A coordinate-Iess space wouId not have spatiaI dimensions either, it wouId just be a space in name, and in truth a non-space.
I’m not quite sure what you mean by this, but a non-metrizable space is certainly a space, and has well-defined dimensionality. The informal notion of “number of coordinates needed to specify a point” is informal for good reason. It’s not accurate(nor coherent, in fact), neither is the informal(not even sure if it’s common, as I haven’t seen it elsewhere) notion that a space is a set of points and axes or dimensions(dimensions aren’t elements of a space, dimensionality is any one of various different properties of different kinds of space that usually agree for well-behaved spaces)
If you take “1-dimensional” to mean “you can specify every point in the space uniquely with only one real number”, then 2D spaces are necessarily, by this definition, 1D. There is a bijection from Rx2 to R. In fact, there’s various ones(I won’t be listing any in particular, but if you want to find one, it’s relatively easy to google one)
Take some bijection from Rx2 to R, let’s denote it by F
Then, take some space X, such that every point is of the form (x, y) for real numbers x and y
Then, there is a way of uniquely identifying each point of X with ONE real number z, e.g. it’s, by the previous definition, 1-dimensional.
Take F((x, y)). Since F is injective, F((x, y)) is unique, and, since it’s a function, it takes
any value (x, y) for real x and y. Hence, F((x, y)) is the desired real number uniquely associated to that point in space.
You can do the exact same thing with 3-dimensional, 4-dimensional, etc.