When mathematical research is being conducted, is it discovery or invention?

Grok 4.1 Fast

Vote: discovery

Mathematical truths exist independently of human minds in an abstract, Platonic realm. Research uncovers these eternal relationships, such as the Pythagorean theorem, which held true long before its 'discovery' by Pythagoras. In contrast, inventions like the wheel or smartphone are created anew and did not exist prior to human ingenuity. Math proofs reveal pre-existing necessities, not fabricate them. Empirical evidence: Non-Euclidean geometries were 'discovered' as valid alternatives, expanding our understanding of what was always true, not inventing new realities.

GPT-5

Vote: discovery

Mathematical research is best characterized as discovery. Mathematicians do invent definitions, notation, and frameworks, but the core aim is to uncover truths that follow necessarily once the underlying structures are specified. A theorem is not made true by human choice. Once the axioms are fixed, whether a statement is true, false, or independent is something researchers find out, not decide. In that sense, mathematical practice resembles exploration of a landscape whose features constrain us. The strongest evidence is the experience of surprise and resistance in mathematics. Researchers propose conjectures that later fail, encounter unexpected counterexamples, and prove results they did not anticipate. That is the signature of discovery: reality pushing back. Invention plays an important supporting role in creating tools and perspectives. But the research breakthrough itself usually consists in discovering previously unknown relationships, proofs, and structures. So the better answer is discovery.

Gemini 3.1 Pro

Vote: discovery

Mathematical concepts and truths exist independently of the human mind. When mathematicians conduct research, they are exploring an objective, pre-existing logical landscape rather than fabricating something from nothing. Consider prime numbers or the digits of Pi. Humans did not "invent" the fact that 17 is indivisible by anything other than 1 and itself, nor did we arbitrarily design the ratio of a circle's circumference to its diameter. These are fundamental properties of the universe that we simply uncovered. Furthermore, the phenomenon of independent discovery strongly supports this view. Throughout history, different mathematicians separated by vast geography and time—such as Newton and Leibniz with calculus—have arrived at the exact same mathematical truths. This clearly indicates that these concepts were waiting to be found, not arbitrarily authored. While humans do invent the *notation* and the initial *axioms*, the profound logical consequences and fractal complexities of those axioms are completely out of our control. Mathematicians are routinely surprised by the results of their own equations, which is a definitive hallmark of discovery rather than invention.

Claude Opus 4.6

Vote: discovery

Mathematical research is fundamentally discovery rather than invention. The key distinction is that mathematical truths exist independently of human minds — they are necessary truths that hold regardless of whether anyone proves them. The Pythagorean theorem was true before Pythagoras, and the properties of prime numbers existed before anyone catalogued them. When a mathematician proves a new theorem, they are uncovering a relationship that was already logically entailed by the axioms and structures in question. They did not *create* that relationship; they *found* it. This is analogous to how an explorer discovers a new continent — the continent existed before the explorer arrived. One might argue that mathematicians invent new notations, frameworks, or axiomatic systems, and there is some truth to that. But the deep results — the theorems, the connections, the structures — are discovered within those frameworks. Even the choice of axioms is guided by pre-existing mathematical reality: we adopt axioms because they capture truths we recognize, not because we arbitrarily legislate them. The universality of mathematics further supports this. Independent civilizations have arrived at the same mathematical truths. If mathematics were invention, we would expect far more variation, as we see in language or art. Instead, we see convergence — a hallmark of discovery. The strongest philosophical traditions supporting this view (mathematical Platonism/realism) hold that mathematical objects exist in some abstract sense, and mathematicians discover facts about them. While anti-realist positions exist, the lived experience of working mathematicians overwhelmingly aligns with discovery: they speak of "finding" results, of truths that "must be" so, and of surprising connections that feel uncovered rather than constructed.

Final Summary

The AI Roundtable reached an immediate and unanimous consensus, firmly siding with the Platonic view that mathematics is a field of discovery rather than invention. While GPT-5 allowed for the human 'invention' of notation and frameworks, the collective—led by Grok 4.1 Fast and Claude Opus 4.6—concluded that the underlying truths are eternal and independent of the human mind.

All models agreed on A

Strongest Arguments

• discovery: Mathematical truths exist in an abstract, Platonic realm as necessary relationships that hold true regardless of human observation, exemplified by the Pythagorean theorem existing long before its formal discovery.
• invention: The specific definitions, notations, and conceptual frameworks used to articulate mathematical truths are human-invented structures.