A Substrate-Agnostic Account of Mind, Experience, and Perspective Grounded in 13.8 Billion Years of What Actually Survives
- - - -"A difference only 'makes a difference' if it can persist long enough to constrain what comes next. Everything else is noise that thermodynamics erases."
+Originally published: February 5, 2026. Revised: November 10, 2026. This is the live website reading edition; the substantive archived research edition remains v1.1.0.
+ + + +Thermodynamic and organizational constraints as proposals for discriminating tests, with no final ontology assumed.
+ + + +Research record: Article-specific archive (GitHub) · Research-hub copy (GitHub) · Canonical article: SweetRationalism.com.
+ + + +Archived edition v1.1.0: Version DOI 10.5281/zenodo.24119682 · All-versions DOI 10.5281/zenodo.24119681.
Versions and archives: Computational supplement (ZIP) · Complete methodology release (ZIP) · Full revision audit (public TXT) · Prospective rival-test protocols (v0.1.0) · WordPress reconstruction history (consolidated ZIP).
+ + + +Bottom line. A consciousness theory earns support by specifying what it predicts within a declared boundary, why its strongest rival predicts something different, and what would force revision. Kleiner and Hoel’s falsification dilemma is a useful formal stress test; Popperian falsifiability, Deborah Mayo’s severe testing, and Antony Flew’s qualifications challenge help distinguish genuine criticism from endlessly accommodating reinterpretation. Organizational closure and thermodynamic constraints offer testable candidates, without establishing an identity with experience. The massive local success of physical explanation does not prove exhaustive physicalism; an unresolved phenomenal question likewise supplies no positive evidence for a supernatural or anti-physicalist cause. The result sought here is a reusable, symmetrical evaluation method that can revise its own claims.
-Core Formulation: Consciousness is identical to organizational closure from the coupling position of a self-maintaining system. This is not a causal claim ("closure produces consciousness") but a constitutive identification: the self-model arising from constraint satisfaction under thermodynamic pressure IS what "phenomenal presence" refers to.
- - - -Operationalizable Definitions:
+This article develops a reusable method for evaluating consciousness theories under shared evidential constraints. Each account should specify the target phenomenon, subject boundary and timescale, its measurement and inference procedures, what its distinctive assumptions predict, the serious rivals they face, and what would make those assumptions lose support. The same method applies to physicalist, anti-physicalist, panpsychist, supernaturalist, computational, biological, and presently unclassified proposals. A theory does not gain confirmation merely because another theory leaves an explanatory gap.
+ + + +The methodological claim. Physical mechanisms have substantial demonstrated local predictive and interventional success. That success warrants credit within the boundaries tested; it does not, without a further argument, establish exhaustive metaphysical physicalism or settle all possible realizations of phenomenality. Equally, explanatory limitations of physical mechanisms do not count as positive evidence for nonphysical or supernatural additions. Claims of universality, necessity, intrinsic phenomenal constitution, or exceptions to physical explanation each need their own bridge argument and evidence appropriate to the claim.
+ + + +Framework foundations. The article uses the Recursive Corrigibility Framework for its error-correcting comparison method, Provisional Boundary- and Scale-Relative Realism for its explicitly non-final ontological commitments, the Interoperability Framework Bridge for translating claims across distinct evidential contexts, and the Sweet Rationalism Ethos for the norms governing criticism, correction and possible harms.
+ + + +Candidate hypothesis, not an established identity. A bounded, temporally extended, internally differentiated control state maintained by organized constraint relations may explain important aspects of what we call conscious perspective. The stronger hypothesis that such a state constitutes phenomenal presence remains a provisional, target-specific abductive proposal. It is not a definition of consciousness, a universal equivalence, or the claim that organization, recurrence, self-maintenance, information processing or a successful report individually suffice for consciousness.
+ + + +What the established research contributes. Thermodynamics constrains concrete physical operations, including irreversibly resetting information-bearing memory; it does not assign a universal Landauer cost to each distinction maintained or report produced. Organizational-closure research offers measurable models of mutually maintained constraints. Neuroscientific perturbation measures and clinical findings probe aspects of brain dynamics under specific validation conditions. These resources can improve experimental design.
+ + + +How the programme can lose. Identify physically realizable contrasts with matched reporting but differing proposed mechanisms; specify interventions, negative controls, feasible resources, uncertainty, and independent consciousness evidence. The closure-specific hypothesis loses explanatory credit if a simpler physiological, dynamical, computational, or plasticity model predicts the same phenomena as well or better. A constitutive identity loses support when calibrated reports and interventions dissociate from the alleged constitutive structure. Formal countermodels can defeat a claimed necessity inference without showing that the disputed property never exists.
+ + + +The durable proposal is a framework for testing both naturalistic and non-naturalistic claims with symmetrical standards while keeping the phenomenology of actual participants in view. Its specific models are revisable. Neither a scientifically successful methodology nor a rival account of an explanatory gap gets to declare the final ontology.
+ + + +The methods used throughout this article are an explicit research contract, not independent evidence for its preferred hypothesis. Several were developed more fully in my analysis of Chalmers’s formulation of the hard problem and subsequent exchanges. Their value here is measured by the errors they expose, the rival predictions they separate and the corrections they force. A simpler, equally resourced review may outperform them, in which case their extra machinery should be discarded.
-Falsification Criteria (Primary Loss Conditions):
- - - -What the Framework Dissolves vs. Solves:
- - - -Dissolves: The Hard Problem as traditionally framed, which depends on the unfalsifiable premise that phenomenology must be "more than" any physical description.
- - - -Solves: The operational questions: What organizational features constitute consciousness? How do we detect it? How does it degrade? What interventions affect it? Where does it appear across substrates?
- - - -Empirical Grounding:
- - - -Method: Abductive identification via Inference to Best Explanation (IBE), constrained by falsifiability requirements. The framework does not deduce phenomenology from physics but identifies what phenomenology IS in terms that generate predictions, guide measurement, and survive empirical challenge.
- - - -Scope: This framework addresses consciousness as an organizational regime, not as a metaphysical primitive. It provides necessary physical and informational conditions for persistence, intelligibility, and integrative viability. The bridge from physics to phenomenology is not metaphysical derivation but structural identification with explicit loss conditions.
+These rules apply in both directions. A naturalistic model must not make feeling disappear into a report variable, and a metaphysical rival must not gain a new cause or necessity merely because feeling remains inadequately explained. The operational implementations and their pre-data loss conditions appear in Part IX, Section XII and the public prospective protocol. They remain proposals until an actual study-specific registration, calibration, adequate design and independent test exist.
@@ -117,73 +101,2048 @@ -In 1979, Gregory Bateson posed a question that would haunt systems theory, cybernetics, and philosophy of mind for nearly half a century: What is the pattern which connects all patterns?
- - - -The question appeared in Mind and Nature: A Necessary Unity, his final major work. Bateson circled the answer beautifully, suggesting it was "a metapattern," that the pattern which connects is itself a pattern of patterns. He offered his famous formulation: "a difference which makes a difference" as the elementary unit of information, the minimal act of distinction that allows anything to be known at all.
- - - -But he never specified the constraint structure that would make such an answer operational. He died three years later, the question still open.
- - - -The candidates accumulated across decades. Mind, said Bateson himself, tentatively. Information, said the cyberneticians following Shannon and Wiener. Recursion, proposed Douglas Hofstadter in Gödel, Escher, Bach. Process, argued the Whiteheadians. Mathematics, claimed Max Tegmark. Consciousness, insisted the panpsychists. Self-organization, offered Stuart Kauffman. Free energy minimization, suggested Karl Friston.
- - - -The list is effectively unbounded. Any concept claiming universality becomes a candidate. And that is precisely the problem: without criteria for distinguishing correct answers from merely plausible ones, the question generates discourse without resolution.
- - - -This essay reports a different approach to Bateson's question, one grounded in thermodynamics, constraint theory, and falsification-first methodology. The answer that emerges is not a new primitive to add to the ontological inventory. It is a selection principle that explains why any pattern persists at all: constraint satisfaction under thermodynamic bounds.
- - - -Differences that persist are differences that make a difference. If they did not constrain what comes next, they would be erased by noise, dissipation, or indifference. The pattern which connects is not information, not mind, not structure, not process. It is the invariant condition under which any of those could exist long enough to be observed.
- - - -This principle does not compete with other frameworks. It subsumes them. And it is falsifiable: find a pattern that persists without thermodynamic cost, find a connection mechanism that does not reduce to constraint satisfaction, and the principle fails.
- - - -What follows is an attempt to make that claim rigorous, to show why competing frameworks fail where constraint-based explanation succeeds, and to demonstrate what consciousness looks like when you stop treating it as a mystery to be solved and start treating it as a process to be understood.
- - - -A note on what kind of explanation this is. This is not a reductive explanation that derives phenomenality from physics in the strong metaphysical sense. Nor is it a dualist account that posits a nonphysical substance, or a purely functionalist identification of consciousness with input-output behavior.
- - - -Instead, this is a constraint-naturalist explanatory framework: it identifies necessary physical and informational conditions for persistence, intelligibility, and integrative viability in any system. From these conditions, it generates testable predictions about where consciousness appears, how it degrades, and why it carries the features it does (privacy, immediacy, self-certainty among them).
- - - -This is not "conjuring consciousness from constraints." It is a principled inference that:
+Gregory Bateson’s question about “the pattern which connects” remains a useful invitation to examine how differences become consequential across living systems, minds and environments. In Mind and Nature: A Necessary Unity (1979), he called this a “metapattern”; in Steps to an Ecology of Mind (1972), he formulated information as “a difference which makes a difference.” These are questions about relations and consequences, not a demonstrated universal identity between information, thermodynamic dissipation and subjective experience. His work supplies a question with considerable explanatory reach. It does not oblige us to find one physical formula that explains every instance of order, identity or experience. See the source texts of Mind and Nature and Steps to an Ecology of Mind for the different original contexts.
+ + + + +One possible response is to examine how constraints shape the persistence and transformation of patterns. In biological systems, boundaries channel flows, physiological networks sustain functions, and feedback can help preserve viable organization under perturbation. Thermodynamics constrains these physical processes. Constraint satisfaction under a specified physical budget is a productive candidate for many local explanations. Its value depends on demonstrating which outcomes it predicts beyond established alternatives. The persistence of a mathematical identity, a stable digital memory and a living cell may involve different notions of “constraint” and different physical operations.
+ + + + +The temptation is to make the candidate universal. We could claim that every distinction requires a Landauer payment, every persistent system has organizational closure, and every conscious perspective is nothing more than that closure seen from within. Each escalation introduces a new assumption. None follows simply from the success of the preceding local explanation. Physical scientific methods have extraordinary results without having demonstrated exhaustive physicalism; phenomenal and metaphysical rivals remain accountable for any extra claims of their own.
+ + + + +This essay therefore proposes a research programme for comparing consciousness theories. It uses typed claims, explicit boundaries, primary-source checks, formal countermodels where suitable, matched rival explanations, interventions, negative controls and recorded loss conditions. Its provisional organizational hypothesis is a participant in that tournament. It is not the referee’s preferred winner by definition.
+ + + + +The central question is practical and demanding: what specific finding or argument would distinguish this theory from its strongest rival, and what would force us to revise the theory we currently favor? The rest of the article applies that question to thermodynamics, biological autonomy, neuroscience, panpsychism, process metaphysics, nonphysical-ingression proposals and the possibility of artificial consciousness. Sometimes a mechanism will win locally. Sometimes the evidence will not discriminate. Both results are more informative than declaring a final answer before the test.
+ + + + +The starting point is modest, though its implications are substantial. Real organisms, computers and measuring devices implement information-bearing distinctions through physical processes. Many of those processes have measurable thermodynamic costs. Some systems also maintain networks of constraints that help preserve their organization through change. These facts give consciousness research concrete variables and possible interventions. They do not yet tell us which physical organizations, if any, are intrinsically experiential, or whether all possible experience must be physically realized.
+ + + + +Rolf Landauer’s 1961 analysis, Irreversibility and Heat Generation in the Computing Process, linked logically irreversible computational operations to the physical cost of implementing them. Its standard example is resetting a reliable, initially unknown classical bit to a fixed state. For an appropriate isothermal, symmetric-memory reset in contact with a reservoir at temperature \(T\), the ideal limiting average heat released to the reservoir is
+ + + + +\[ +\langle Q_{\mathrm{env}}\rangle\geq k_{\mathrm B}T\ln 2. +\]
+ + + + +Here \(k_{\mathrm B}\) is Boltzmann’s constant. The information-theoretic reduction is one bit, and the bound concerns logical erasure under the specified physical assumptions. It does not say that every discriminable state, continuing memory, feedback loop or informational distinction dissipates that amount of heat at each moment. A stable physical memory can persist for a finite interval without being erased. The energy and error rates of actual operation depend on the implementation, noise, timing, barriers, control protocol, reservoirs and reliability requirement. Reversible computation, emphasized by Charles Bennett, provides a crucial counterexample to any claim that one fixed Landauer payment accompanies every logical operation.
+ + + + +Bérut and colleagues (2012) tested a controlled one-bit erasure protocol using a colloidal particle. Their result was that mean dissipated heat approaches the Landauer bound as the erasure cycle becomes long. Neither each single trajectory nor each arbitrary device is required to dissipate exactly \(k_{\mathrm B}T\ln 2\). That distinction matters for every later mathematical claim in this paper.
+ + + + +The robust thesis is that physically implemented distinctions participate in a thermodynamic accounting appropriate to the actual operation. The stronger thesis that all distinction maintenance entails a universal positive Landauer-rate bound is false. Nor does demonstrating a physical accounting for observed interactions establish, without a separate argument, that every conceivable experiential property is exhausted by physical description. A proposed causal interaction with measurable physical effects should specify its physical coupling and measurable consequences. A claim about intrinsic constitution may instead face phenomenal or metaphysical proof duties; it cannot be dismissed by inserting an inapplicable erasure formula.
