Well, the counterargument is that in theory, you can imagine a way to create structural color regardless of substrate. So imagine a technology that shines a laser on a car or a block of concrete and makes it blue; I'd argue that's correctly "without chemicals".

Of course, I doubt you can do that to any random substrate, since the color will depend on the properties of the material.