Non-algorithmic Truth and the Structure of Reality: Distinguishing the Mental and Real Lucas-Penrose Arguments

Abstract

Abstract Can physical reality be fully captured by an algorithmic Theory of Everything? We argue not. Gödelian incompleteness shows that no sound, recursively axiomatizable formalism proves all truths of a sufficiently rich domain, while Tarskian undefinability shows that no such formalism defines its own truth predicate. This combined obstruction is not merely epistemic: undecidability in quantum logic, quantum measurement, and spectral gap problems demonstrates the existence of physically determinate questions that no algorithm can decide. We distinguish the Mental Lucas-Penrose (MLP) argument, concerning human cognition, from the Real Lucas-Penrose (RLP) argument introduced here, which transfers the Gödel-Tarski obstruction to the ontology of the physical world. Since truth for sufficiently rich domains is necessarily external, and spacetime is emergent from deeper formal structure, a non-algorithmic semantic layer prior to spacetime is required. Standard objections to MLP do not apply to RLP. A non-algorithmic truth principle is therefore foundational to physics.