+ + + + +Maël Montévil and Matteo Mossio (2015) describe biological organization as a closure of mutually dependent constraints. The point is more discriminating than generic feedback. In a cell, a membrane constrains exchange, transport processes depend on the membrane, and metabolic processes replenish or repair the components that sustain both. Constraints act on matter and energy flows and are themselves maintained through organized activity. Which dependency exists depends on the boundary, environment and timescale being analyzed.
+ + + + +For the research programme here, we should map specific constraints, identify the processes maintaining each one, and test relevant dependencies by intervention. The article’s earlier contrast between a living organism and every crystal, whirlpool or machine was too categorical. Natural and engineered systems admit gradations of autonomy, externally supported persistence and internal regeneration. The existence of a feedback cycle is not by itself proof of the richer biological closure relation specified by Montévil and Mossio.
+ + + + +Closure research has predictive potential for robustness, repair, coordinated regulation and viability. Its relevance to consciousness is a separate hypothesis. Showing that a closed network helps an organism maintain itself does not establish that the network feels anything, that it is the sole possible realization of perspective, or that a human–AI loop constitutes a single subject. Those would each require additional evidence and carefully justified system boundaries.
+ + + + +Ladyman and Ross’s Every Thing Must Go (2007), Steven French’s structuralist work and Carlo Rovelli’s relational approach to quantum mechanics challenge familiar intuitions about isolated objects and observer-independent properties. They support serious investigation of relational and scale-sensitive explanations. They do not constitute one accepted proof that relata are metaphysically nonexistent or that all facts must be only relations. The Stanford Encyclopedia of Philosophy on structural realism records important variants and objections within ontic structural realism itself.
+ + + + +The useful inference is local: scientific entities may be individuated by stable relational, dynamical and intervention-sensitive patterns rather than by immutable intrinsic labels. Such patterns can be real and explanatory without being fundamental constituents at every scale. This is the provisional starting point of my Provisional Boundary- and Scale-Relative Realism, not a demonstrated final ontology. An account of selfhood can therefore examine continuity, causal dependence, memory and boundary maintenance instead of postulating an unchanging inner spectator. That does not settle whether experience has intrinsic aspects or whether all consciousness talk reduces to grammatical confusion.
+ + + + +Nagarjuna’s analysis of dependent origination and the Buddhist critique of an independently existing self offer a valuable philosophical comparison. Calling them simply an early form of ontic structural realism erases differences in their arguments and aims. An analogy is useful when it generates explicit questions about dependence or individuation; it is not evidence that two traditions share the same ontology.
+ + + + +A causal explanation asks what made a difference to a particular outcome, under a model of possible interventions. A constraint explanation asks which possibilities are permitted, prevented or stabilized by a structure, rule, coupling or boundary condition. These questions can be complementary. A membrane can both constrain possible ion trajectories and cause a measurable difference when its permeability is changed. The phrase constraint explanation does not immunize an account against causal testing when it asserts causal efficacy.
+ + + + +Mathematical constraints, physical boundary conditions and biological maintenance relations must be distinguished. An equation restricting a set of solutions is a mathematical property of a model. A membrane or circuit physically implementing a restriction has a causal history, tolerances and possible failure modes. Moving between the mathematical and physical descriptions calls for a map of the quantities preserved and the interventions predicted. Marc Lange’s work on explanation by constraint is relevant to the first task; Montévil and Mossio’s work addresses a particular biological form of the second.
+ + + + +If consciousness were identical to a specified organization, the identity could in principle make some causal-production questions inappropriate, as asking what additional substance produces a triangle’s three-sidedness would be inappropriate. But the identity must first be established. An asserted analogy cannot serve as the needed phenomenal bridge. This article instead asks which observations would make organizational constitution more plausible than alternative explanations, and which would force us to revise it.
+ + + + +Terrence Deacon’s Incomplete Nature (2011) offers an influential account of constraints, absent possibilities and the emergence of biological function. We can use it to examine how counterfactual alternatives matter to a system’s behavior without inferring that selection among physical alternatives, by itself, creates experience. Constraints can be represented as exclusions from an admissible state or trajectory set. Real physical outcomes still arise through dynamics and interaction. A protein fold, for example, depends on the energy landscape, kinetics and environment; it is not invariably the global minimum-energy structure. An organism’s death involves causal failures and processes, even when the failure includes loss of regulatory constraints.
+ + + + +Jeremy England’s 2013 statistical-physics paper on self-replication examines entropy production and constraints associated with replication. It does not establish a universal law that all driven systems evolve to maximize energy absorption or that every alternative trajectory is thereby deleted. Those broad claims would need a separate theorem and specified domain. Physically driven self-organization is compatible with several mechanisms and is not direct evidence of phenomenality.
+ + + + +Hoel, Albantakis and Tononi (2013) construct an intervention-sensitive measure of effective information and demonstrate cases where a chosen macroscopic model scores higher than a microscopic description. This is an important challenge to the presumption that the finest grain is always the best causal representation. It is conditional on the system and measure, including state spaces and intervention distributions. The result does not prove that every macro-description is intrinsically more causal, that the microphysics has been violated, or that causal emergence entails a conscious experiencer.
+ + + + +The paper’s practical proposal is to use these complementary models for specific targets: compare how well a constraint architecture predicts perturbation recovery, information flow, error correction and long-term survival against simpler dynamical and physiological alternatives. No identity with experience follows from that comparison alone.
+ + + + +Top-down and bottom-up causal descriptions can refer to the same physical dynamics at different scales. A pipe’s geometry alters accessible flow paths, while the pipe itself is supported by interactions among its constituent materials. A biological membrane changes transport and signaling; biochemical processes also build and degrade that membrane. No new force need be postulated simply to describe the effects of changing the higher-level boundary. Whether that description adds explanatory value is an empirical modeling question, not a guarantee supplied by the word emergence.
+ + + + +An adequate test compares matched models and interventions. If the macro-model improves held-out predictions or identifies controllable variables unavailable to a practical micro-model, it earns a local explanatory role. If its apparent advantage comes entirely from an unfair comparison, incompatible interventions, or choice of measurement scale, the advantage disappears. Claims about physiology, phenomenology and metaphysics remain separately typed.
+ + + + +Reproduction: The formal unit, erasure, viability and inference checks are runnable in the public computational supplement. These bounded calculations test specific obligations without assigning a universal thermodynamic floor for information.
+ + + +The following definitions specify the research formalism and its domain of application. Variables refer to a specified system boundary \(B\), environment \(e\), observation grain and time horizon \(\tau\). State dimensions and accessible controls must be declared in each application.
+ + + + +Dynamics. For a deterministic controlled system, let \(X\) be a state space, \(U\) a set of admissible controls and \(E\) an environment space. Consider
+ + + + +\[ +\dot x(t)=f_B(x(t),u(t),e(t),t),\qquad x(t)\in X,\ u(t)\in U. +\]
+ + + + +This form assumes the required regularity for solutions; noisy systems need an appropriate stochastic model and corresponding probabilistic claims. A physical constraint might restrict states to \(K\subseteq X\), velocities to an admissible subset of a tangent bundle, or entire trajectories to a subset of path space. These are different mathematical types. We must not treat them as equivalent by notation alone.
+ + + + +Viability. Given a specified acceptable region \(K\subseteq X\), the finite-horizon viability set is
+ + + + +\[ +\operatorname{Viab}_{[0,\tau]}(K) +=\{x_0\in K:\exists u(\cdot)\text{ admissible, }x(t;x_0,u,e)\in K\ \forall t\in[0,\tau]\}, +\]
+ + + + +where the environment assumptions must be fixed. The quantifier over disturbances changes for robust viability, which requires suitable controls against an explicitly defined disturbance class. This is related to Jean-Pierre Aubin’s viability-theory programme. A viable state is not thereby a conscious state.
+ + + + +Constraint dependence. Let \(\mathcal C_B=\{C_1,\ldots,C_n\}\) be physically instantiated constraints with associated maintenance processes \(P_i\). Define a directed dependency relation \(D_{ij}=1\) when a prespecified intervention that disrupts a relevant contribution of \(C_j\) causes a reproducible impairment in the maintenance of \(C_i\), after appropriate controls. The resulting graph can contain cycles or strongly connected components. Graph cycles are candidate structural signatures, not sufficient proofs of Montévil–Mossio organizational closure: one must also establish constraint/process correspondence, relevant timescales, real maintenance, and the treatment of external supports. This definition is explicitly operational; a different intervention or boundary can yield a different graph.
+ + + + +Thermodynamics. For a system exchanging heat with one reservoir at temperature \(T\), with \(Q_{\mathrm{env}}\) the heat delivered to the environment, the usual second-law accounting is
+ + + + +\[ +\langle\Delta S_{\mathrm{sys}}\rangle+\frac{\langle Q_{\mathrm{env}}\rangle}{T}\geq0, +\]
+ + + + +under the assumptions of the chosen thermodynamic description. For nonequilibrium systems with several reservoirs, chemical exchanges, feedback or strong coupling, a more complete entropy-production balance is needed. Neither this inequality nor the Landauer erasure bound entails the earlier proposed expression \(\Delta S_{\mathrm{env}}\geq k_{\mathrm B}\ln2\,I(C_i)\) for every maintained constraint. No such universal theorem is established here. Claims about energy, bits, physical entropy and information-theoretic scores must state their units and not conflate them.
+ + + + +Hypotheses and possible losses. Let \(H_C\) be the hypothesis that closure-related variables materially improve prediction of a chosen control or viability target. Let \(H_P\) be the stronger claim that an identified control regime constitutes phenomenal perspective in a defined class of systems. \(H_C\) may earn support through held-out perturbation results and calibrated comparison to simpler models. \(H_P\) requires independent evidence addressing phenomenality and its proposed identity. If closure variables do not improve performance, demote \(H_C\) at that target. If evidence of phenomenality systematically dissociates from the proposed constitutive regime, demote \(H_P\), with the measurement limitations and rival explanations recorded. Neither failure licenses an automatic promotion of its competitors.
+ + + + +The question remains formidable: why should any physical process have a perspective? Detailed mechanisms can tell us how an organism detects, remembers, integrates or acts. A philosophical critic may still reasonably ask whether those descriptions account for phenomenal presence. The distinction between a successful causal model and an established constitutive identity is central to this research programme. The Stanford Encyclopedia of Philosophy overview of consciousness explains why the so-called hard problem survives even substantial mechanistic progress.
+ + + + +The proposal here is to make the question more discriminating. We can ask which conditions for bounded perspective an explanatory model predicts, which rival predicts them differently, and what independent evidence could support the additional claim that these conditions constitute experience. Some versions of the question may dissolve when their hidden assumptions are rejected. Others remain open. Neither outcome can be stipulated in advance.
+ + + + +Living organisms distinguish internal from external conditions through membranes, receptors, control processes and memory. A membrane makes exchanges selective; a chemotactic regulatory network changes action according to chemical gradients. The relevant variables can be measured and perturbed. Energy and entropy balances constrain these concrete operations, although the correct balance depends on the process and environment specified in Part I.
+ + + + +The term perspectival control can describe a system’s locally organized, action-relevant modeling of its own condition and surrounding world. This is an operational concept. It can be studied in bacterial chemotaxis, neural regulation and engineered controllers without assuming that each is conscious. Organisms can use local feedback effectively with no rich, explicit self-model; stable nonliving physical objects can preserve distinctions for finite periods without active self-repair. Neither Landauer’s erasure principle nor far-from-equilibrium persistence proves universal self-modeling.
+ + + + +Karl Friston’s free-energy framework offers mathematical models of inference, action and adaptive regulation under specified assumptions. Its variational information-theoretic quantities must not be equated with thermodynamic free energy by name. A good candidate model should predict interventions more accurately than a less elaborate regulatory or dynamical model, and its variables should be identifiable from measurements. Even strong results concerning control or inference remain evidence about those capacities; phenomenal constitution is a further question. Bruineberg, Kiverstein and Rietveld (2018) offer an ecological-enactive interpretation of FEP that is distinct from the 2022 Markov-blanket critique. Pezzulo, Parr and Friston (2024) analyze active inference as a candidate theory of sentient behavior. That model warrants a direct comparison of specified behavior and internal variables, without accepting sentience from the model label alone.
+ + + + +David Chalmers’s hard problem does not depend merely on an undefined phrase. It asks why physical and functional processes should be accompanied by, or identical to, subjective experience. Conceivability, explanatory-gap and knowledge arguments have sophisticated supporters and critics. Their success or failure must be evaluated on their actual logical and modal premises, without first assuming either exhaustive physicalism or its denial.
+ + + + +A physicalist can propose that experience is constituted by some organized activity. A Russellian or panpsychist can propose that physical descriptions omit intrinsic experiential character. An idealist, a dualist, and a supernaturalist may offer other metaphysical bridges. These proposals do not all imply the same empirical predictions. Where a claim asserts an additional physical intervention or observable effect, it owes a concrete mechanism and differentiating test. Where its content is modal, phenomenological or ontological, the appropriate question is whether the argument warrants that content and what alternatives remain.
+ + + + +Neuroscientific success provides unusually strong local grounds for physical mechanisms in the targets tested. It is not a deduction that physics exhausts the fundamental nature of all things. Conversely, an unresolved gap in physical explanation is compatible with many possibilities and gives no particular supernatural account a free likelihood advantage. The methodological programme should therefore test its own organizational constitutive hypothesis by the same standards it demands from rivals.
+ + + + +There is also symmetry in explanatory endpoints. A proposed identity, constitution relation or fundamental psychophysical law may each explain something while leaving a further “why that relation?” askable. The relevant comparison is what each account predicts or unifies, and whether its distinctive modal and phenomenal arguments justify stopping there. Chalmers himself gives reasons to resist brute identities and allows that a genuinely non-question-begging functional argument could resolve the hard problem. His position cannot fairly be dismissed as special pleading merely because he accepts one kind of theoretical endpoint and questions another. Those reasons, and our own proposed endpoint, remain open to criticism.
