This is a compelling article about the nature of mathematical objects (here focusing the exposition on the natural numbers). Essentially, the paper argues that numbers cannot be any of their possible particular definitions (e.g., as particular sets, Church numbers, etc). Instead, when we talk of numbers, we speak of the abstract structure that relates them. So 2 is neither $(s (s 0))$ nor $\{\varnothing, \{\varnothing\}, \{\varnothing, \{\varnothing\}\}\}$ -- not these particular objects, but the relation that 2 has to 1 and 0 and 3 and so on, in whatever definition you want to give to the whole system. Though not explicitly mentioned, the paper argues for the categorical point of view of mathematics, where numbers would be defined by their universal property, which all of the particular definitions can be shown to satisfy.