> But there is nothing to support the claim that narrower lanes speed up traffic
My impression is that the opposite is actually true: narrower lanes tend to make drivers feel more nervous/wary, and they slow down as a result.
Lane-narrowing (and "perceptual" narrowing, adding features that make it seem narrower to drivers without actually changing the amount of space) seems to be a fairly common traffic-calming technique.
This comes from simple fluid dynamics. When a fluid is forced into a narrower pipe, it goes faster. (Venturi also tells us there is a drop in pressure, but it's not clear how this advanced aspect is applicable here.)
> One is that, for the most part, driving lanes weren't actually eliminated when they bike lanes were built — they were simply narrowed.
But there is nothing to support the claim that narrower lanes speed up traffic. Also this phrase is copy pasted verbatim further down the article.