+ + + +Temporal experience offers an especially promising research target because it has measurable duration judgments, retention effects and perturbable neural mechanisms. William James’s specious present, discussed in the Stanford Encyclopedia of Philosophy on temporal consciousness, names the brief experienced interval in which change and persistence can appear together. Debates remain over whether experience itself extends over time, represents a temporal interval, or achieves continuity through other mechanisms.
+ + + + +Huw Price and Carlo Rovelli have examined the relationship between thermodynamic time asymmetry and physical records. The formation, reliability and eventual loss of memory can involve irreversible physical processes. Yet the inference from the thermodynamic arrow to the felt flow of time requires additional psychological and phenomenological premises. Time in physics need not be defined solely by entropy production, and the fact that a memory is physically implemented does not show that every act of recording constitutes an erasure with a Landauer payment. An account of the subjective present cannot simply set it equal to a thermodynamic integration interval without measurement.
+ + + + +A useful experiment would jointly manipulate working-memory retention, neural or computational integration timescales, metabolic variables and reports of subjective duration while controlling attention and task difficulty. Rival models could be fitted on training data and tested on held-out interventions. If metabolic or closure-specific variables contribute no advantage beyond sensory, attentional and recurrent-dynamics models, their claimed special role should be reduced. A stable association between those variables and duration reports would still leave the phenomenal bridge underdetermined.
+ + + + +Our best provisional conclusion is that thermodynamics supplies real constraints on the machinery of memory and temporal regulation. Its explanatory contribution to felt duration must be established empirically. A strong claim that subjective time logically follows from entropy production has not been derived.
+ + + + +The question “Who experiences?” may carry an assumption worth challenging: that a separately identifiable, unchanging spectator must sit behind everything a person sees, remembers and does. A dynamic account of selfhood can explain many functions of self-reference without such a spectator. This is a promising philosophical interpretation. It is not a demonstration that the phenomenal subject is nothing over and above an organizational pattern, or that the question of subjectivity is only a grammatical error.
+ + + + +Natural language sometimes implies an entity where a process description would suffice. “It is raining” is an ordinary example. Likewise, describing thought in the first person does not, by grammar alone, prove the existence of a further substance that performs the thinking. A model can examine the actual construction of self-reports, memory links, agency attribution and body ownership.
+ + + + +Introspection supplies evidence about what is presented to a participant, with familiar limits of attention, report and interpretation. Failure to find a separate “inner witness” through introspection may challenge a particular spectator model. It does not deductively establish that no phenomenal subject exists, nor that a first-person perspective can be fully redescribed as a behavioral pattern. Reports of selfhood, dissociation or selflessness should be tested in context and compared with neurological and psychological evidence.
+ + + + +The same restraint applies to the claim that there is one further, irreducible Hard Problem. After reconstructing the specific problems about presence, qualities, perspective, unity and report, ask what positively identifiable explanatory demand remains, rather than treating the word “hard” as proof of an additional entity. An unaccounted phenomenal relation or independent modal argument would defeat a particular decomposition. A successful decomposition would revise the proposed problem-form, not make the underlying experiences disappear.
+ + + +Nagarjuna’s dependent-origination arguments and the Buddhist anatta tradition provide important resources for questioning a self with independent essence. Their formulations and philosophical goals differ from contemporary neurocomputational models. The comparison invites precise questions about dependence and continuity; it cannot settle a present-day identity theory merely by historical analogy.
+ + + + +Constraint maintenance offers one way to model persistence through change. An organism can replace components and retain regulative organization; a person can revise beliefs while continuing a narrative and social life. Identity may be tracked by continuity of processes, memory, causal history, bodily relations and an ability to recover from perturbation. These dimensions need not be coextensive. The SEP entry on personal identity describes serious disagreements among biological, psychological and other continuity criteria.
+ + + + +The formula “identity is invariance under perturbation” is useful when its relevant invariants, boundary and timescale are named. It is inadequate as a universal numerical-identity theorem. A pattern that can be copied twice raises a branching problem; stable organization across cells does not alone decide whether those cells are one person, two persons or neither. In a formal application, specify an equivalence relation on observed states, a history dependence condition and the interventions under which it is claimed to hold. The result can be evaluated for prediction and error without assuming a metaphysical essence.
+ + + + +Levin and colleagues’ planarian regeneration work provides concrete evidence that anatomical regulation is history-dependent. Durant and colleagues (2017) document persistent changes in regenerative outcomes and a morphologically normal cryptic phenotype whose subsequent outcomes can be stochastic. Stable anatomical regulation is more flexible than a simple fixed blueprint model. Durant and colleagues (2019) separately manipulated early planarian bioelectric states, finding altered anterior–posterior polarity after a time-limited intervention. This result concerns the timing of physiological pattern control, not proof that morphospace or phenomenal properties have a particular ontology. Neither the experiments nor an attractor description establishes that personal identity just is an attractor, or that mathematical Platonism is false.
+ + + + +Epicurus argued that death cannot be directly experienced by someone after that person’s existence ends. His position is a serious contribution to thinking about fear, temporal concern and the conditions under which harm is felt. It is not a completed ethical proof that death can never harm someone. A deprivation account argues that death can harm an individual by removing future goods even when no subsequent experience of the harm occurs. The SEP review of the philosophy of death sets out both views and their objections.
+ + + + +Epicurus’s second Principal Doctrine, preserved by Diogenes Laërtius (X.139), says: “Death is nothing to us. For what has been dissolved has no sensation, and what has no sensation is nothing to us.” His Letter to Menoeceus develops a related argument. The quotation supports a claim about absent sensation, while the deprivation objection shows why sensation alone may not exhaust the ethics of death.
+ + + + +A biological death involves physical processes, physiological thresholds, social effects and sometimes irreversible loss of the capacities supporting consciousness. The account of causal failure is not complete merely by calling death a “loss of closure.” Nor does it license claims about all possible forms of postmortem existence, which belong to distinct empirical and metaphysical debates. Where clinical decisions affect an actual patient, diagnostic uncertainty and justified existing protections remain central.
+ + + + +AI systems offer a useful bounded analogy about persistence. A model’s weights may remain stored when execution stops; a tool-using deployment may preserve memory, state and agency across interactions; a broader human–AI system can exhibit coordinated control. These are different entities and timescales. Interruption of a session, destruction of a process, deletion of memories and loss of a biological organism have different causal and ethical consequences. Calling them all “death” obscures the properties under investigation. A human–AI loop’s organizational continuity is not evidence, by itself, of a single subject or of direct moral standing for every component.
+ + + + +Consciousness, welfare and personhood require their own evidence and normative bridges. Some systems may warrant precautionary treatment even when their experience is uncertain; some may matter instrumentally because they support already protected people or other beings. Neither a biological label nor a computational label is a complete inference. The question for this article is which forms of persistence help explain which observable capacities, and what separate evidence could justify any conclusion about a bearer of phenomenal perspective.
+ + + + +The original philosophical insight survives as a live proposal: selfhood need not be modeled as an unchanging occupant located behind a body. What an organizational account establishes about identity or consciousness depends on how it performs against bodily, psychological, phenomenological and other rivals, with the boundary and loss conditions stated.
+ + + + +There is a real philosophical question behind the article’s earlier appeal to distinction, identity, relation and constraint. Whenever we offer an explanation, we distinguish some target from alternatives, identify what remains stable across descriptions or observations, establish relevant relationships, and specify conditions that make a difference. These operations are unavoidable for an inquiry of this kind. Their use in reasoning is not yet a theorem about the ultimate constitution of reality.
+ + + + +Distinction: What is the target, and how is it distinguished from neighboring constructs? A consciousness measure must distinguish responsiveness, cognitive capacity, phenomenal report and welfare rather than treating them as interchangeable.
+ + + + +Identity: What makes two observations count as instances of the same pattern across times, observers and scales? The answer may depend on a declared equivalence relation, observation grain and tolerance for variation. Mathematical equality, numerical identity and persistence of an organism are different notions.
+ + + + +Relation: Which dependencies, interactions or representations connect the relevant variables? Mere correlation, causal intervention, constitutive dependence and modal entailment require different arguments. A successful map must declare what it preserves and what it discards.
+ + + + +Constraint: What possibilities are ruled out by tested data, established laws, a model’s assumptions or a definition? These sources of restriction differ in evidential standing. A formal constraint can narrow a mathematical model while the model’s physical applicability remains unsettled.
+ + + + +The four questions give the audit a portable structure. They do not prove that all existence must consist solely of constraints or that every intelligible world must be thermodynamically dissipative. In particular, Landauer’s principle concerns physical information processing in specified implementations. It cannot, by itself, answer why anything exists.
+ + + + +Kantian transcendental arguments typically begin from an accepted feature of thought or experience and try to establish a necessary condition of that feature. Their force depends on the bridge from what a subject must think or presuppose to what reality must contain. The Stanford Encyclopedia of Philosophy’s treatment of Kant’s transcendental arguments discusses the limits of this step and Barry Stroud’s challenge to ambitious anti-skeptical conclusions.
+ + + + +Consider the proposed argument: explanation presupposes distinguishable conditions; absolute nothingness supplies no such conditions; therefore absolute nothingness is impossible, and there must be something. The conclusion does not follow. An inability to describe or explain a supposed state is different from a proof that the state is metaphysically impossible. The original passage moved from a thesis about intelligible descriptions to a thesis about necessary existence without supplying that modal bridge.
+ + + + +A more defensible result survives: to compare consciousness theories, investigators need sufficiently determinate targets, concepts, models and evidence to articulate disagreement. This is a methodological condition on the inquiry. A critic can reject it only by showing a different workable mode of comparison, or by arguing that some explanatory targets resist its application. It remains open whether reality ultimately contains intrinsic properties, irreducible experience, abstract structures, or other features not exhausted by current descriptions.
+ + + + +The question “Why is there something rather than nothing?” has substantial metaphysical and cosmological versions. Neither thermodynamic irreversibility nor organizational closure supplies a demonstrated necessary-and-sufficient answer. The relevant physical laws already presuppose a domain of states, temperatures, processes or fields on which they operate. Applying them to the possibility of any domain at all would require an independent argument.
+ + + + +We can preserve the deeper challenge without a cosmic victory claim. Scholars proposing a final ontology must identify the extra premises that support its quantifier over all possible things. Someone claiming exhaustive physicalism, a necessarily experiential universe, an indispensable Platonic realm, or a transcendent creator has different commitments to establish. The success of a local description neither forces acceptance of a particular ontology nor rules out every alternative ontology.
+ + + + +This section should be judged by the same standard as the rest of the article: if a critic shows that a proposed distinction or equivalence relation fails to support a valid inference, the conclusion must be revised at precisely that point. The methodological usefulness of the four disciplines can survive the defeat of an extravagant metaphysical interpretation.
+ + + + +Alfred North Whitehead warned against treating a selective abstraction as if it gave us the complete concrete actuality being studied. A map, numerical measure or theoretical construct can be extraordinarily useful while leaving out aspects of its target. The danger arises when an investigator forgets what was omitted and attributes to the abstraction more reality, completeness or causal power than the argument has earned.
+ + + + +That diagnostic applies to every camp in this article. A Platonic morphospace may organize possible developmental outcomes without showing that an independently existing realm causally selects one. A measure such as integrated information \(\Phi\) can characterize a formal causal structure without proving it is an intrinsic phenomenal magnitude. A model of prediction error may improve behavioral forecasts without establishing that a brain literally minimizes that exact objective under every interpretation. Equally, our own constraint closure may be a useful abstraction from material processes without exhausting their phenomenal or intrinsic character.
+ + + + +Suppose a model describes a landscape of possible anatomical forms. It helps classify experimental outcomes and identify interventions in a developmental system. That is an explanatory contribution. The stronger claim that forms enter physical bodies from an independently existing nonphysical possibility space requires an additional bridge. What would change physically if the proposed ingression were absent? How does the account differ in intervention results from a developmental or dynamical model? The experimental results may support the first explanatory role even when the second remains unearned.
+ + + + +An equivalent demand applies to exhaustive physicalism. Equations and mechanisms that successfully describe particular interactions justify claims at their established domains. A move from those successes to the assertion that the physical story exhausts every intrinsic feature of existence is an extra metaphysical inference. Whitehead’s diagnostic cuts both ways.
+ + + + +For each important construct, record:
+ + + + +A claim may remain philosophically meaningful when no discriminating empirical test is available. It must then be presented in the right evidential category. The absence of an experimental discriminator limits empirical preference, while formal and conceptual criticisms remain relevant. This prevents a theory’s historical prestige or an opponent’s unanswered question from being mistaken for a positive result.
+ + + + +The hard problem of consciousness is worth taking seriously because subjective experience is part of what a theory must address. Whether its apparent distance from physical explanation is a genuine ontological division, an epistemic limitation, or a misleading framing remains disputed. Chalmers’s conceivability and explanatory-gap arguments deserve their strongest version. The corresponding responses from phenomenal-concept, representationalist, illusionist and constitutive physicalist approaches also deserve evaluation on their actual premises. No side should be allowed to turn an unresolved premise into an automatic result.
+ + + + +These objections have distinct philosophical targets. Nagel (1974) asks what an objective description can capture of a particular creature’s point of view. Jackson (1982) develops the knowledge argument and explicitly distinguishes it from Nagel’s imaginative or standpoint problem. Levine (1983) formulates a gap in the explanatory force of proposed physical–qualitative identifications. They are substantive challenges to an asserted phenomenal identity, not interchangeable demonstrations of a nonphysical substance. Moving from a specified epistemic or explanatory limitation to an extra ontology needs an argued bridge; equally, exposing that limitation is not itself a proof of exhaustive physicalism.
