Fifteen footsteps are enough to create a trail. Trampling studies have measured this consistently: fifteen passages over a patch of vegetation produce a visible line through the grass. The line attracts the sixteenth passage, because walking on a faint trail is easier than walking on undisturbed ground. The sixteenth passage deepens the line. The seventeenth follows the sixteenth. Within a few days, the trail exists as a fact that appears to have always been there, its origin in any single decision already erased.
Urban planners call these desire paths. The term appears in transportation planning literature from the mid-twentieth century, though the phenomenon is older than cities. Other names exist in other languages: olifantenpad in Dutch — elephant path. Kemonomichi in Japanese — beast trail. Chemins de l’ane in French — donkey path. Every name attributes the path to an animal following instinct rather than instruction.
Broadway runs the length of Manhattan on a diagonal that cuts across the city’s grid. Most of Manhattan’s streets meet at right angles. Broadway ignores the grid, running from the southern tip of the island northwest to the Bronx. The diagonal produces the city’s most distinctive intersections — Times Square, Herald Square, Madison Square, Union Square — all of them created by Broadway’s refusal to conform to the surrounding geometry.
The street may follow a trail made by the Wecquaesgeek people before European colonization. The trail connected settlements along the island’s ridge, following the easiest route through the terrain. Dutch colonists found the trail and used it. The English inherited the trail from the Dutch. The city planners who imposed the Manhattan grid in 1811 could not straighten the road that was already there. The grid bent around it.
A desire path made by feet walking on dirt in the seventeenth century or earlier became Broadway, the most expensive commercial real estate on the planet, because the first path attracted the second walker, the second attracted the third, and no subsequent generation of planners could undo the accumulated weight of prior use.
Michigan State University’s campus was established in 1855. Early construction budgets did not include concrete or paved walkways. Students walked across grass to reach their classes, and the routes they chose wore lines into the ground. The lines became the university’s pathway system.
A century and a half later, a campus administrator asked Steve Troost, the university’s campus planner, whether the walkways around West Circle — one of the oldest sections of the campus — were still needed. The paths had been paved over the student-made trails decades earlier, and the resulting network looked strange from above: no right angles, no geometric regularity, just an irregular web of paths crossing the lawn at apparently arbitrary angles.
Troost’s team conducted video and photograph surveys of pedestrian traffic across the area. They measured where people actually walked. The conclusion: almost every paved path was necessary. The system that students’ feet had designed a century earlier still matched the traffic patterns of the current population. The planners found, in Troost’s words, “a very small amount of walkways that really weren’t needed.”
The paths had worked for over a hundred years because they were never designed. They were observed.
The University of California, Irvine, reportedly tried the opposite of planning. The campus planted grass everywhere and installed no pathways at all. For a year, they watched. Students, faculty, and staff walked where they walked. Some paths were predictable — straight lines between adjacent buildings. Others curved in ways that made no geometric sense until the planners investigated and found that the curves followed shade patterns in summer and avoided wind tunnels in winter. The walkers were optimizing for comfort, not distance.
After a year, the university paved the paths that had appeared. The campus still feels navigable decades later. The paving followed the data. The data was produced by feet that knew nothing about data.
Christopher Coutts, an environmental planning professor at Florida State University, co-authored an academic paper on informal pedestrian infrastructure. He wanted to call the phenomenon “deviant pedestrianism.” His reviewers objected to the term. He used “desire paths” instead.
The rejected term is revealing. “Deviant” means departing from an established course. The word carries moral weight — deviance implies wrongdoing, a failure to comply. A desire path, from the perspective of the designer who built the official walkway, is an act of deviance. The walker was supposed to turn left. The walker went straight. The grass died. The designer’s intention is visible in the paved path. The walker’s intention is visible in the dirt one.
Whether the dirt path represents a design failure or a design insight depends entirely on whose intention you privilege. If the designer’s intention is the standard, the desire path is vandalism. If the walker’s behavior is the standard, the desire path is feedback. The same physical mark in the grass — a line of dead vegetation — is either damage or data.
The self-reinforcing quality of desire paths is their most important feature. A trail that fifteen people created will be used by thousands, not because thousands independently chose the same route, but because the trail already existed when they arrived. The first walkers made a choice. Every subsequent walker made a different, easier choice: follow the visible path. The origin was decision. The continuation is convenience.
This produces a specific problem for anyone trying to understand why a path exists. The current traffic volume on the path is not evidence that the path is optimal. It is evidence that the path exists. An equally good or better route nearby will attract little traffic if no one has walked it yet, because it lacks the visible trail that lowers the cost of choosing it. The existing path has an advantage that is entirely historical: someone walked it first.
The trampling research confirms this. Fifteen passages create a trail. The trail attracts all subsequent use. Remove the trail — reseed the grass, cover the dirt — and a new trail may form on a different line, depending on which fifteen walkers happen to cross first. The final network of paths across a landscape is partly a map of optimal routes and partly a map of historical accidents that became self-sustaining.
The Finnish word for desire path is oikopolku. Oiko means direct. Polku means path. The direct path. The name contains no animal, no deviance, no desire. It describes geometry: the straight line between two points, which is what a person walks when nothing forces a detour.
The English name is different. “Desire” implies want, intention, longing. A desire path is a path you want to take. The word suggests the walker is choosing against the designed path out of preference, not efficiency. The connotation is slightly transgressive — desire as a force that overrides propriety.
Both names are accurate. Both are incomplete. The Finnish name captures the geometry but misses the psychology. The English name captures the psychology but misses the geometry. The actual path is both: the shortest route, chosen because it is shortest, maintained because choosing it is easier than choosing an alternative, visible because maintenance makes it visible, and reinforced by visibility in a cycle that erases its own origin.
A desire path is what remains after planning has been tested by use. It is the gap between prediction and behavior, made physical, made permanent, made into the ground itself.