Abstract Mathematical exactness is commonly treated as a uniform notion, applying equally to all numbers once formal definition is fixed. This paper argues that this assumption obscures a structurally significant distinction in mathematical practice. I introduce a qualitative–quantitative distinction between numbers that terminate specification as self-contained magnitudes and numbers whose exactness is sustained only through defining relations, procedures, or constraints. The distinction does not concern numerical identity or membership in the real numbers, nor does it rely on epistemic limitations or representational choices. Instead, it concerns the conditions under which numerical exactness is realized and maintained. I show that this distinction is invariant across base choice, computation, algebraic closure, and limit behavior, and that it clarifies the role of exact relationships in both pure mathematics and bridge equations connecting mathematics to physical theory. The framework is compatible with structuralist accounts of numerical identity while addressing a different explanatory question, namely how exactness persists for different kinds of numbers in use.
Ian D. Reynolds (Fri,) studied this question.