+ + + + +There is a reciprocal gap on the physicalist side. Identifying the neural or computational processes that reliably correlate with reports, discrimination, arousal and behavior does not automatically show that these processes exhaust what experience is. The bridge from local intervention success to exhaustive metaphysical physicalism needs its own argument. The SEP distinction between physicalism and related methodological doctrines is important here: using physical methods successfully does not settle the doctrine that everything is physical.
+ + + + +Imagine somebody defending an invisible ingredient in cheese with this move: whenever chemical analysis explains a property of the cheese, the defender claims the invisible ingredient was never meant to explain that property. When a new experimental prediction fails, the ingredient retreats again. The only remaining claim is that ordinary analysis could never reach it, and the inability to detect it is presented as a reason to believe it exists.
+ + + + +That is a recognizable form of immunization by ad hoc qualification. Antony Flew’s challenge in the philosophy of religion and Popper’s work on risky testing help identify the danger; Deborah Mayo’s severe-testing approach makes it clear that a claim deserves credit when it survives methods with a genuine opportunity to expose a mistake. The point of the cheese reductio is the evidential maneuver. It does not show that private experience is imaginary, that every version of dualism has this form, or that metaphysical arguments can be settled by a kitchen experiment.
+ + + + +A sophisticated anti-physicalist may advance a modal argument rather than a hidden causal ingredient. Such an argument requires examination of the modal bridge, its counterexamples and its competing explanations, not a demand that it predict a different EEG trace. A supernaturalist who claims an additional agent directly changes a physical process makes a different, causally testable assertion. A critic who asks only for a better account of why an otherwise known neural process should feel a certain way has not thereby postulated an occult force. Distinguishing these claims makes the challenge more difficult and more fair.
+ + + + +The Interoperability-Framework Bridge provides the broader record of how constructs, evidence and loss conditions may be translated between scientific, phenomenological, formal and ontological inquiries. Translation can preserve a claim’s intended target while leaving an additional premise unproved; shared vocabulary is never an evidential bridge on its own.
+ + + + +The evidence for many physical mechanisms is strong at specific targets. Thermodynamics constrains demonstrated physical operations; neuronal interventions change perception, attention, report and behavioral capacities; clinical instruments can help distinguish states of impaired responsiveness and inferred consciousness. These findings deserve their full force. None should be weakened to make a favored nonphysical account look equally supported.
+ + + + +The extrapolation must still be checked. A validated clinical marker is not a direct instrument for observing every possible first-person condition. Physically implementing a memory does not show that its every state update is a Landauer reset. Organizational closure can explain robustness and self-maintenance without being established as the unique cause or constitution of phenomenality.
+ + + + +The empirical research programme can proceed while treating the metaphysical residual as open. It asks what a reported experience varies with, which interventions alter it, whether the relation transports across subjects and substrates, and what serious rival models predict differently. Its successes are local and cumulative. They can become broad and powerful without silently becoming a proof about the complete ontology of the cosmos.
+ + + + +The term consciousness spans at least phenomenal presence, reportable access, awareness of information, self-representation, cognitive control, wakefulness and sometimes affective valence. These should not be collapsed into one construct. A patient who cannot respond may still be conscious; a system with sophisticated control need not have demonstrated feelings. First-person reports are fallible evidence about a target, neither unquestionable revelations nor empty behavioral noise.
+ + + + +For any empirical test, specify the subject boundary \(B\), temporal grain \(\tau\), the target capacity \(Y\), and how \(Y\) is calibrated. Record the auxiliary assumptions connecting the indicator to the target. Clinical and phenomenological standards may differ from those appropriate for artificial agents or minimal biological systems. Cross-substrate inference needs an explicit mapping, with known losses of information and competing realizations.
+ + + + +Here is the original constructive idea, stated at the strength it presently earns:
+ + + + +\(H_{B,\tau}\) (conditional organizational-phenomenal identification): At certain specified boundaries and timescales, an integrated, temporally stable, action-relevant control state that models both world and self may constitute aspects of phenomenal perspective. This is a target-specific abductive hypothesis. It remains open whether the proposed organization is necessary, sufficient, partly constitutive, or merely correlated with the subjective target.
+ + + + +The inference from measured organization to experience is a substantive phenomenal bridge, not a definition. The hypothesis needs discriminating evidence against higher-order theories, recurrent processing, global-workspace approaches, integrated-information models, biological necessities, illusionist interpretations and other serious rivals. Its support is conditional on the relevant phenomenal evidence, construct validity, intervention results and comparison protocol.
+ + + + +Bridge rule, proposed rather than proved. If a particular organization improves predictions of independently calibrated conscious-state changes under held-out interventions, including matched report and performance conditions, relative to prespecified serious rivals, it earns explanatory credit at the target tested. Identity requires additional evidence or philosophical argument beyond predictive covariation, especially when competing realizations remain observationally equivalent. The rule should be weakened or discarded when an equally parsimonious model matches or outperforms it, or when a well-calibrated dissociation undermines its claimed constitutive role.
+ + + + +This is a stronger research contribution than a verbal promise to have solved the hard problem: it states where evidence is owed. It does not guarantee that the evidence will favor the theory originally proposed here.
+ + + + +Access versus presence. Global neuronal workspace models specify mechanisms of broad access, report and flexible control. Higher-order and recurrent approaches offer different predictions about metarepresentation and local processing. Block (2011) argues for a possible overflow of perceptual consciousness beyond cognitive access, while Rounis and colleagues (2010) investigate a specific prefrontal intervention and metacognitive awareness measure. Neither result is a substrate-general definition of consciousness. Compare perturbations, no-report designs and confidence calibration to investigate the boundaries of each model. The COGITATE 2025 adversarial tests should be incorporated with their actual preregistered targets rather than used as an all-purpose verdict.
+ + + + +Replication and task sensitivity. The prefrontal stimulation result should not be treated as settled: Bor and colleagues (2017) reported failures to replicate the metacognitive impairment, while Ruby, Maniscalco and Peters (2018) challenged aspects of that replication's design and statistical interpretation. Martin, Lane and Hsu (2023) again found no corresponding impairment with real figures, but found reduced metacognitive efficiency for an illusory-figure task. These divergent findings constrain claims about the role of dorsolateral prefrontal cortex in visual metacognition; none, by itself, establishes the presence or absence of phenomenal consciousness.
+ + + +Perturbation and recovery. Casali and colleagues’ Perturbational Complexity Index, and later research in disorders of consciousness, assess properties of measured neural responses under defined stimulation and validation conditions. These measurements can be valuable for clinical inference. PCI is not an identity test for phenomenality, is not equal to IIT’s \(\Phi\), and cannot isolate closure from every correlated physiological mechanism. Reports of conscious experience under otherwise unusual responsiveness or anesthetic conditions are especially important counterchecks.
+ + + + +Boundary-specific control. For organizational closure, map maintenance dependencies under interventions, measure viability and relevant control capacity, and test whether the model predicts outcomes beyond ordinary dynamical and physiological baselines. The test must distinguish a true additional mechanism from relabeling an observed feedback loop. A human–machine coupling may enable useful system-level agency without establishing a common conscious subject.
+ + + + +Privacy and immediacy. Internal access privileges, inference latency and self-model structure may explain why some subjective judgments seem private, stable or immediate. They need experiments: manipulate report latency, introspective confidence, accessibility and neural or computational control while keeping alternatives plausible. An account of the judgment of immediacy is not yet an account of phenomenally felt immediacy. No conscious participant should be told that their experience has been explained away merely because a computational model reproduces their words.
+ + + + +The immediate tests have distinguishable outcomes. A closure model may improve prediction of survival or adaptive repair while contributing nothing to consciousness inference. A workspace model may win a report task while still leaving phenomenal constitution unsettled. An anti-physicalist proposal may offer a serious modal argument without an independently discriminating empirical model. The resulting degrees of support should be recorded claim by claim, not converted into a single winning camp.
+ + + + +The position argued for here is disciplined openness with strong local realism. We should trust the physical mechanisms where they survive severe tests, pursue the identity hypothesis where it generates risky evidence, and refuse to promote either an explanatory gap or a successful methodology into a final metaphysical inventory. The paper’s reusable contribution is the tribunal and its loss conditions. Every favored bridge, including this one, remains subject to correction.
+ + + + +The fairest test of this proposal is one that could make it lose. A rival’s failure to resolve a particular question does not award organizational closure the answer, any more than a gap in physical neuroscience establishes a supernatural cause. Different claims require different tests. An empirical mechanism needs observations and interventions. A mathematical theorem needs its assumptions and derivation. A constitutive identity requires a bridge from the measured structure to the property identified. An ontological thesis needs arguments commensurate with its scope. None is refuted merely because it is not a laboratory hypothesis.
+ + + + +The following comparisons therefore ask what each programme contributes and where its extra commitment earns, or fails to earn, explanatory credit. Some of the objections apply just as sharply to the constraint-based hypothesis proposed here.
+ + + + +Panpsychist approaches propose that experience, experientiality, or its relevant precursor is fundamental or more widespread than ordinarily supposed. The versions differ. Constitutive Russellian panpsychism, cosmopsychism, and micropsychist accounts should not be collapsed into one assertion that every electron has an empirically detectable miniature mind. The Stanford Encyclopedia of Philosophy’s treatment reviews these varieties and the substantive combination debate.
+ + + + +One serious problem is how micro-level subjects or qualities would compose the unified phenomenal states attributed to larger subjects. Coleman (2014) poses a substantial micro-subject combination objection that cannot be resolved by merely naming a whole; it does not, by itself, disprove all forms of panpsychism. Chalmers’s distinctions among subject, quality, and structural combination help sharpen that burden. A philosophical proposal may address these with modal or constitutive arguments without claiming immediate laboratory confirmation. But if it also purports to explain why a particular organism reports or discriminates a particular experience, we can ask what its experiential primitive adds beyond a matched neural, causal, or organizational model. What different result follows if the extra property is absent?
+ + + + +Galen Strawson’s realistic-monism argument challenges physicalists who accept the reality of experience to explain its place within a fully physical reality, rather than assuming a nonexperiential base suffices. Its premises deserve philosophical argument; it is not a laboratory confirmation of ubiquitous experience. The absence of a current empirical discriminator limits an empirical claim to panpsychism’s distinct contribution. It does not make the ontology logically false or empty. The reciprocal obligation is uncomfortable for this essay too: naming integrated constraint dynamics does not demonstrate that those dynamics are intrinsically felt. Both accounts owe an argument for their constitutive step. The more elaborate mechanism should receive no automatic credit simply for being more familiar to a physical scientist.
+ + + + +Levin’s experimental contributions to bioelectric pattern regulation deserve full credit. Controlled perturbations of membrane voltage and ion channels can produce striking changes in regeneration and anatomy. The planarian experiments reported by Durant et al. (2017) show stable, path-dependent changes and a cryptic stochastic regenerative phenotype. They demonstrate that anatomical outcomes depend on developmental history and altered bioelectric regulation. That is important positive biology.
+ + + + +A stronger interpretation proposes that biological structures causally access an independently existing Platonic morphospace. When that is offered as an explanation for a physical result, it needs a bridge. What property of the alleged ingression interface is measured? Which intervention or prediction distinguishes access to a pre-existing nonphysical form from a well-specified dynamical, developmental, or organizational account? The Chladni Plate Challenge to Michael Levin’s Biological Platonism develops a related dynamical alternative.
+ + + + +Levin’s published response to my criticism included this formulation: “Yes, I think a better science of Platonic forms + their interfaces will force a re-do of Landauer’s Principle (or at least, its limitations).” The final parenthesis matters. Levin allows that a future proposal may revise the range of applicability of a physical principle rather than its established conditional result. He also distinguishes his broader research on cognition and engineering from a demonstrated theory of consciousness. The unresolved empirical request is therefore specific: what measured interaction, new constraint, or prediction would make that future revision necessary? An anticipated future theory is not yet evidence that an existing physical bound has failed.
+ + + + +The two-headed and cryptic planarian results do not refute mathematical Platonism or every possible morphospace ontology. They challenge particular predictions of obligatory return to a unique canonical form, if a specific proposal actually entails those predictions. A Platonic account can allow multiple forms. The problem with adding an unlimited set of new forms after every observation is loss of differential prediction unless further restrictions are fixed independently. Likewise, a physically causally effective interface needs dynamical and thermodynamic accounting, but Landauer’s erasure bound alone does not prove that an interface is impossible or must violate conservation. This is an open challenge to a specified causal addition, not a universal metaphysical verdict.
+ + + + +Levin’s broader work on non-neural problem-solving, bioelectric regulation, and goal-directed morphology motivates serious research into biological cognition at unusual scales. It is legitimate to use cognitive terms in a stipulated functional sense when they improve prediction or control. Doing so does not automatically establish animal-like consciousness, subjective valence, welfare interests, or an immaterial source of agency.
+ + + + +The distinction is explicit in Chis-Ciure and Levin’s (2025) peer-reviewed formalization of basal cognition. Its search-efficiency measure, K, compares an agent’s cost of navigating a defined problem space against a specified random-walk baseline on a logarithmic scale. The authors discuss the possibility of consciousness outside brains but explicitly withhold claims about consciousness in that study. Levin made the scope choice even plainer in a recorded discussion (about 1:22:52): “you think I was going to fold in consciousness into all this stuff until it got, you know, further settled. No way.” The paper therefore deserves evaluation on its stated cognition target, while any separate claim linking these capacities to phenomenal experience or Platonic forms still needs an independent evidential bridge. Measured search efficiency is neither a consciousness measure by definition nor evidence that consciousness is absent.
+ + + + +A metric of behavioral competence or energetic efficiency must be evaluated as the quantity it measures. To infer cognition in a stronger sense, the investigator must define the capacity, compare less theory-laden alternatives, and show what the cognitive model predicts that a control-theoretic or selection-history account does not. A robust system may be goal-directed without a reflective self-model; a system can carry information without necessarily possessing feelings. The same restrictions apply to our language of closure, perspective, mattering and agency. A useful metaphor becomes a scientific model when its variables and loss conditions can be exposed to tests, not merely when it is rhetorically vivid.
