Gödel's Theorem Beyond Incompleteness: Closed Algorithmic Systems and Open Living Systems

Gödel's incompleteness theorem has often been used  to argue that the human mind or human cognition cannot be equivalent to a machine that implements closed algorithmic systems  because humans can recognize truths that closed algorithmic systems  cannot derive (Gödel, 1931/1986; Lucas, 1961; Penrose, 1989; Dietrich et al., 2021; Russell & Norvig, 2021). The most prominent arguments are Lucas-Penrose Arguments. The anti-mechanist arguments advanced by Lucas (1961) and later developed by Penrose (1989) have generated an extensive critical literature (see Meyer, 1997, for references). Most of these criticisms focus on Lucas's interpretation of Gödel's theorem, incompleteness in human cognition itself, and on the algorithmic properties of formal computational systems (Megill, 2012; Dietrich et al., 2021).

This paper argues that both the Lucas–Penrose interpretation and most of its subsequent criticisms misinterpret the philosophical significance of Gödel's theorem. The central issue is not the incompleteness of particular axiomatic systems or cognitive system but the distinction between closed algorithmic systems (implied by Gödel’s theorem) and open, self-organizing, normative systems (Kauffman, 2019). The concept of normativity is related to form of organization of these systems. Open, self-organizing systems are living systems (Longo et al., 2012; Kauffman, 2019). Living systems are thermodynamically open, exchanging matter and energy with their environment (Kauffman, 2019). Such systems maintain their organization under far-from-equilibrium conditions as non-ergodic, dissipative structures (Prigogine, 1967; Prigogine & Stengers, 1984; Schrödinger, 1992). Living systems must continuously regulate their interactions with the environment in ways that preserve their organization (Longo et al. , 2012). We use here 'normative' in the minimal biological sense that the system distinguishes states that contribute to or undermine its continued organization.

By a closed algorithmic system we mean a system whose operation is exhaustively specified by a finite formal description in the Gödel–Turing sense. A closed algorithmic system is one whose transitions, dynamic or statistical, are fully determined by a finite rule specification, regardless of the number of possible states or inputs. Although algorithms may receive novel inputs or adapt internal parameters, these changes remain constrained by the predefined formal rules (deterministic or statistical) that constitute the system. Their openness is therefore limited to interaction with the environment rather than to changes in their own constitutive organization. Every state transition of such a system is determined by its formal specification, even if predicting particular outcomes may be computationally intractable.

Consequently, Gödel's conclusion cannot be circumvented simply by proposing more powerful algorithms or richer formal systems. This is because the diagonalization at the heart of the proof applies to any recursively axiomatized extension of the system: recognizing the Gödel sentence as true requires stepping outside the system's own rule-governed transitions to a semantic standpoint the system cannot occupy from within; each axiomatic extension produces a new closed algorithmic system rather than a transition to a different organizational class of system.

We do not deny that open, self-organizing systems realize numerous algorithmic processes. Our claim is that these processes do not exhaust the organization of the system (Longo et al., 2012). The organization and continued evolution of open systems cannot be completely specified by a finite set of formal axioms and algorithmic transition rules.

The real issue is therefore not whether incompleteness can be overcome but whether these two organizational classes should be regarded as equivalent. Gödel's theorem should therefore be interpreted not merely as identifying a limitation of closed algorithmic systems but as revealing a more fundamental distinction between closed algorithmic systems and open, self-organizing, normative systems.

We do not claim that incompleteness alone entails the closed/open distinction, which is grounded independently in the thermodynamic and normative organization of living systems (Kauffman, 2019); rather, Gödel's theorem corroborates that distinction by showing that formal closure survives the most natural strategy for escaping it, namely axiomatic strengthening. Every algorithmic system, regardless of its sophistication, remains formally closed, whereas human cognition emerges within an open, self-organizing living system whose continued organization depends on normative regulation.

In this respect, Lucas-Penrose's intuition that Gödel points toward a fundamental difference between human and machine intelligence is preserved, although the difference is interpreted here as arising from the ontological distinction between closed algorithmic systems and open, self-organizing, normative systems rather than from incompleteness alone.

References

Dietrich, E., Fields, C., Sullins, J. P., Van Heuveln, B., & Zebrowski, R. (2021). Great philosophical objections to artificial intelligence: The history and legacy of the AI wars. Bloomsbury Academic.

Gödel, K. (1931/1986). On formally undecidable propositions of Principia Mathematica and related systems. In S. Feferman et al. (Eds.), Kurt Gödel: Collected works (Vol. 1, pp. 144–195). Oxford University Press. (Original work published 1931)

Kauffman, S. A. (2019). A world beyond physics: The emergence and evolution of life. Oxford University Press.

Longo, G., Montévil, M., & Kauffman, S. (2012). No entailing laws, but enablement in the evolution of the biosphere. In Proceedings of the 14th annual conference companion on Genetic and evolutionary computation (pp. 1379-1392). https://arxiv.org/abs/1201.2069.

Lucas, J. R. (1961). Minds, Machines and Godel. Philosophy. 36 (April–July): 112–127. doi:10.1017/s0031819100057983.

Megill, J. (2012). The Lucas–Penrose argument about Gödel's theorem. In J. Fieser & B. Dowden (Eds.), Internet Encyclopedia of Philosophy. https://iep.utm.edu/lp-argue/.

Meyer, J. (1997). J.R. Lucas: References for Criticisms of the Gödelian Argument. https://jamesrmeyer.com/ffgit/lucas-references-godel

Penrose, R. (1989). The emperor's new mind: Concerning computers, minds, and the laws of physics. Oxford University Press.

Prigogine, I. (1967). Introduction to thermodynamics of irreversible processes (3rd ed.). Interscience Publishers.

Prigogine, I., & Stengers, I. (1984). Order out of chaos: Man's new dialogue with nature. Bantam Books.

Russell, S., & Norvig, P. (2021). Artificial intelligence: A modern approach (4th ed.). Pearson.

Schrödinger, E. (1992). What is life? The physical aspect of the living cell; with Mind and Matter and autobiographical sketches. Cambridge University Press