It sounds like positive feedback loop. Some reason (usually jobs) results in higher density, which results in a shortage of housing and higher prices, and that’s an incentive to build denser housing so that more people can move in. The available workforce might attract more employers, and the loop continues.

You might as well ask how cities get started and grow larger. Contrast with farmers.

A positive feedback effect will keep going until there’s negative feedback counteracting it. That’s almost tautological. The specifics will be more interesting.

I hadn't thought of it like this before, but I think he's arguing that there is a network effect. High density is more convenient, which induces demand.

Yes exactly it is obvious. The article attempts to address this by saying this effect is not the full story, which seems highly unconvincing to me. His argument is stripped of density, Manhattan is just another island like lots of islands in Maine that didn’t end up having very expensive real estate. But firstly lots of islands (eg Martha’s Harbour, Long Island, which isn’t really an island but close enough) in the area are significantly less dense and have very expensive real estate so I don’t think that argument bears any scrutiny at all.

In general his thought experiments with data in the article just seem completely handwavy. “Imagine if I added 5x more houses to Oakland”. Well we can imagine that but we would likely be wrong about what would happen so it didn’t prove anything. Besides anything the most likely outcome would be you would have a lot of empty houses.