Consciousness as Constraint: A Critical Engagement with Integrated Information Theory
Zenodo (CERN European Organization for Nuclear Research) February 5, 2026 DOI: 10.5281/zenodo.18492680 (opens in new tab)
Study at a glance
AI-extracted from the abstract| Characteristics | Theoretical or philosophical paper Peer reviewed |
|---|---|
| Keywords | Consciousness Reflexivity Integrated information theory Existentialism Field mathematics Property philosophy Vagueness Epistemology Cognitive science Typology Convergence economics Transitive relation Information processing Artificial intelligence |
| Key points | Argues that IIT's structural tensions arise from treating consciousness as a bearer-property and that the Triaxial Existential Field (TEF) offers a constraint-based alternative with clinical, cross-domain, and AI architectural payoffs. |
Abstract
Abstract This paper applies the bearer/constraint audit methodology developed in “From Substance to Constraint” to Integrated Information Theory (IIT); the most rigorous contemporary attempt to treat consciousness as a quantifiable property (Φ) instantiated by physical systems with specific causal architecture. IIT’s bearer-commitments generate stable structural tensions: the exclusion problem, the grain problem, the small-Φ problem, and temporal instability. These persist across theory versions because they are downstream consequences of treating consciousness as a bearer-property rather than a constraint-role. The Triaxial Existential Field (TEF) retypes consciousness as integrated reflexive availability under constraint, articulated through three irreducible roles; Coherence, Reflexivity, and Participation. This retyping yields three differential payoffs: (1) a clinically testable typology of consciousness disorders classifying by axis-failure rather than scalar degree; (2) cross-domain convergence with constraint-based treatments of time and quantum measurement; and (3) implementable architectural criteria for artificial consciousness where IIT’s Φ-maximisation remains computationally intractable. Explicit falsification conditions are stated. TEF is positioned as a methodological bet with concrete stakes in clinical classification and AI architecture.