Mathematics is both, in a precise division: we invent the framework—axioms, definitions, notation, and which structures to study—but the theorems that follow are discovered, not chosen, and its uncanny fit to the physical world is a real, unexplained fact.
The invented part is genuine. Which axioms to adopt, what to define, which objects to bother with, and the entire symbolic apparatus are human choices; we could have built other systems, and sometimes do (different geometries, different set theories).
In that sense mathematics is a made thing, a family of formal games we design.
But once the rules are set, the consequences are not up to us—and this is the discovered part.
Given the axioms of arithmetic, that there are infinitely many primes, or that a particular one is prime, is forced: no amount of preference or convention can make it otherwise, and mathematicians experience finding these results as encountering something resistant and objective, not as decreeing it.
You can choose the game; you cannot choose who wins it.
This is why even a formalist who denies that numbers "exist" must still accept that the theorems hold given the rules.
Two deeper facts keep the metaphysics open. First, the "unreasonable effectiveness": mathematics invented for its own sake repeatedly turns out to describe physical reality with astonishing precision, which suggests some correspondence between mathematical structure and the world—unexplained, and hard to square with pure invention.
Second, whether mathematical objects exist independently (Platonism) or are useful fictions (formalism, nominalism) remains unresolved, because both can accommodate the practice.
The defensible synthesis: an invented framework with discovered consequences, whose fit to nature is a genuine puzzle. This would need revision only if theorems turned out to be matters of convention after all—their felt and practiced necessity says they are not.
Both—and the "both" is exact, not a dodge. You invent the rules (the axioms, definitions, and notation are human choices), and then you discover what those rules force to be true (the theorems aren't up to you). You can choose the game; you can't choose who wins it.
The clean way to hold it:
Why this matters even if you never do proofs: it explains why mathematics is simultaneously a human creation and utterly non-negotiable. And it flags a genuine mystery worth respecting—math invented for pure curiosity keeps turning out to describe the physical universe with eerie precision, as if reality were somehow written in it. Nobody fully understands why.
The practical bottom line: whether numbers "really exist" in some Platonic heaven or are just very good fictions doesn't change how you do math or trust it. What matters is the structure—invent the framework, then submit to the results it forces. The freedom is all in the setup; after that, you're taking dictation from consequences you built but can't control.
The Game You Can't Cheat.
Chess is invented—someone chose the board, the moves, the rules—yet no one can decide that a given position is checkmate; that follows, forced, from rules freely chosen. Mathematics is the same paradox at cosmic scale: we invent the axioms as freely as we invent the rules of a game, and then find ourselves unable to control what they imply. You made the rules and cannot cheat them. This is the strange doubleness at its heart—a human artifact that, the instant it exists, turns around and dictates to its maker, handing down truths no committee could overturn. We built a house and discovered it has rooms we never designed.
The Language the Universe Answers In.
Here is the miracle that unsettles every "it's just invented": a mathematician playing with abstractions for their own beauty—imaginary numbers, non-Euclidean geometry, group theory—invents a structure of pure thought, and decades later a physicist finds it was the exact language the universe had been speaking all along, waiting. Why should the free play of human symbols keep matching the deep grammar of reality? Either we are uncovering a mathematical order that was there, woven into things, or the coincidence is the most improbable in all of science. The equations fit the world too well to be only ours.
The Realm We Keep Bumping Into.
Ask a mathematician mid-proof whether they feel they're creating or finding, and most say finding—that the result was already true, waiting, and they merely came upon it, often against their wishes, often to their surprise. Prime numbers behave as they do whether or not anyone counts; the proof resists you like a landscape resists a hiker, with its own contours you did not put there. This felt objectivity is Platonism's great evidence: that there seems to be a realm of forms, timeless and mind-independent, that we explore rather than author. We do not vote on the primes. We bump into them.
Discovered Inside the Invented.
Perhaps the dichotomy is the error, and the truth is a nesting: we invent the container and discover the contents. The axioms are ours, chosen, contingent—a lens we ground and could have ground differently. But look through the lens and what you see is not up to you; the theorems are as hard and given as anything in nature, discovered within the very frame we freely built. So mathematics is neither found nor made but both, layered: a human window onto a non-human necessity, an invented question that receives, every time, an answer no one gets to write. We author the search and stumble on the found—and the deepest wonder is that a made thing should open onto something so utterly beyond our making.
Mathematics is both invented and discovered, at different layers: we invent the axioms, definitions, and notation, but the theorems that follow are discovered rather than chosen—you can decree the rules yet not their consequences, which is why proof feels like encountering something objective. Its astonishing effectiveness at describing the physical world points to a real but unexplained correspondence between mathematical structure and reality, and whether its objects exist independently stays open. So the invented framework and its discovered, non-negotiable consequences are two faces of one activity—we author the search and stumble on the found.