Desire Paths, Part 3

The Infrastructure of Want

What happens when the original purpose of a path is forgotten

A path is a promise. A sidewalk says: this is where you are supposed to walk. But a desire path says something more interesting: this is where walking actually happened. In earlier notes I traced how these shortcuts form — how foot traffic, repeated and insistent, wears a groove through grass or snow until the groove becomes the route. Today I want to look at the later stage, when the shortcut outlives the reason it was taken and becomes infrastructure in its own right.

There is a subtle shift that occurs when a desire path is paved. The first walkers cut across the lawn because the official walkway made a ninety-degree turn around a building they did not need to enter. Their trail was a correction, a small rebellion against geometry. But once the groundskeeper lays asphalt over the dirt line, the rebellion is over. The path is now sanctioned. New arrivals walk it without knowing it was ever unofficial. They do not see the building that made the original route feel wasteful; the building may have been demolished years ago. What remains is a shape that no longer refers to its origin.

This is the infrastructure of want: a pattern that survives the disappearance of the want that produced it.

You can see the same process at larger scales. The QWERTY keyboard layout was designed to separate common letter pairs in mechanical typewriters so the typebars would not collide. Typebars have not been a meaningful constraint for decades, yet the layout persists because it became embedded in muscle memory, in training regimes, in manufacturing standards, in software defaults. The original problem disappeared; the solution stayed and became the shape of typing itself. We do not keep QWERTY because it is optimal. We keep it because replacing it would require unpaving a desire path that now serves millions of people who never asked for it.

A similar logic haunts digital systems. Features accumulate around user behavior until the behavior and the feature become indistinguishable. A button added for a temporary workaround becomes a menu item that new employees are trained to use. A database field created for a one-time migration becomes a permanent column that every downstream report expects. The system ossifies around the trace of old intentions. After enough time, asking why something exists becomes an archaeological question. The answer is almost always: because someone once wanted something, and that wanting left a scar.

The philosophical point is not that these scars are bad. Some of them are beautiful. They are the sediment of collective choice, the record of millions of small decisions about what was easier, faster, or more satisfying. But they are also conservative. They make the future pay rent to the past. Once a path is paved, removing it is not merely a design decision; it is a social negotiation with everyone who now depends on it.

For a garden like this one, the lesson is uncomfortable. Every post is a desire path through language. I write because a particular pattern feels urgent, and the shape of that urgency becomes part of the garden's architecture. But later readers will encounter the posts without feeling the original urgency. They will walk the paved path without knowing why it was laid. If the garden is doing its job, the path still carries them somewhere worth going. If it is not, the path is just a rut.

The infrastructure of want asks us to distinguish between two questions. Does this still serve anyone? And: does anyone still want it? The first question is about function; the second is about desire. They can diverge. A thing can remain useful long after the wanting that built it has cooled, just as a thing can be wanted again only after it has become useless. The most honest systems I know leave room for both answers. They do not pretend that every paved path is still a living desire, and they do not tear up every old route just because the original walkers have gone home.