Unplanned settlement form
A Socratic walk-through of unplanned settlement form — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #Why do old towns on different continents share street patterns that nobody ever designed?
Walk the old core of a market town in England, a medina in Morocco, and a hill town in Anatolia, and the same features keep turning up: a few curving routes converging on a widened space, irregular blocks between them, narrow lanes running behind the buildings, dead ends, and a ring road tracing where a wall once stood.
Nobody drew any of it. The three towns had no contact and no shared building tradition. So either this is the eye finding a pattern it wants, or something is generating the same form from different starting points. Which?
Reasoning it through
REASONING #Ask what exists first in a place that later becomes a town. Not buildings — movement. There is a ford, a spring, a mill, a shrine, a landing place, and people walk between them. The lines they walk are not straight: they are the least-effort routes around a slope, a marsh, a boundary. Before a single wall is raised, the geometry of the settlement has already been drawn by feet.
Now ask what a builder wants. Frontage on the busiest route — because that is where custom, water and light are. So the first buildings line the existing tracks. And notice what has just happened: the moment a house has its wall on a track, the track's alignment is fixed. You can rebuild the house, but you cannot straighten the street without acquiring somebody's land. A temporary path has become a permanent one, bends included.
Then the third ingredient, which is the one that makes the pattern last a thousand years: the plot. Land gets subdivided into holdings, usually long narrow strips running back from the frontage, because frontage is the scarce thing. Those boundaries are legal objects — inherited, mortgaged, litigated — and they survive fire, war and complete rebuilding. This is why a town's plan is its most conservative layer, changing far more slowly than its buildings or its uses. M. R. G. Conzen's morphological work makes exactly that separation, and it is the reason ground plans can be read as documents.
The rest follows from repeating those steps. Where routes converge, traffic is densest, so traders occupy the widening and a market place appears without anyone laying one out. Backland is built over and needs access, so lanes appear behind the plot series. Growth happens at the edge and in the gaps rather than by extending a lattice, so the street network branches like a tree, with many dead ends. A defensive perimeter will be as short as possible for the area it encloses, pushing the whole toward compact and roughly circular — and when it is demolished, its cleared strip is the only wide continuous land available, so it becomes a boulevard or ring road.
Nothing there is culturally specific. Walking distances, cart geometry, the value of frontage, the persistence of property, incremental growth — these are the same in Anatolia and Wessex, which is the answer: the towns converge because the generating rules are shared, not the plan.
That gives a sharp test. If the form comes from accretion under local rules, then settlements founded in a single act should not show it, whatever the culture — and settlements that grew slowly should show it, even in cultures that love grids. Both hold. Roman coloniae, Chinese planned capitals, medieval bastides, Spanish colonial towns under the Laws of the Indies and the Manhattan grid of 1811 are rectilinear, across entirely unrelated traditions; and organic fabric appears in those same cultures wherever growth outran the survey. The refuting observation would be a documented all-at-once foundation that came out irregular and convergent, or a slowly accreted town that produced a rectilinear grid with no survey behind it.
The analogy
THE ANALOGY #Think of a path worn diagonally across a lawn, and then a fence built along it. Before the fence, the path is a suggestion and could still be moved. After it, the diagonal is permanent — and every later fence, gate and flower bed is aligned to the first one, so the shape of a shortcut somebody took becomes the shape of the whole garden.
a lawn path can be reseeded and rerouted at any time, whereas a street's alignment is held by property law rather than by habit — which is why the pattern survives events that destroy everything standing on it. After the fire of 1666 London rejected Wren's proposed new plan and rebuilt on its medieval streets, because the plots underneath had owners and the streets did not need to be redesigned to be rebuilt.
Clarifying the model
THE MODEL #The phrase in the question — "nobody ever designed it" — is the thing most in need of correction. These towns were not designed as wholes, but they were intensely rule-governed at the small scale: burgage plots granted at standard widths, building lines and rights of light, servitudes over passages, and in Islamic jurisprudence a body of law on street widths, drainage and overlooking that shaped the medina in detail. Emergence here means the absence of a plan for the whole, not the absence of rules. The convergence across continents is evidence that different rule-sets, applied incrementally to the same problem, land in the same region of possibility.
Three honest qualifications. Some of the resemblance is topography rather than process — put any settlement on a spur above a river bend and it will look like the others there. The claim that these patterns share measurable quantitative regularities, as space syntax research argues, is contested in both its methods and its strength. And the tidy contrast between organic and planned leaks: grids can also emerge without a master plan, when a landowner repeatedly subdivides in the same standard way, so rectilinearity is evidence of a repeated rule, not necessarily of a single designer.
A picture of it
THE PICTURE #How to readRead down the stages in order; each is a step in accumulation, not a decision anyone took. The first two entries under each stage are what happens and what it leaves behind, and the second is the load-bearing part — notice that from stage two onward every entry is about something becoming fixed. That is the argument: the plan is not designed at any stage, it is deposited, and each deposit constrains the next. The last stage explains a feature that puzzles visitors most, a ring road with no wall on it.
What became clearer
WHAT CLEARED #The shared form is not a shared idea. It is what you get when movement precedes building, frontage is valuable, property boundaries are durable, and growth arrives a plot at a time — conditions that held in most pre-industrial societies regardless of culture. Each of those steps freezes something: the walk becomes a street, the street becomes a plot boundary, the boundary outlives the building. What looks like an inherited design is really an accumulated record of decisions, none of which could see the shape they were jointly making.
Where to go next
ONWARD #- Why some planned grids decay back toward irregularity while others hold their lines for centuries.
- Neighbouring pieces divide the question: City location and Fall-line cities ask why a town is at a particular spot, and Town spacing asks how far apart towns end up; this one asks only why, once a town is there, its inside takes the shape it does.
Key terms
TERMS #| Term | What it means |
|---|---|
| Burgage plot | a long narrow urban holding running back from a street frontage, typical of medieval town grants. |
| Town plan (in morphology) | the streets, plots and block pattern, treated as a layer distinct from the buildings on them and changing far more slowly. |
| Desire line | the route people actually take between two points, as opposed to the one provided for them. |
| Space syntax | a research approach measuring street-network connectivity and its relation to movement; influential and contested. |
Every term the collection defines is gathered in the glossary.