Wayfinding
A Socratic walk-through of wayfinding — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #How do people build a mental map of an unfamiliar place?
The phrase "mental map" does a lot of quiet work. It invites you to picture a small chart in the head that you unfold and consult. But test it. You can walk a route to work daily and be unable to draw it; be certain one place lies north of another and be flatly wrong; know two shops for years and be astonished to find they back onto each other. A map does not fail in those particular ways. So what is the thing in your head, if it is not a map?
Reasoning it through
REASONING #Start from the raw material. What can you acquire from one walk through a strange town? Not positions — you were never above it. You get a sequence: a red door, then a hill, then a church where you turned left. So the first thing to accumulate is distinctive things you can recognise again, the second what to do when you see them. That second kind is a procedure, not a picture, and you can hold a great deal of it with no sense of layout — the state of someone who can get to work but cannot say which way is north.
Notice what a procedure cannot do. It cannot be run backwards without effort, since you learned the turns from one direction and the scenery looks unfamiliar from the other. And it cannot answer a question it was never a recording of: if a road is closed a route gives you nothing, holding nothing about what lies between its steps.
So how does the third kind — the sense of relative position that lets you cut a corner you have never walked — appear? Two ingredients do most of the work. One is a running self-motion estimate: as you move you keep a rough tally of your displacement, so a bend and a bend and a bend sum into "I am roughly back where I began". The other is stitching: two routes that pass the same recognisable landmark can be joined there, and once joined the geometry between them becomes available for the first time. Do you see why landmarks matter more than their information content suggests? They are not memory aids; they are the seams.
The classical account presents this as a sequence — landmarks, then routes, then survey knowledge — which captures the ordinary case but is worth flagging as contested. The evidence does not support a strict staircase: survey knowledge can arrive early, especially from a map or a view from height, and many never develop much of it for a place they have lived for years.
The analogy
THE ANALOGY #Think of a jigsaw assembled not from a box but from photographs taken on separate walks. Two can be joined only where something recognisable appears in both, and until a joining shot exists, two well-known districts float as unrelated islands.
jigsaw pieces are rigid and either fit or do not, whereas the mind will cheerfully force a join and deform the pieces to make it work — and it fills gaps with confident inference rather than holes, which is why your map almost never feels incomplete even where it is badly wrong.
Clarifying the model
THE MODEL #That last point matters most, because the distortions are not noise — they are systematic fingerprints of the assembly. People encode places hierarchically, a city inside a state inside a country, then answer about the small things by reasoning about the big: asked whether Reno is east or west of San Diego, most reason that Nevada is east of California and answer east, when Reno is in fact west of it. Layouts also get regularised — bends remembered as straighter, junctions as more nearly right-angled, neighbourhoods rotated to line up with a coastline or a grid. And routes with many turns are remembered as longer than smooth routes of equal length, distance inheriting its estimate from the effort of walking it.
One correction. This is not purely abstract psychology: cells in the hippocampus fire when an animal is in a particular place, and cells nearby fire in a repeating spatial lattice. But the step from those recordings to "this is your mental map of Lisbon" is one the research has not closed, and they are more honestly the substrate than the explanation.
A picture of it
THE PICTURE #How to readRead the four inner branches as ingredients of the thing in your head rather than stages in a queue; the fifth is what a designer can do about them. The first two are what a single walk gives you, the third what becomes possible only once separate walks have been joined — which is why landmarks reappear there as seams. The "seams that show" branch is the diagnostic one: each item is a predictable error a real map would not make, evidence that this one was assembled rather than copied.
What became clearer
WHAT CLEARED #There is no map in the head — only a heap of recognisable things, procedures for getting between them, and a patchy sense of layout stitched wherever two experiences overlapped, with the gaps filled by confident inference. The characteristic errors of navigation are not failures of the map; they are the shape of how it was built.
Where to go next
ONWARD #- Why the "you are here" map is easier to use when rotated to match the way you face.
- How satellite navigation, by supplying routes without requiring stitching, appears to leave weaker survey knowledge.
Key terms
TERMS #| Term | What it means |
|---|---|
| Landmark, route and survey knowledge | recognising places, knowing procedures between them, knowing relative positions. |
| Path integration | keeping a running estimate of your position from your own movement, unaided by landmarks. |
| Hierarchical encoding | storing places nested in larger regions, then inferring their relations from the regions'. |
Every term the collection defines is gathered in the glossary.