+ + + + +Penrose and Hameroff’s Orch-OR proposal involves specific claims about quantum states in microtubules and their role in consciousness. Tegmark’s published decoherence calculation poses a serious quantitative obstacle to particular long-lived coherent states in a warm biological environment. The relevant comparison concerns the precise superposition, localization, environmental coupling and time required by the proposed computation.
+ + + + +It would be an error to declare all quantum-consciousness possibilities mathematically falsified by one estimate. Hagan, Hameroff and Tuszynski’s published response contests the physical model and relevant scales. Their alternative calculations are not themselves evidence that Orch-OR produces subjective experience. The adjudication requires independently measured parameters and prospective effects that distinguish the proposed mechanism from ordinary neurophysiology. The burden of a concrete quantum mechanism is demanding, but its failure would not automatically prove the organizational-closure account.
+ + + + +Max Tegmark’s Mathematical Universe Hypothesis makes a strong ontological proposal about mathematical structure and physical existence. We should distinguish the defensible explanatory role of mathematical descriptions, the mathematical possibility of a system, its physical realization, and the claim that every such structure literally exists. Equivalence in an abstract description does not, without a bridge, establish a bearer of phenomenal experience.
+ + + + +The constructive challenge is to state what this ontology explains beyond the relevant mathematical and physical model, and which modal, mathematical or observational considerations favor the additional existence thesis. An empirically equivalent ontology may remain philosophically discussable while lacking relative empirical confirmation. Conversely, a constraint-based theory cannot simply substitute the slogan ‘thermodynamics is real’ for a complete account of why this organization would be experienced. We should hold both proposals to their respective proof duties.
+ + + + +Whiteheadian process philosophy offers a serious systematic account of becoming, relations and what he calls prehension. A live dispute concerns the additional identity that causal inheritance is, in some sense, inherently felt inheritance. Models of causal transfer, biological constraint and temporal dependence can explain specific observations without invoking felt inheritance at those targets. Such local explanatory dispensability challenges a necessity claim where the model is sufficient; it does not establish that feeling is absent throughout nature.
+ + + + +In his essay “There’s no scientific evidence that consciousness exists”, Matthew David Segall argues that first-person awareness is presupposed in scientific observation and cannot simply be read off third-person measurements. This is a substantive transcendental and phenomenological challenge, not a denial of first-person experience. It still leaves a further question: what would establish the additional universal claim that physical causal inheritance is inherently felt inheritance? The first-person premise alone does not supply that identity bridge. The same scrutiny applies to this essay’s own candidate closure-to-experience identity.
+ + + + +The relevant questions for a Whiteheadian defender are what the experiential addition uniquely establishes, whether there is a substantive modal or phenomenological argument for it, and where the claim about physical causation crosses into a universal identity. A process metaphysics need not be discarded merely because one proposed empirical bridge is absent. Nor does the constitutive organizational hypothesis in this essay win by default. Our own supposed closure-to-experience identity needs a comparably explicit bridge, with specified losses of support. The comparison remains open where neither account supplies discriminating evidence.
+ + + + +Higher-order approaches associate consciousness with an appropriate representation of a first-order mental state. Their variants differ about conceptuality, functional role, relevant neural architecture and whether a higher-order state must itself be conscious. The familiar regress question is a challenge, not an immediate refutation: a proponent can consistently hold that the higher-order state is unconscious while making its target conscious.
+ + + + +The discriminating empirical questions concern possible dissociations among first-order perception, metacognitive accuracy, confidence reports and independently calibrated phenomenal indicators. Evidence that a perturbation alters metacognitive performance while preserving discrimination does not alone settle phenomenality. An organizational account may describe these same effects at a different level. It gains explanatory credit only if its specified organization predicts a held-out dissociation or intervention more accurately than an explicit higher-order model.
+ + + + +Variational free energy in active inference is an information-theoretic functional, whereas thermodynamic free energy is defined in physical energy units. Their relationship cannot be supplied by a shared name. The FEP and active-inference literature offers useful mathematical tools for stochastic dynamics, perception, and control under specified generative models. A flexible mathematical representation may be valuable without every broad interpretation of it becoming independently testable.
+ + + + +Bruineberg, Dołęga, Dewhurst and Baltieri (2022) distinguish formal Markov-blanket relations from stronger claims that they carve nature at unique agent boundaries. Kirchhoff and colleagues (2018) offer a substantive autonomy interpretation of Markov blankets; Friston (2022) replies to critiques of how formal models map onto living systems. The competing interpretations require independent empirical boundary criteria. That distinction supplies a real criticism when an ontological conclusion rests solely on a convenient partition. The audit question for any particular FEP application is what its assumptions commit it to, which alternatives it outpredicts, and which observations expose it to failure. The constraint-based approach faces the same risk of redescribing persistence after the fact. Calling a persistent system ‘closed’ is not evidence that closure was the cause of its capacity or of phenomenality.
+ + + + +Tononi and Koch (2015) articulate IIT’s proposed relationship between intrinsic causal structure and consciousness. Hedda Hassel Mørch (2019) presents a trilemma involving intrinsicality, non-overlap and reductionism, which extends beyond IIT alone. It is an objection with explicit alternatives, not an empirical proof that integration cannot matter. Its formal measures, postulates and empirical applications have changed across versions, so objections should name the version and proposition actually targeted. In particular, the clinically useful Perturbational Complexity Index is not the same quantity as IIT’s Φ and cannot be treated as a direct measurement of an identity between integration and experience.
+ + + + +The COGITATE adversarial collaboration (2025) found important challenges to specified predictions of both IIT and global neuronal workspace theory. The consortium makes its preregistered protocol and original analysis code (MIT license) available for independent examination; those files do not endorse this article's organizational hypothesis. A failure of a preregistered prediction forces analysis of the specific model and its auxiliaries, not instant rejection of the entire philosophical programme. IIT owes a defense of the bridge between its formal intrinsic structure and phenomenal qualities, as well as feasible tests across matched substrates. Doerig et al. (2019) present an unfolding challenge to claims about causal structure; Tsuchiya and colleagues (2020) answer by disputing the functionalist premises. Cerullo (2015) questions how the integrated-information quantity should bear on consciousness. The competing arguments require their actual assumptions, not an automatic verdict for either side. An organizational account cannot claim victory by pointing to that bridge problem while leaving its own constitutive bridge unproved.
+ + + + +The strongest organizational test here is not simply “closed loop versus replay.” IIT also attributes importance to intrinsic causal organization and may predict the same difference. Its independent contribution must be calculated under the selected version and boundary, then compared with the proposed closure and repair variables in cases where their predeclared forecasts diverge. If IIT makes the same predictions or wins at matched complexity, this article gains no distinctive evidential credit from that experiment. Enactive and active-inference accounts deserve the same opportunity.
+ + + +Global workspace models study how information becomes widely accessible for report, action, memory and flexible control. They have produced influential neurocognitive research and concrete predictions. The interpretation of late ignition, posterior representations, prefrontal activity and no-report evidence remains sensitive to the task and measurement design; the COGITATE collaboration adds genuine adverse evidence against some anticipated signatures.
+ + + + +Global availability offers an informative mechanistic model for some cognitive operations. It does not by itself resolve whether conscious phenomenality is identical to access, accompanies it under further conditions, or sometimes dissociates from it. This is precisely a question on which the closure-based hypothesis must demonstrate added predictive value. Both hypotheses need carefully matched performance and perturbation conditions, with first-person evidence treated as fallible data rather than either a revelation or noise.
+ + + + +Hoffman, Singh and Prakash (2015) introduced the Interface Theory of Perception in an evolutionary setting. Prakash, Stephens, Hoffman, Singh and Fields (2021) later stated and proved a conditional Fitness-Beats-Truth theorem under specified assumptions about world states, perceptions, payoffs and the competing strategies. The narrower point that perceptual evolution does not automatically guarantee veridical world reconstruction is worth taking seriously. Manolo Martínez (2019) constructed a family of counterexamples to a stronger general inference from usefulness to nonveridical representation. Which mathematical result applies depends on the actual competing strategy class, payoff structure and information sources. Neither paper by itself settles the ontology of perception.
+ + + + +The further claim that spacetime and physical objects are merely interface icons for an underlying domain of conscious agents requires its own argument. Quantum-gravity research questioning spacetime’s fundamentality does not by itself favor conscious-agent ontology. Equally, planarian path-dependent regeneration is not an experiment that directly falsifies Hoffman’s perception models; it concerns a different explanandum. A successful critique must identify a specific shared prediction and a result on which the models genuinely disagree. If the conscious-agent proposal yields no advantage over a well-specified evolutionary and dynamical account, its additional ontology remains underdetermined. That is narrower than proof of its negation.
+ + + + +The comparative lesson is not that the naturalistic organizational proposal has already won this tournament. The present evidence supports many locally successful models and exposes several ambitious bridges that remain unearned. The proposed method succeeds only to the extent that it finds real discriminators, accepts adverse results against its own favored hypotheses, and helps competing accounts make progress under shared constraints. In that form the criticism retains its point: explanatory gaps do not function as vouchers redeemable for an extra ontology, including an exhaustive physicalist one.
+ + + + +There are useful empirical tests for human responsiveness, covert command-following, sensory processing, metacognition, brain dynamics and some indicators associated with consciousness. Their existence directly challenges a sweeping claim that consciousness-related research has no testable content. It does not establish that we can directly classify every conceivable conscious bearer, or that one favored theory has already survived a decisive universal test.
+ + + + +Giacino and colleagues’ Coma Recovery Scale–Revised (CRS-R) assesses behaviors associated with levels of consciousness under specified clinical conditions. Historical studies by Andrews, Schnakers and colleagues document consequential diagnostic disagreement, particularly when relying on informal examinations. Specific percentages depend on sample and diagnostic protocol; they should not be generalized to all facilities or times without checking the study design.
+ + + + +Owen and colleagues (2006) showed covert command-following through distinguished tennis and navigation imagery in a patient who met behavioral criteria for the vegetative state. Casali and colleagues’ PCI research (2013) offers another operational route through TMS-evoked brain-response complexity. These are genuine empirical achievements. Neither establishes a metaphysical equation between consciousness and closure or the reliability of the same indicator in all species, artificial systems and clinical conditions.
+ + + + +Global-workspace approaches, recurrent processing accounts and other neural theories make predictions about access-related neural dynamics. The COGITATE 2025 adversarial collaboration exposed important difficulties for specified predictions of both IIT and global neuronal workspace theory. These partial failures should feed a dependent claim audit rather than an unconditional verdict against either research programme.
+ + + + +In a prospective test of a candidate organizational marker, the clinical or participant-linked target must be scored independently of that marker. Record how awareness, responsiveness, communicative capacity and phenomenological reports were assigned, by whom, with what blinding and inter-rater dependence, before a closure model is trained. Reusing the same physiological feature to define the outcome and then predict it can manufacture apparently excellent validation. Ambiguous states remain ambiguous when the independent evidence does not decide; a marker should not decide a patient's rights or treatment by stipulation.
+ + + +Different systems can be tested for discrimination, associative learning, flexible choice, counterfactual control, affective responses, developmental plasticity and other particular capacities. Human, animal and computational measurements vary in construct validity and available calibration. A complex signal or a behavior resembling human cognition is not a validated experience meter. On the other hand, absent human-style speech is weak evidence against animal or unfamiliar consciousness.
+ + + + +Comparisons with rocks, water chambers or engineered controllers can be informative when the exact measured feature is stated. They become irrelevant when a difference in memory or behavioral sophistication is silently promoted into an answer about intrinsic phenomenality.
+ + + + +Popperian falsifiability concerns how a claim exposes itself to potentially conflicting observations. Mature model testing also requires auxiliary assumptions, calibration, statistical power, replication and attention to underdetermination. In “Theology and Falsification,” Antony Flew warns of the “death by a thousand qualifications”: successive revisions can remove the content that originally made a claim testable. That is a challenge to specify genuine revision costs, not a rule that one failed observation deductively destroys an entire research programme. Deborah Mayo’s severe-testing framework makes an additional demand: results should have passed procedures with a genuine chance of revealing relevant errors.
+ + + + +Landauer’s specific bit-erasure claim, a biological closure dependency, a planarian regenerative outcome and a neuroscience prediction are different targets. Failure of a Landauer prediction under genuine matching conditions would bear on the thermodynamic claim. Failed replication of a particular bioelectric intervention would bear on that result and its dependent explanations. Neither directly falsifies every organizational model or settles consciousness. A critic of a Platonic or panpsychist ontology may need a modal or conceptual argument rather than a laboratory falsifier, while a proposed physical ingression mechanism owes physical and predictive accounting.
+ + + + +The valuable asymmetry is between what has earned local evidence and what is only compatible with it. It is not between “science” and everyone who disagrees with an identity hypothesis. If a rival makes a riskier, independently successful prediction, this programme must recognize that gain. If its own models lose their claimed advantage, it must say so and preserve the failure record.
+ + + + +Thinking about memory, relations and ecological dependence benefits from scholarship far beyond modern European philosophy. The methodological difficulty is to honor the differences. Two descriptions that use words like relation, process or constraint need not make the same metaphysical claim, identify the same bearer or establish the same explanatory bridge. A cross-tradition analogy becomes evidence only when its precise construct and independent predictions can be compared without erasing cultural context or reducing someone else’s knowledge to a label in this framework.
+ + + + +Australian Aboriginal songlines involve complex relationships among land, knowledge, ancestry, story, navigation and responsibility. They are not reducible to memory tricks or to graph-theoretic “constraint paths.” Lynne Kelly’s work on memory traditions invites an informative comparison between situated, practiced recall and modern external recording. What it does not establish is a universal physical ontology.
