Pattern 001 A Pattern Landscape

In 1977 Christopher Alexander, Sara Ishikawa, and Murray Silverstein published A Pattern Language. The book tried to define something like the human experience of space, and it did so through a long series of related though distinct typologies of encounter: 253 of them, from the scale of the region down to the placement of a window seat. Each pattern named a recurring situation, described the tension at its heart, and proposed a form that resolved it. Each nested inside larger patterns and contained smaller ones. Each came with criteria for success.

The idea was as elegant as it was optimistic. It was also anything but succinct, and, by my account anyway, naive to a fault. This was a decade after the Summer of Love, written in Berkeley, after the wave had crashed on Big Sur and drawn back out. The book still believed that if you could only describe good places well enough, people would build them.

Its success as a design methodology is rightly debated and contested. It gave us Seaside, Florida, after all: the pattern book as a recipe for a town, and the town as a film set. But the book found its true influence somewhere else entirely, in the seemingly unlikely and then still nascent discipline of computer science that washed up, and remains, on the coast of Northern California.

Patterns for machines

Programmers took to Alexander for a simple reason. He had solved a problem they had. Software, like a city, is a system too large for any one person to hold in their head, built by many hands over time, in which local decisions have global consequences. What programmers needed was a way to name recurring problems and their proven solutions so that the knowledge could be shared without being reinvented. Alexander had already written the format.

In 1987 Ward Cunningham and Kent Beck presented a short paper applying pattern languages to object-oriented programming. Seven years later a book called Design Patterns, written by four authors and so known as the Gang of Four, became one of the most cited texts in the field. Its structure was lifted directly from Alexander: name the pattern, state the problem, state the forces at play, give the solution, note the consequences. When Cunningham later built a tool for programmers to collect and edit patterns together, he called it a wiki. It was the first one.

So the book meant to describe the experience of a courtyard ended up as a heuristic for structuring code. This looks like an accident of history. I do not think it is.

The distance is smaller than it looks

It is tempting to treat computation and the city as opposites: one abstract, the other material; one designed from scratch, the other accreted over centuries. But the city has always been an information problem before it was a building problem. Who lives where. Who owns what. Who is moving through which street at which hour, and whether they belong there. Every administrative invention of urban life, the census, the cadastre, the street address, the house number, the zoning map, is an attempt to make the city legible to itself.

Put differently, the city has been continuously remade through striving for perfect knowledge about itself. It wants to see all of itself at once and it never can, so it builds instruments that approximate that view, and then it rebuilds itself around the instruments. Haussmann's boulevards were many things, but among them they were sightlines. The grid is a data structure before it is a street plan.

Seen this way, the programmers were not borrowing from an alien discipline. They were recognizing a relative.

The city as a surveillance apparatus

This is where the argument turns, because the same lens changes how we should read the present. The contemporary city is full of cameras, sensors, license plate readers, and phones that report their position several times a minute. It is common to describe this as an aberration, a symbol of technological overreach, a departure from some earlier city where people moved unobserved. That earlier city did not exist.

The city has always been a surveillance apparatus. Its basic social fact is visual: to be in a city is to be seen by people who do not know you, and to see them in turn. This visual sociality is not a side effect of density. It is the medium through which urban order has been produced for as long as there have been institutions to produce it.

Consider modern policing, which in its recognizable form dates to seventeenth-century Europe. In 1667 Louis XIV created the office of Lieutenant General of Police for Paris and gave it to Gabriel Nicolas de La Reynie. The word meant something broader than it does now. Police was the whole administration of the city: markets, sanitation, fire, morals, the poor, and crime. What La Reynie needed above all was to know what was happening in the streets, and the streets at night were unknowable.

So the same year, and under the same authority, Paris began hanging lanterns. Glass boxes with candles inside, strung on ropes across the street at regular intervals, lit by neighborhood residents on a schedule and paid for by a tax on them. Within a few decades there were thousands. Amsterdam followed with oil lamps designed by the painter Jan van der Heyden. London followed. The lantern was not a convenience for pedestrians. It was an instrument of administration. It extended the hours during which the city could be seen, and therefore governed, and it did so by turning every lit street into a space where anyone might be watching.

People understood this perfectly well. Smashing lanterns became a recognized form of protest. In 1789 the Paris crowd hanged officials from lantern brackets, and à la lanterne entered the language as a call for justice, which tells you what the lantern had come to stand for.

Public lighting, then, is the first sensor network. It did not record, but it did the thing all sensor networks do: it converted a portion of the city that was dark to the state into one that was legible. Every expansion of lighting since, from gas to arc lamps to LED, has been justified in the same terms, safety and order, and has produced the same result, more of the city visible to more eyes for more of the day.

What follows

None of this is an argument that current surveillance is harmless, or that a camera is the same as a candle. Scale, persistence, and the capacity to search change what visibility means. But it is an argument against nostalgia. There is no prior city to return to in which seeing was innocent. The city was built to see. The question worth asking is not whether it should, but who holds the view, what it costs to be seen, and what it costs to be invisible.

Alexander thought that describing patterns of good experience would let people build good places. The programmers who took his book thought that describing patterns of good structure would let them build good systems. Both were half right, and for the same reason: a description of a place is also a means of controlling it. The pattern language was, from the beginning, a way of seeing the city whole. It should not surprise us that it was adopted by the people who were about to build the machines that see the city whole.