The Rule-Transition Field: A Receiver-First Architecture for Cognitive Change Beyond Fixed Formal Systems

Accounts of cognition commonly begin inside an already stabilized formal world: objects are recognized, evidence has an admissible form, a task and objective are declared, and rules can transform inputs into judgments or actions. This paper argues that such accounts leave a distinct scientific object under-described. Cognition also changes the formal worlds through which later cognition becomes possible. Rules lose authority, receivers become sensitive to previously excluded evidence, ontologies and objectives are revised, meanings are transported or broken, and consequences alter what may govern next. We introduce the Rule-Transition Field as a receiver-first architecture for representing this interval. The proposal is motivated by Roger Penrose's refusal to identify mathematical understanding with the blind continuation of one fixed formal procedure, while rejecting any direct inference from incompleteness to physical noncomputability, objective reduction, microtubular realization, phenomenal consciousness, or wisdom. Penrose is treated as a central precursor rather than an exclusive target. Clark and Chalmers, enactivist and predictive traditions, cognitive-flexibility neuroscience, and meta-learning provide non-equivalent partial views of the same wider burden. The paper defines a Formal World Constitution Fₜ = ⟨Rₜ, Ωₜ, Vₜ, Eₜ, Gₜ, Bₜ⟩; a Rule-Transition Operator Rₜ₊₁ = ΠΦ(Rₜ, Eₜ, Kₜ, Mₜ, Bₜ, Sₜ, Qₜ); a Rule-Transition Depth Ladder separating parameter, policy, representation, ontology, objective, and authority revision; and a Provisional Closure Constitution Cₚ = ⟨scope, duration, evidence, authority, return⟩. It further introduces Receiver Nonclosure, Rule-Authority Separation, Reciprocal Rule Qualification, External Consequence Authority, Counterpoint Abundance, and Field Doppelgängers. Generative AI is positioned as a new instrument for constructing and comparing formal worlds at scale, not as an authority capable of validating the worlds it generates. The benchmark constitution tests whether Field variables add predictive or interventional gain beyond task switching, model selection, meta-learning, or fixed hierarchical control. The bounded thesis is not that mind escapes computation or that every transition is conscious. It is that a theory of cognition remains incomplete when it explains operations within a formal world while omitting the structured Field through which that world becomes contestable, revisable, provisionally replaceable, and answerable to consequence.