats it as absent. The closure dial enforces exactly this rule on the instrument's web — raise χ and open strong-strong pairs get sealed into triangles, while any strong tie that tries to bridge two villages is either closed into triangles itself or demoted to weak.
The consequence is a theorem about geography. Strong ties, forced into triangles, can only live inside dense clusters. A bridge — the sole line between two clusters — has endpoints with no mutual friends; were it strong, closure would already have manufactured that mutual friend and the bridge would be a triangle instead. So bridges are, almost necessarily, weak. Sweep χ to 1.0 and watch the structure line: the count of strong bridges falls to zero. And note what the model concedes: per contact, strong ties genuinely transmit better — the σ:ω dial gives them three times the rate. The weak ties carry the world anyway, because they are the only ties standing where the new information is.
What to try
Load Granovetter's Boston and press New lead. Strong ties fire at 3× the weak rate, yet carried by a weak hop settles near 80% — everything outside the home village crossed a thin thread to get where it went.
Press Cut all weak ties: the web falls into fragments and reach collapses toward one village in five. Now Cut same # strong instead: the same edge count removed, the villages thin — and the world stays 1 component, reach barely dented.
Drag closure to 0%: a few thick ties survive as bridges (the structure line counts them). Push it to 100%: every bridge in the web is weak. The forbidden triad is doing the sculpting — closure decides where weak ties can live.
The society's circulatory system
In Granovetter's Boston sample, of the professionals who found their job through a personal contact, more than four in five saw that contact only occasionally or rarely. The finding scandalises our intuitions about loyalty, and the mechanism is bloodless: your close friends' knowledge already overlaps yours almost completely. Ask the ledger — by the end of a run, most strong-tie contacts are redundant, warm repetitions landing on people who already know, while the weak ties spend most of their few firings delivering something new. The acquaintance is valuable to you for precisely the reason they are distant: they are embedded in a cluster you are not, and every cluster is a room where different things are said.
Scale this up and it becomes a claim about what holds a society together. A community knit only of strong ties is cohesive and closed — a warm room with the windows painted shut; trust circulates, and so does the same information, forever. The weak ties between such communities are the circulatory system of the larger body: jobs, ideas, technologies, marriages, and rumours move between worlds on threads too thin to bear any other weight. That is why the surgery is so asymmetric. Cut every weak tie and the components seal shut — the lead never leaves its village, however lovingly it circulates there. Cut the same number of strong ties at random and clusters grow threadbare, but the world remains one world. What we experience as the fabric of society frays from the middle of friendships far more slowly than it shatters at the acquaintances.
Next doors
The mathematics of why a handful of thin threads makes the whole world reachable in a few hops lives in small worlds — weak ties are the shortcuts that collapse the distances. What happens to the person who occupies a bridge — the profit they extract and the loneliness they carry — is the subject of structural holes. And Granovetter built a second machine with the same signature move — outcomes decided by structure rather than sentiment — in thresholds & the crowd, where two identically angry crowds meet opposite fates. Both papers make the same wager: the pattern of connection explains what the feelings inside it cannot.
The mapping
| In the model | In the world |
|---|---|
| strong ties | The people who'd help you move. |
| weak ties | The people you'd email after three years, opening with an apology. |
| the forbidden triad | Why your two best friends know each other. |
| a bridge | The old colleague in another industry — your only line into that world. |
| redundancy | What your close friends can tell you that you don't already know: little. |
| the surgery | What actually holds the big world together, and what merely thins it. |
Where it tears
A thin thread will hand you a job lead; it will not lend you money, mind your children, or sit with you in a crisis. Granovetter said so himself, and the later work on core discussion networks sharpened it: news and support travel on different ties, and for everything that requires sacrifice you need the people who'd help you move. This page must not romanticise acquaintance — the thin ties carry news precisely because they carry so little else.
The job-finding result comes from white-collar Boston in the early 1970s, and replications vary by class, country, and era: where work moves through formal credentials, union halls, or family obligation, the weak-tie advantage shrinks or inverts. In the platform age it has held up strikingly well — a 2022 field experiment on twenty million LinkedIn users found that moderately weak ties caused the most job moves — but the honest statement is a distribution of results, with this page's clean asymmetry at its centre rather than everywhere.
Here a weak tie costs nothing to hold. In life, the person who is the bridge absorbs its costs: divided loyalties between clusters that mistrust each other, endless translation labour, and a structural loneliness — a member of every camp, a citizen of none. The simulation counts what bridges deliver and stays silent on what they endure; that story continues in structural holes.