The map that nobody drew

He built the smallest possible city. Two kinds of households. Some empty space. One mild preference: I am comfortable if enough of my neighbours are like me. If not, I move.

Then he repeated the decision. The result was not a gently textured neighbourhood. It was a border. The unsettling claim is not that cities are simple. It is that some of their largest visible patterns can be generated by a model with almost nothing in it.

Density gradients, districts, ethnic enclaves, dead seams between neighbourhoods: again and again the macro-pattern falls out of the same machine. You do not need to model a city. You need to model a decision, repeated.

The Map That Nobody Drew

Urban history is full of explicit decisions. Zoning boards draw hard edges. Banks ration credit. Highways are pushed through the wrong neighbourhood. The point of Schelling's model is not to erase any of that. It is to isolate a second force.

Suppose nobody plans the final map. Suppose each household only asks a local question about comfort. Not purity. Not domination. Just enough familiarity to stay put.

Braitenberg's trick was always to build the vehicle first and only then name what it seems to feel like from outside. The Schelling city rewards the same approach. Look at the map first. Ask who wanted it. Only then inspect the rule.

The Model With Almost Nothing In It

The vehicle is almost insulting in its simplicity. Each household occupies a square on a grid. It inspects its immediate neighbours. If the share of similar neighbours drops below a threshold, it moves to an empty square. Repeat.

That is enough. No one sees the whole city. No one reasons about demographics. No one intends a ghetto. Yet as the rule iterates, the line appears anyway.

The Gap

The emotional centre of the model is the gap between what feels mild and what looks extreme. A household that is content with only a third of its neighbours being similar does not sound fanatical. But iterate that threshold across a whole city and the collective outcome drifts toward near-total sorting.

The reason is mechanical. Each move that relieves one household slightly changes the composition around somebody else. Local comfort is contagious. So is local discomfort. The city is not complicated. The city is iterated.

Three Is Enough

Once you see the grammar, Schelling stops looking unique. Other systems with similarly tiny rule sets produce the same kind of surprise: rich geometry with almost no intelligence inside it.

The city starts to look less like a masterpiece of planning and more like a stable pattern produced by repeated local updates.

Boids give you spontaneous lanes on a crowded pavement. Cellular automata give you pulses of crowding, abandonment, and return. Reaction-diffusion gives you stripes and bands, as if a city could grow districts the way a fish grows markings.

The common structure is small. A local state, a local rule, a loop. The wonder is that this is enough. The horror is that it is enough.

Vacancy, Districts, and the Death of Streets

Movement requires slack. In Schelling's world, empty squares are not background. They are the channels that let households reshuffle. A little vacancy lubricates sorting. More vacancy creates visible seams: wide buffers, islands of similarity, streets that stop serving as bridges and start serving as borders.

That is one reason abandonment feels urban rather than merely empty. It changes the dynamics of who can move where, and which local decisions have somewhere to go.

What This Means for Blame

This is the philosophical payload, and it cuts both ways. If nobody explicitly chose the final pattern, the pattern can still be unjust. Emergence is not innocence.

History still matters because history sets the initial conditions. It decides who can move, who can borrow, which streets are cut in half, where the empty housing stock is, and how expensive discomfort is. The model does not absolve redlining. It shows how a city can keep rewriting an old intention long after the planner has left the room.

That is why Schelling's little grid still unsettles. A city can end up looking designed when all you really gave it was repetition.