A city functions when the things its inhabitants need are proximate to them, and nothing generates proximity like density. Put additional dwellings, shops and workplaces on the same ground and the average distance between any two of them diminishes; put fewer, and it increases. That is arithmetic, not ideology. Everything a city is asked to deliver -- short journeys, practicable transit, streets sufficiently busy to be pleasant -- is a consequence of it, because each is contingent on enough people being within reach of the same place concurrently. A bus route requires passengers per mile of route, and a shop requires customers per hour. Distribute the same population over four times the ground and both fail, not because anyone planned incompetently but because the arithmetic no longer functions. The appropriate measure of a city's success is how long it takes a resident to reach what she needs, and density is the one lever that moves it everywhere at once. The point is not that density is agreeable; many people find it disagreeable, and that is an independent argument. It is that every alternative to density operates by duplicating destinations rather than by relocating people, and there are limitations on how many schools, surgeries and workplaces a town can afford to duplicate. Beyond those limitations the only remaining variable is how far apart things are positioned. A city that disperses is not making a different trade-off; it is making the identical one worse.
[[PASSAGE B]]
The measure is appropriate and the lever is mistaken. What residents want is accessibility -- how much they can reach in twenty minutes -- and density is one mechanism for producing it, not the thing itself. A town of low density with a school, a surgery, a market and a bus stop on one street furnishes more accessibility than a dense district where residential use has been zoned away from everything else, and such districts are commonplace. Density raises accessibility only where uses are permitted to intermingle and where the network interconnects them; where either fails, the additional people are additional distance. Nor is the lever costless. Construction at high density is protracted and expensive, and the least expensive accessibility gains in most towns come from relocating a bus stop, opening a through path, or permitting a shop on a residential street. Aim at the measure directly and density will follow wherever it is genuinely the answer. None of this makes density a mistake. It makes it a means, and a means with conditions attached is not a lever that moves everything simultaneously. Where a district already intermingles its uses and interconnects them, additional people will indeed shorten journeys, and doing so is worth the expenditure. The error consists in treating that case as though it were the only one.
