The paper sketches and defends two instances of the strategy Let N’s be whatever they have to be to explain our knowledge of them—one in which N’s are natural numbers and one in which N’s are propositions. The former, which makes heavy use of Hume’s principle and plural quantification, grounds our initial knowledge of number in (a) our identification of objects as falling under various types, (b) our ability to count (i.e. to pair memorized numerals with individuated objects of one’s attention), (c) our (initially perceptual) recognition of plural properties (e.g. being three in number), and (d) our predication of those properties of pluralities that possess them (even though no individuals in the pluralities do). Given this foundation, one can use Fregean techniques to non-paradoxically generate more extensive arithmetical knowledge. The second instance of my metaphysics-in-the-service-of-epistemology identifies propositions (i.e. semantic contents of some sentences, objects of the attitudes, and bearers of truth, falsity, necessity, contingency, and apriority) with certain kinds of purely representational cognitive acts, operations, or states. In addition to providing natural solutions to traditionally un-addressed epistemic problems involving linguistic cognition and language use, I argue that this metaphysical conception of propositions expands the solution spaces of many of the most recalcitrant and long-standing problems in natural-language semantics and the philosophy of language.