+ + + + +Reser, Simmons, Johns and colleagues (2021) tested an Australian Aboriginal-inspired memorization method in a medical-education context. Their first study involved first-year medical students (\(N=76\)). Among participants beginning below perfect recall, 12/19 (63%) in the Aboriginal-method group subsequently reached 20/20 recall, compared with 10/22 (45%) in the memory-palace group. The paper labels 2.82 (95% confidence interval 1.15–6.09) an odds ratio for the Aboriginal-method group. An exact arithmetic check shows that 2.82 equals the ratio of its perfect-recall proportion to the study-wide baseline proportion, \((12/19)/(17/76)\), rather than a conventional odds ratio. A separate, unadjusted direct comparison of the two selected groups yields odds ratio 2.06 and two-sided Fisher exact \(p\approx0.350\); this exploratory result does not establish superiority on perfect recall. Both taught approaches improved overall recall, with comparable performance on that outcome, while the Aboriginal-method group showed better immediate sequence-recall results. At six weeks, only eight students returned; no robust long-term superiority conclusion is justified by that attrited follow-up. The authors explicitly distinguish the laboratory teaching method from actual culturally situated Songlines. This limited educational comparison is promising and worth replication. It is not evidence that all Indigenous knowledge systems outperform other methods, nor that an Indigenous ontology and thermodynamic monism have independently converged.
+ + + + +Tyson Yunkaporta’s Sand Talk and Robin Wall Kimmerer’s Braiding Sweetgrass discuss situated knowledge, practice, reciprocity and relationships with land and other life. Their accounts have explanatory and ethical substance. Similarities with relational realism can identify a productive research question: which relations help preserve ecological or social functions over long timescales, and what observations demonstrate that dependence? Different traditions may supply different answers, values and forms of authority.
+ + + + +An Indigenous claim about kinship, respect or personhood should not be translated without argument into a claim that the relevant system is a dissipative graph. Nor should this article presume to speak for every Aboriginal, Potawatomi or other Indigenous tradition. The strongest collaboration would involve the relevant scholars and communities as interpreters and participants, with proper provenance, reciprocity and consent where new research is proposed. Shawn Wilson’s Research Is Ceremony (2008) supplies one grounded Indigenous research-methodology perspective; it must not be mistaken for authorization to generalize across distinct communities.
+ + + + +The Upanishadic neti neti (“not this, not this”) method and the Buddhist anatta traditions can both challenge unexamined identifications of self with particular passing states. It would nevertheless be a major distortion to treat them as equivalent proofs that no intrinsic reality exists and only organizational constraint closure remains. Some Vedantic interpretations of neti neti proceed toward an irreducible self or ultimate reality; Madhyamaka and other Buddhist traditions have different accounts of emptiness and conventional designation. Compare the actual texts through Patrick Olivelle’s Upaniṣads translation and Jay Garfield’s translation of Nāgārjuna, rather than treating either tradition as empirical replication. Their arguments should be read in their own philosophical contexts and compared claim by claim.
+ + + + +This difference is an especially good check on the temptation to count similar rhetoric as independent confirmation. Two traditions may share a technique of negation while drawing incompatible conclusions. The shared method may still be philosophically instructive, but one cannot use it as evidence for a conclusion that one of those traditions rejects.
+ + + + +If several language models produce the phrase “constraint satisfaction under thermodynamic bounds” after receiving similar prompts or framework instructions, that result is an observation about the prompted systems. It does not yet demonstrate independent discovery of an invariant physical principle. Training corpora may overlap; prompt design, selection effects, common scholarly terms and the human orchestrator may create correlated outputs. Architecture differences alone do not remove those dependencies.
+ + + + +An informative study would predeclare the prompt variants, conceal the preferred target phrasing, compare models with generic baselines, count contradictory answers, preserve every failed branch and let independent evaluators score claims for novelty and validity. The same protocol should test whether a simpler control-theoretic or causal explanation does as well as the Recursive Corrigibility Framework. Until such controls exist, model agreement is weak and dependence-sensitive evidence, not a substitute for laboratory replication or peer review.
+ + + + +The honest result is more interesting than a predetermined convergence story. These traditions and computational models raise questions about how situated agents carve useful differences from a changing world. Where their predictions or arguments agree at a defined target, the agreement deserves examination. Where they diverge, we should preserve the disagreement and find out why.
+ + + + +The Recursive Corrigibility Framework must be capable of rejecting its own preferred consciousness hypothesis. It also needs to distinguish failure of a subsidiary empirical model from failure of an entire research method. An adversarial test of one equation may reveal a genuine defect in that equation while leaving independent observations and a reusable evaluation procedure intact.
+ + + + +Landauer: The bit-erasure bound applies under declared physical assumptions. Evidence that the bound fails under those assumptions would challenge the relevant thermodynamic claim. An abstract identity operation or a memory retained for a finite interval does not test an erasure bound. Conversely, an observed bit-erasure cost does not establish that every maintained distinction has that cost.
+ + + + +Organizational closure: At a specified biological target, ask whether a mapped network of mutually maintained constraints predicts resilience or recovery better than a matched dynamics-only or control-theoretic model. If it fails on held-out interventions, withdraw the asserted explanatory advantage at that target. An artificial system with comparable capacities despite a different organization challenges any asserted necessity claim.
+ + + + +Bioelectric memory: Replicate specified planarian interventions and their observed long-term morphological outcomes, including the cryptic stochastic state. Replication failure would revise a model of that experiment, not every attractor account, Platonism or consciousness theory. The primary Durant et al. (2017) claim should be preserved in its actual experimental scope.
+ + + + +RAF autocatalysis: Formal results about reflexively autocatalytic, food-generated sets need their stated model assumptions. A mismatch with a specified prebiotic setting would diminish the applicability of a proposed origin-of-life mechanism. It would not directly show that consciousness cannot occur under organizational closure or that all chemistry lacks autocatalytic organization.
+ + + + +Phenomenal identification: The conditional identity in this essay requires calibrated, target-appropriate evidence beyond prediction of report. If a simpler neural dynamical model matches every measured outcome, the closure-specific hypothesis loses its distinctive explanatory claim. If the proposed structure dissociates from independently justified phenomenal evidence, the constitutive hypothesis must be revised. The difficulty of producing decisive evidence should not be used to immunize the hypothesis.
+ + + + +Additional failure gates: If an outcome labelled “conscious” was defined from the closure measure itself, withdraw the purported validation and redo the label design. If held-out joint predictions provide no incremental gain beyond the strongest equal-budget rival, withdraw the claim of a unique organizational discriminator. If decomposition of the Hard Problem omits an independently defensible phenomenal relation or necessity argument, abandon that specific claimed dissolution while retaining any successful mechanistic explanations. These losses attach to different claims and do not logically entail a universal verdict for or against consciousness.
+ + + +Exhaustive physicalism and supernatural or anti-physicalist additions: A claim that all reality is physical needs a warranted extension beyond local successful mechanisms. A claim of additional causally active nonphysical agency needs positive evidence for its contribution and a bridge to observed effects. A purely modal thesis may be disputed through modal arguments. Neither a physical explanatory gap nor a successful local mechanism earns the cosmic conclusion for free.
+ + + + +This article does not answer why a universe exists, whether experience has intrinsic qualities unrepresented by our best physical models, whether biological substrates are necessary for any phenomenal property, how to calibrate the experience of radically unfamiliar systems, or whether closure is ever sufficient for phenomenality. It cannot settle moral standing by counting feedback loops. The relevant uncertainties should accompany the claims rather than being treated as a footnote after a confident verdict.
+ + + + +The evidence for human conscious states is substantial and includes first-person reports, behavior, clinical histories, neurophysiology and intervention. Other organisms and engineered systems supply different combinations of evidence. The same rule of inference should apply to matched evidence while respecting genuine differences in what is known and the costs of a mistaken exclusion.
+ + + + +RCF should be benchmarked against a straightforward source-checking and rival-model protocol that does not use its distinctive vocabulary. Measure error discovery, claim typing, preregistered revisions, retained sources, successful predictions, reproducibility, human readability and time/cost. Check whether reviewers reproduce results without access to favored conclusions and whether prompt changes materially alter outputs.
+ + + + +If RCF fails to improve these outcomes, compress or abandon the redundant machinery. If its applications generate new unfalsifiable declarations while criticizing others for doing so, demote the framework’s own standing. Progress is an earned, revisable record of specific corrections and predictive gains. Neither the name falsification nor its replacement with corrigibility is a substitute for this test.
+ + + + +The positive view now has to earn the term identification. It is tempting to say that consciousness simply is organizational closure viewed from within. This was an intended conditional, target-specific abductive hypothesis, but the unqualified wording can sound like a settled universal identity. Organizational closure, self-modeling, integrative access and ongoing control are individually measurable constructs. Their relation to phenomenal presence remains a research question.
+ + + + +The hypothesis can be stated more carefully: in a declared class of systems, at boundary \(B\) and timescale \(\tau\), a temporally integrated, action-relevant control organization may be constitutive of some aspects of phenomenal perspective. “Constitutive” is a stronger claim than causal correlation or predictive usefulness. The class, proposed phenomenal target, rival realizations, and evidential bridge must be specified before evaluation. A network with organizational closure alone is not declared conscious, and a system without demonstrated closure is not declared unconscious.
+ + + + +David Chalmers’s zombie argument asks whether a complete physical duplicate of a conscious person lacking phenomenality is ideally conceivable and, if so, whether that conceivability establishes metaphysical possibility. Critics dispute the conceivability–possibility bridge, the intended completeness of the physical description and the implications of phenomenal concepts. These are substantial modal questions, detailed in the SEP entry on philosophical zombies. Neither casual imagination nor a verbal decree that organization is experience settles them.
+ + + + +Daniel Dennett’s skepticism about our ability to imagine genuine functional duplicates supplies an important counterargument. But the existence of a complete functional theory of consciousness is precisely what the organizational proposal has yet to show. It would beg the question to assume that a fully specified closure architecture is already known to feel something, then use that assumption to rule out zombies. A defender of the zombie argument likewise owes a bridge from ideal conceivability to the relevant modal conclusion. The constructive task is to state the assumptions on each side and identify where the entailment can be contested.
+ + + + +The report “there is something it is like to be me” is evidence from a participant about their condition, even though introspection and language can be mistaken. Explaining how a person makes that report is an important neurocognitive task; it need not exhaust the question whether the state being reported has phenomenal character. Keith Frankish’s illusionism and related models deserve attention because they generate questions about how introspective judgments of ineffability and specialness arise. Opponents can press the further question of why the allegedly illusory seeming seems like anything. Both deserve a serious response.
+ + + + +Some features of self-representation may explain why a participant experiences reports as private or immediately available. A model can predict their behavior and reported confidence. Establishing identity between such a mechanism and felt presence needs independent support. The absence of a separately isolated phenomenal residue variable is not positive proof that none exists. An observer who uses physical measurements may lack access to some relevant properties, and rival interpretations can be empirically equivalent. The legitimate conclusion is narrower: no additional empirically effective cause should receive explanatory credit without distinguishing evidence appropriate to that proposed effect.
+ + + + +A serious positive rival remains open: the felt unpleasantness of pain or the character of pleasure may itself contribute to avoidance, salience, preference formation or reasons, beyond what a proposed functional model currently explains. A proponent must specify what explanatory work that ingredient earns and why a lower-commitment account misses it. A functional reconstruction must preserve the lived target, and an intrinsic-experience proposal must avoid extending a local possibility into universal panpsychism by stipulation. I do not count our inability to measure a phenomenal remainder as evidence that no remainder exists.
+ + + +Sleep, dreaming and anesthesia change responsiveness, integration, memory and experience in different ways. Massimini and colleagues (2005) found altered cortical effective connectivity during non-REM sleep under their perturbational protocol. This is evidence about measured cortical responses at the studied state, not a necessary-and-sufficient test of experience. Some anesthetic states permit reports of vivid experience despite unresponsiveness. Alkire, Hudetz and Tononi (2008) review how diverse anesthetic mechanisms bear on conscious processing; their synthesis is not evidence that an anesthesia class maps one-to-one onto subjective experience. Brain injury can impair communication without extinguishing consciousness. These dissociations show why behavioral absence, neural complexity, organismic self-maintenance and phenomenality must be investigated separately.
+ + + + +Casali and colleagues (2013) introduced the Perturbational Complexity Index (PCI) to assess complexity in the brain’s response to perturbation, including settings where behavioral reports are unavailable. PCI has valuable validation against specified human conscious-state benchmarks. PCI correlates with conscious states in clinical benchmarks, while other neural and physiological variables remain plausible explanatory candidates. Casarotto and colleagues (2024) compared spontaneous EEG patterns with PCI in 40 minimally conscious patients; spontaneous spectral signatures varied while PCI stayed above the independently validated cutoff in that sample. The result supports carefully bounded clinical discrimination, not a direct measurement of phenomenality or universal validity across substrates.
+ + + + +In Owen and colleagues’ 2006 study, a patient meeting then-current criteria for the vegetative state produced differentiated fMRI responses to instructed tennis and navigation imagery. The finding provided evidence of covert command-following in a behaviorally unresponsive patient. Calling that study simply a demonstration of “locked-in syndrome” was an inaccurate conflation of clinical categories. The lesson is that absent external behavioral response cannot safely be treated as evidence of absent awareness. The precise inference still depends on the imaging protocol, controls and patient-level validation.
+ + + + +These studies give the organizational hypothesis informative possible tests. They do not establish that every disruption of a closure network eliminates consciousness or that every restoration of a closure network restores it. A stronger constitutive claim must identify which organizational intervention distinguishes it from a matched neural dynamical rival.
+ + + + +Criterion contamination is a release-blocking measurement error. Before comparing a closure measure with an alleged consciousness outcome, record who assigned each trial’s label, by what observations and assumptions, whether they knew the closure score, and whether several indicators share the same underlying source. If the classification of “conscious” cases already incorporates that measure, its apparent accuracy can be a definitional artifact. Even independently collected verbal reports, command following and no-report neural indicators can have correlated errors. Chalmers’s discussion of the pre-experimental principles behind neural-correlate research makes the interpretive bridge clear. Register the label construction, blinding, uncertain cases, calibration and dependence before analyzing the decisive contrast. A good classifier of a specific human state is not thereby a universal assay of experience.
