Some Mathematicians Are Not Turing Machines

TL;DR

The paper proves that a faultless mathematician who recognizes certain proofs cannot be fully replaced by any Turing machine, under specific conditions, highlighting fundamental limits of computational simulation of mathematical reasoning.

Contribution

It demonstrates that certain faultless mathematicians with proof recognition abilities cannot be simulated by Turing machines, under specified conditions.

Findings

Faultless mathematicians cannot be fully replaced by Turing machines.

Recognition of certain proofs imposes limits on computational simulation.

The proof establishes fundamental boundaries of algorithmic reasoning in mathematics.

Abstract

A certain mathematician M, considering some hypothesis H, conclusion C and text P, can arrive at one of the following judgments: (1) P does not convince M of the fact that since H, it follows that C; (2) P is the proof that since H, it follows that C (judgment of the type "Proved"). Is it possible to replace such a mathematician with an arbitrary Turing machine? The paper provides a proof that the answer to the question is negative under the two following conditions: (1) M is faultless, namely his judgment "Proved" always implies that since H, it actually follows that C; (2) M recognizes a certain P' as the correct proof of the fact that for certain H' and C', if H', then C' (where P', H', and C' are stated in the paper).