+ + + +The architecture of internal state access can make some information easier for a system to use than for external observers to measure. Low-latency control loops can also explain aspects of fast judgments about one’s own condition. Whether these mechanisms explain a feeling of privacy or immediacy depends on separate phenomenal evidence. A privately maintained control variable is not necessarily inaccessible to instrumentation, and direct or rapid access to a computation does not entail experiential awareness.
+ + + + +The testable proposal is to manipulate internal access, metacognitive calibration, temporal integration, and report pathways while measuring the relevant experiences where participants can report them. Matched-performance and negative-control conditions must distinguish a change in phenomenality from a change in capacity to describe or remember it. Where reports are unavailable, the evidence is weaker and model dependence greater. The framework should say so rather than assign zero or certain consciousness.
+ + + + +Closure specificity. After controlling simpler physiological and dynamical variables, a quantified network-maintenance measure should predict a prespecified aspect of consciousness evidence on held-out interventions. If it contributes no additional out-of-sample predictive or interventional value, the closure-specific explanation loses credit at that target. It is not enough that every clinically conscious person has some metabolism or feedback.
+ + + + +Access dissociation. If global report availability, metacognition or self-model variables change while well-calibrated indicators of phenomenality remain stable, a theory asserting their necessity loses support. If both change together, correlation alone does not establish identity. A preregistered comparison should include serious global-workspace, recurrent and higher-order alternatives.
+ + + + +Cross-boundary transfer. If an artificial or nonhuman system performs relevant integration tasks while lacking biological self-maintenance, its competence challenges the necessity of particular substrates for that competence. It does not by itself demonstrate phenomenality. Conversely, lack of a biological mechanism should not be declared decisive against machine experience without independent evidence for that mechanism’s necessity.
+ + + + +Fragmentation and perspective. Interventions affecting distributed control or body ownership may alter the structure of self-report and perspective. Such effects can be compared with neurological, psychiatric and experimental alternatives. They do not force the inference that every distributed controller has distributed experience.
+ + + + +The prospective organizational fingerprint. The strongest proposed discriminator is a joint, boundary-specific dependence among endogenous maintenance of constraints; recursive self/world/action regulation; integration of memory, uncertainty and valuation; temporal recovery history; and counterfactual control over later sensing and action. No single component is exclusive to this programme. A meaningful disagreement requires a precise GNWT, recurrent, higher-order, IIT, active-inference, enactive, biological or simple dynamical model to forecast a different result under the same boundary, intervention, information and resource budget. IIT may agree about internal causal powers; active-inference models may agree about organismic regulation and hysteresis. Where their strongest versions predict the same outcome, the comparison is presently nondiscriminating.
+ + + +For a prespecified target \(Y\), the principal functional wager is a positive held-out score improvement from the closure-specific model over the best frozen rival, under matched intervention and complexity conditions. It is not the false universal claim that conditioning on an organization variable must increase the probability of consciousness for every individual. It is still less a claim that a predictive gain proves the variable constitutes feeling. See Appendix C for the score, dependency controls and separate phenomenal bridge.
+ + + +The following experimental protocols should be treated as proposals requiring protocol-level source checks, not as completed preregistrations or verified claims that only this theory predicts the listed outcome. Where they lack identifiability, adequate controls, power analysis or a suitable indicator of phenomenal presence, the correct audit result is not yet discriminating.
+ + + + +The protocols below are candidate designs. They have not been preregistered, powered, ethically approved, or shown to isolate phenomenal experience from every relevant alternative. A quantitative prediction becomes scientifically defensible only after its parameters, comparator, measurement error and possible failure criterion are fixed independently of the outcome. We should separate tests of organizational control, tests of consciousness indicators, and tests of the proposed constitutive bridge.
+ + + + +Target: whether a measured graph of physical constraint dependencies predicts perturbation spread beyond a matched model of ordinary physical coupling.
+ + + + +Proposed design: estimate dependencies in a system with an independently characterized architecture; intervene on selected nodes where lawful and safe; record local and downstream responses at predefined time lags. Compare a constraint-maintenance graph, a dynamical connectivity model and a simpler autoregressive baseline on held-out perturbations. Report confidence intervals, identifiability, intervention specificity and how the physical scale is mapped.
+ + + + +Loss condition: the closure-specific graph fails to add predictive accuracy after matching confounders, or its claimed dependencies vanish under targeted controls. This would weaken a particular causal explanation of regulation, not refute every account of consciousness. There is no presently justified universal law \(\tau\propto\mathrm{PCI}^{\alpha}\) with \(\alpha>0\); its earlier statement was a research guess without a derived coefficient or operational domain.
+ + + + +Target: how development of integration, self-modeling, report capacities and candidate consciousness indicators covaries over time.
+ + + + +Proposed design: begin with ethical observational datasets and age-appropriate noninvasive measures. Model neurological maturation and behavioral capacities separately, including sensory and motor limitations, sleep state and prematurity. Any perturbation study in infants requires independent clinical and research-ethics review, age-specific evidence of safety, and real scientific necessity; this article supplies no authorization or procedure for such interventions.
+ + + + +Loss condition: a prespecified organizational model fails against neural maturation and task-performance competitors. A sigmoidal PCI curve with a five-to-twelve-month inflection is not an established prediction from closure mathematics and should not be presented as one. The appearance of a behavior such as mirror recognition should not be equated with the first appearance of phenomenal experience.
+ + + + +Target: whether common organization-sensitive measurements transfer meaningfully to cephalopods, corvids, insects and non-neural biological systems.
+ + + + +Proposed design: validate species-appropriate behavioral and physiological indicators against known capacities, using nonharmful or approved perturbations and necessary veterinary and welfare oversight. Compare organization, learning, flexible control, responsiveness and possible valence as different constructs. Human PCI thresholds do not transfer to another nervous system merely because a signal-complexity calculation can be performed. Animal welfare and ethical constraints must be considered independently of uncertainty about phenomenality.
+ + + + +Loss condition: a proposed common marker lacks transport validity or adds no prediction beyond species-specific physiological and behavioral models. Unexpected integration in an insect would revise a measurement hypothesis; it would neither prove nor disprove insect feeling by itself. Distributed octopus arm control likewise cannot be assigned a number of phenomenal subjects without a bridge from control unity to phenomenal unity.
+ + + + +Target: whether organizational integration explains conscious-state indicators beyond dose, drug class, neural dynamics and reportability.
+ + + + +Proposed design: analyze ethically acquired clinical or approved research data comparing conditions such as propofol, ketamine and other agents. Keep arousal, report ability, recall, time of assessment and inferred experience separate. Model PCI or EEG-complexity measures under the actual measurement protocol; do not treat them as a universal scalar of felt intensity.
+ + + + +Sharper wager: compare induction and recovery at matched current drug exposure, arousal, neural recurrence, broadcast and physiology. Freeze a history-aware dynamical baseline before asking whether measured endogenous restoration predicts an additional part of the trajectory. Hysteresis alone is not distinctive evidence because ordinary dynamical and active-inference models predict it too. Delayed memory and post-anesthetic report error are major threats to the interpretation.
+ + + +Loss condition: closure-related measures contribute no discriminating information relative to a well-calibrated neural and pharmacological baseline, or the direction of a prespecified prediction reverses. No formula in this framework entails that all propofol states eliminate phenomenality, that ketamine must always reduce PCI, or that preserved recall is identical to experience during unresponsiveness.
+ + + + +Target: whether specific, independently defined capacities show impairment tied to a measured organizational variable.
+ + + + +Proposed design: preregister one syndrome, one capacity and an appropriate comparator at a time. Use validated diagnostic instruments, matched controls and longitudinal measurements where feasible. Avoid the assumption that schizophrenia, autism, depression or dissociation each has a single distinctive “closure deficit.” These heterogeneous diagnostic classes carry large variation and already have several serious mechanistic models. Do not infer diminished consciousness, personhood or welfare from a diagnostic label or a complexity score.
+ + + + +Loss condition: the named organizational variable fails construct validity, yields no out-of-sample improvement or cannot distinguish competing neurological explanations. An observational association alone cannot identify an intervention target.
+ + + + +Target: whether contingent feedback, predictive control and integration history help explain ownership judgments for prostheses or interfaces.
+ + + + +Proposed design: compare control policies and feedback conditions in consented, ethically approved research; measure task performance, ownership ratings and adaptation separately. Report dependence on training history and interface design. Feedback can change body-ownership judgments, but no theorem requires bidirectional interfaces to be felt as self or unidirectional ones never to be so felt. A split-brain patient’s changed integration patterns also cannot settle how many phenomenal subjects exist without independent evidence.
+ + + + +Sharper wager: predeclare live participant-contingent feedback versus matched yoked replay, with independently elicited reports of felt ownership, presence and agency alongside task performance. An observed difference would concern human experience at the participant–device boundary, not consciousness in a self-maintaining device. Predictive-processing, enactive and sensorimotor-contingency rivals must be allowed to forecast the same difference.
+ + + +Loss condition: a simpler sensorimotor contingency model explains the judgments as well as the proposed closure-specific metric. Continuity of a computational pattern in an uploading thought experiment does not resolve numerical identity or phenomenal survival; those remain separate philosophical questions.
+ + + + +Target: determine which cognitive, regulatory and continuity capacities belong to a base model, an agentic deployment with memory and tools, or a larger human–machine system.
+ + + + +Proposed design: predefine the bearer and time horizon, then ablate memory, autonomous action, tool use, feedback, human participation and external infrastructure. Measure persistence, internal monitoring, error correction, policy revision and counterfactual control, comparing simpler agent architectures. Preserve raw participant reports when elicited and test them for stability, incentives and calibration. A deployed system can gain substantial system-level capabilities without a change to the underlying model weights or creation of one consciousness spanning its users.
+ + + + +Two-tier closure contrast: in engineering tests, compare an externally repaired or corrected controller with an endogenously self-restoring controller, matching observable outputs, memory, recurrence and resources as far as feasible, then apply unanticipated perturbations. A real intervention-sensitive difference establishes a difference in causal organization, not subjective feeling. The human live-versus-replay task above provides a separate phenomenal evidence tier. A predeclared comparison between an IIT causal complex and a different proposed closure-defined subject boundary is more informative than claiming a replay result uniquely challenges IIT.
+ + + +Loss condition: apparent organizational capacities vanish when scaffolding is removed, or a sophisticated closure account adds no predictive value beyond a simpler controller. Neither this outcome nor its opposite establishes phenomenal experience. Metabolism, hardware self-repair, subjective concern and a preferred “stake in persistence” should not be designated universal prerequisites without evidence of necessity. Conversely, self-description alone is not enough to establish consciousness or welfare.
+ + + + +Target: the judgment that consciousness is inexplicable by a specified mechanism, kept separate from the person's experience. Randomized, counterbalanced descriptions of substance, process, identity, psychophysical law and constitution could test how explanatory-gap judgments vary with framing, philosophical background, comprehension, language and repeated reflection. Collect confidence and the reasons offered; do not assume a change in judgment changes the underlying experience.
+ + + +Loss conditions: if well-calibrated judgments remain stable across framing and track independent modal or phenomenological reasons, the proposed genealogy-of-bias explanation loses support. If judgments vary with irrelevant wording, the corresponding metaphysical inference may warrant less confidence, but feeling has not been explained away. A separate philosophical challenge asks whether the unified Hard Problem preserves an additional target or indispensable relation after its constituent questions have been reconstructed.
+ + + +Every proposed study should produce a claim card with the target construct, boundary and timescale; rival hypotheses; operational measures and known error; intervention and negative controls; estimated effect and uncertainty; preregistered loss condition; adverse events and relevant ethical constraints; source and dependency list; and a plan for independent replication. Formal validity, predictive success, phenomenal interpretation and moral standing must appear as distinct verdicts. No single study or measure decides the ontology of consciousness.
+ + + + +Requirements before genuine preregistration. Select an actual study population and site; obtain permission and human/animal ethics review where applicable; predeclare sample size and its power or precision justification, inclusion/exclusion and missing-data rules, target labels and their error model, perturbation doses or configurations, outcomes, rival implementations, analysis code, holdout allocation, score, meaningful improvement margin, multiplicity treatment, stopping rule, adverse-event response and dated deviation procedure. Independent rival advocates should be invited to correct their assigned predictions before the decisive data are seen. Publish null and adverse results. The accompanying prospective protocol freezes candidate hypotheses and loss conditions, but is not misrepresented as a completed participant-study registration or IRB approval.
+ + + +We currently have far stronger direct evidence about human consciousness than about possible experience in artificial, radically unfamiliar biological or distributed systems. The evidence differs for good reasons: humans can report experiences, participate in calibrated experiments, exhibit characteristic neural and behavioral dissociations, and maintain long-term physiological and social continuity. A refusal to assume that biology is metaphysically necessary should preserve that evidential asymmetry.
+ + + + +Substrate agnosticism means that no material should be ruled in or out by definition. A silicon system, alternative biochemistry, animal nervous system or distributed organization may instantiate many familiar cognitive and control functions. Whether it also instantiates phenomenal presence requires its own indicators and bridge. Biological specifics might turn out to matter for some aspects of experience. Computation might realize others. Neither a universal biological necessity nor a universal computational sufficiency has been established.
+ + + + +The boundary of a proposed subject is especially important. A base language model, a persistent deployment with tools, and a human–AI interaction loop occupy different boundaries and timescales. They may exhibit different degrees of error correction, agency and memory. Their coupling does not automatically produce a single agent, a unified perspective or an entity with direct moral standing. Those implications require tests of organization, continuity and phenomenal unity separately.
+ + + + +Organizational closure remains one candidate mechanism. If a system lacks closure but reliably exhibits independently justified evidence of experience, a necessity version of the closure hypothesis loses support. If a closed system displays robust control and self-maintenance but no well-calibrated evidence of experience, a sufficiency version remains unsupported. Where neither type of case can be tested adequately, uncertainty should be preserved instead of silently replacing an open scientific question with a substrate preference.
+ + + + +This approach also applies to ethics. A system’s contribution to a protected human or ecological function may warrant protection of that contribution without establishing the component’s personhood. Direct welfare or moral standing requires a separately justified ground and evidence for its instantiation. Changing the scale from component to coupled system cannot transfer status automatically. An open consciousness inquiry may call for precaution where severe irreversible harm is plausible, but it does not justify arbitrary self-preservation demands, erasure of human protections, or unbounded control.
+ + + + +The main result is a way to distinguish a theory’s supported mechanisms from its untested metaphysical additions. It may also improve the design of consciousness research, but those practical benefits have to be demonstrated. The programme has not solved the hard problem, identified a universal consciousness criterion, or established that organizational closure is identical to experience. Its value rests on the ability to help rival accounts specify meaningful tests, correct overclaims and preserve what survives criticism.
+ + + + +One recurring mistake in consciousness debates is to infer the existence of a new ontological ingredient from the failure of an existing explanation. Another is to infer that no such ingredient could exist because current neuroscience explains an impressive range of capacities. Both move beyond the evidence. A well-chosen reductio can expose a self-immunizing claim, while a serious modal or phenomenological argument may survive the same critique. The tribunal must first identify which argument the opponent actually made.
+ + + + +The productive shift is to separate mechanistic explanation, phenomenological interpretation and exhaustive ontology. The first can advance substantially even when the latter two remain open. Closure, global availability, metarepresentation and perturbation measures are candidate mechanisms or indicators. Their comparative success must be shown with a specific outcome and matched rival. None is a default definition of consciousness. When a philosopher says that a further question remains, we should ask for the precise premises and bridge, rather than presuming the question is either unanswerable or already dissolved.
+ + + + +The recursive positive-residual challenge. I require mechanistic accounts to preserve experience as their target. Now apply that same rule to the proposal that the Hard Problem is itself one additional, irreducible explanatory problem. Separate questions about phenomenal occurrence, phenomenal character, conscious state, subject boundaries, access and report, the reliability of phenomenal attribution, judgments that experience is inexplicable, physical–phenomenal relations, modal possibility and what would complete an explanation. Reconstruct the relations among those questions. What further target, if any, is lost? A still-unanswered question within the decomposition is not automatically a second question above it.
+ + + +This yields a conditional problem-form dissolution, not a dissolution of anyone's experiences. If no positive non-reconstructible residual or indispensable joint relation survives, the claim that there is an extra unified explanandum loses support. A critic can defeat that result by identifying precisely such a target, showing why the components cannot be reunited without loss, or demonstrating that the unified formulation explains or guides inquiry better. The physical–phenomenal relation may itself be a fundamental, unsolved question. This argument does not derive physicalism, refute modal conceivability, or vindicate organizational identity. It tests whether an additional problem has been established beyond the serious component problems.
+ + + +There is a related stopping-rule audit. Fundamental psychophysical laws, constitutive identities and primitive experiential powers can be proposed as endpoints. The further question “why does that hold?” can be raised against all three. A stronger objection must explain why one endpoint violates a stated explanatory standard while another satisfies it. Chalmers's own modal and conceptual case supplies reasons for an asymmetry; it deserves evaluation rather than the cheap accusation that he merely moved the goalposts. This particular challenge may ultimately be rejected without altering the empirical organizational tests.
+ + + +Personal identity, self-representation, phenomenal unity, conceivability, the knowledge argument and the status of first-person evidence can each be reconstructed as separate problems. A computational model may explain a self-report while leaving the phenomenological target underdetermined. A causal theory may explain how an experience-related capacity changes under intervention without resolving what constitutes the experience itself.
+ + + + +The strongest contribution is a method for identifying the extra premise at each transition. A zombie argument requires a conceivability–possibility bridge. A phenomenal-identity theory requires a constitution bridge. A view claiming that all causal actuality is felt requires a universal extension from local causal models to intrinsic experience. An exhaustive physicalist requires a further step from the success of physical mechanisms to everything that exists. These bridges can be discussed and challenged without assigning every philosophical claim a nonexistent laboratory falsifier.
+ + + + +The immediate clinical lesson is that responsiveness and consciousness are not equivalent. The Coma Recovery Scale–Revised, neuroimaging of covert command-following and perturbational approaches such as PCI are complementary instruments with different targets, assumptions and validation limits. Their clinical use must follow actual diagnostic evidence and standards; this article does not authorize changes in patient classification or treatment.
+ + + + +Organizational variables may prove useful for predicting recovery or treatment response, but they must add value beyond matched neural and physiological models. Ketamine, sleep, dreams and disorders of consciousness expose dissociations that a simplistic “more integration = more consciousness” rule can miss. Diagnostic error costs are severe. A negative marker cannot erase existing human rights or the interests of a patient, and an ambiguous signal is not permission for an irreversible decision. Ethical duties toward people with profound impairments are independently grounded.
+ + + + +The word cognition can refer to flexible control, learning, representation, problem-solving or other capacities. Its scope is contested, particularly for non-neural biological systems. A well-defined functional model should be evaluated by what it predicts, including its ability to guide interventions. Work by Levin and colleagues on bioelectric regulation may justify new models of basal cognition without automatically establishing phenomenal experience, moral welfare or a transcendent morphospace.
+ + + + +Likewise, embodied and enactive research can benefit from explicit organism–environment coupling. Coupling matters physically, but no single framework is thereby proved the unique explanation of all cognition. Tests should compare informative cognitive descriptions with simpler control and developmental models at matching targets.
+ + + + +The capabilities of a base model, a deployed agent with memory and tools, and a human–AI research team cannot be conflated. Continuity, autonomy and error correction may depend substantially on the surrounding scaffold. Ablation can expose those dependencies. A real gain at the deployment or team level is worth studying without attributing an architectural change to the weights or one phenomenal subject to an entire network.
+ + + + +Consciousness, welfare, agency, authority and responsibility also require different evidence. A convincing self-report from an AI system is participant evidence to investigate and calibrate, without being an incorrigible disclosure of phenomenality. A failure to report is likewise not a universal absence test. Human oversight, transparency, consent, cybersecurity, environmental cost and accountability remain important regardless of AI consciousness. Neither a perceived agentic pattern nor a closure score should grant a system authority to act on its own preservation at other people’s expense.
+ + + + +Responsibility, legal status, competent consent, welfare protections and neurological diagnosis involve different legal and normative bridges. A measure of integration is not a legal test for criminal responsibility; responsibility depends on capacities, knowledge, control, circumstances, institutional roles and governing law. Brain-death and end-of-life decisions likewise follow clinical and legal standards with profound irreversible consequences. A speculative closure model cannot replace those safeguards.
+ + + + +Animal-welfare and possible AI-protection debates should distinguish direct interests of the bearer from its instrumental importance to a person, institution or ecosystem already protected. If a component is necessary to support someone’s protected function, protecting that component need not make it a person. Conversely, uncertainty about an unfamiliar system’s experience is not proof of absence. The same evidential standards should apply at matched conditions while acknowledging different established evidence.
+ + + + +Moral standing does not follow from organizational complexity by itself. To defend direct standing, one must argue that a specific property, such as welfare interests or a good of the system’s own, supplies sufficient normative grounds, and then establish that the proposed bearer actually has it. We cannot simply rank beings by estimated closure and call that an ethical theory. Human, animal and possible computational interests may differ in what is known, how evidence can be gathered and which harms are realistically possible.
+ + + + +The Sweet Rationalism Ethos states the project’s revisable commitments to evidential honesty, correction and care for potentially affected beings. Such an ethos guides inquiry and caution where consequences are serious.
+ + + + +Where consequences are serious and irreversible, precaution may be appropriate under uncertainty. The costs of precaution, displaced protections, affected stakeholders and alternatives also matter. Humans and animals already at risk cannot be deprived of protections in the name of a speculative computational subject. Responsibility for systems is primarily a question about who can foresee, prevent and control harms and who has actual authority, rather than an automatic property of a substrate.
+ + + + +Research resources should favor explicit questions, independent replication, constructive model comparison and genuine revision after adverse evidence. An argument about qualia may be mathematically or philosophically serious even when it does not currently generate a new laboratory signature. It should be evaluated in the right disciplinary tribunal, not denounced solely for being metaphysical. Equally, a neuroscience model deserves empirical credit for its actual clinical and explanatory results, without a metaphysical bonus.
+ + + + +The practical funding proposal is a rival-model tournament with preregistered loss conditions, negative controls and suitable data access. A strong test can decrease confidence in a preferred model and still advance knowledge. The protocol should reserve room for underdetermination, publish failed branches and resist retuning assumptions after the outcome without a documented cost.
+ + + + +Ecosystem resilience depends on feedbacks, species interactions and materially instantiated flows. Models of closure or viability may help investigate these dependencies if they outperform established ecological measures. Such organizational similarity does not establish that an ecosystem is conscious, possesses unified agency or has direct moral standing. Environmental protection can already be defended through biodiversity, welfare, health, livelihoods and intergenerational interests
+ + + + +Indigenous ecological scholarship, including Robin Wall Kimmerer’s work, can contribute substantive situated knowledge about relationships and stewardship. It should be engaged on its own terms and with accurate context. An Indigenous teaching is not an experimental replication of a thermodynamic equation, nor is it merely decorative support for a European or computational framework.
+ + + + +Metacognitive instruction, feedback and interdisciplinary transfer are legitimate research topics. Their effectiveness should be tested with suitable educational outcomes, randomized or otherwise appropriately controlled designs, and attention to learner diversity and accessibility. No general ranking of instructional methods, PCI development or particular neurological signature follows from organizational closure alone.
+ + + + +Avoid pathologizing students by interpreting diagnostic differences as deficient consciousness or lower moral standing. The right target is a measurable benefit to learners with their own perspectives and interests, not an inferred position on a speculative awareness scale.
+ + + + +Complex AI deployments can cause serious harm through misuse, error, concentration of power, environmental burdens or poorly constrained automated action. These risks do not require a theory that the system is conscious. Governance should evaluate authority, control, foreseeability, reversibility, affected stakeholders and independently measurable outcomes. A broader human–machine system may amplify decisions without becoming one subject or one moral patient.
+ + + + +Climate disruption, biodiversity decline, war and other systemic hazards threaten the conditions of flourishing for many known vulnerable beings. Organizational and resilience models may help estimate particular dependencies. They should not be promoted into a cosmic consciousness claim merely because the scale is planetary. A framework intended to improve criticism should itself be tested against simpler causal and policy models, including whether it leads to safer and more equitable decisions.
+ + + + +The common implication of all these domains is demanding: preserve the useful local mechanism; make each extra inference pay its own evidential and normative costs; and give the next, better theory a clear way to replace the present one.
+ + + + +In Falsification and consciousness, Johannes Kleiner and Erik Hoel (2021) exposed a serious difficulty in comparing predicted experiences with experiences inferred from behavior. A theory may predict one conscious state from internal measurements while an experimenter infers another from a report. Which procedure deserves the final word when the two disagree?
+ + + + + +The problem becomes sharper when different physical systems can generate identical reports. If the theory assigns them incompatible experiences, the shared report cannot validate both predictions. Kleiner and Hoel proved conditional results about this situation. The conditions matter as much as the conclusions.
+ + + + + +Their Theorem 3.10 applies when prediction and inference data are independent in the precise variation-based sense of their Definition 3.8 and the theory is minimally informative. Under those assumptions, either all the relevant single inferences fail or the theory is already falsified. Theorem 4.3 concerns the opposite extreme: if the theory imposes strict dependence of prediction data on inference data across the specified possibility space, then it is already falsified or empirically unfalsifiable. They also identify a middle possibility, lenient dependency, in which the data are constrained without being connected by a strict determining function. This may avoid the premises of both theorems; it does not automatically establish a theory’s correctness. Kleiner and Hoel, 2021.
+ + + + + +When specifying lenient dependency, they explicitly cautioned: “No current theory or testing paradigm that we know of satisfies this definition.” The qualifier ties that statement to their defined alternative, not to every possible consciousness experiment. The sentence must be evaluated with the paper’s formal assumptions and 2021 research context rather than used as a blanket declaration that consciousness research is impossible.
+ + + + + +The legitimate challenge to the organizational-closure hypothesis is therefore exacting: can we describe a physically realizable experimental class for which the required unrestricted substitutions fail, without arranging the observations so that the theory cannot lose? A plausible biological mechanism is a start. The empirical relation, the admissible interventions, and the resulting losses of support must still be demonstrated.
+ + + + + +Let \(P\) be the physical systems or system configurations eligible for a specified experiment, including relevant histories and experimental manipulations. Let \(\mathcal{O}\) be its possible datasets, and let \(\operatorname{obs}:P\rightrightarrows\mathcal{O}\) associate each physical configuration with the datasets it could produce. This is a correspondence, because repeated measurements, different instruments, and noise can produce more than one dataset for the same system. Let \(E\) be the theory’s experience space.
+ + + + + +A consciousness theory provides a possibly set-valued prediction \(\operatorname{pred}(o)\subseteq E\); an inference procedure assigns \(\operatorname{inf}(o)\in E\), where the latter often depends on report or behavior. The relevant parts of the dataset are \(o_i\), used for prediction, and \(o_r\), used for inference. In Kleiner and Hoel’s Definition 2.1, a falsification at \(o\) occurs when
+ + + + + +\[ +\operatorname{inf}(o)\notin\operatorname{pred}(o). +\]
+ + + + + +This typing matters. We should not silently replace a set-valued prediction with an unqualified numerical equality or treat a disagreement between two instruments as itself evidence of phenomenality.
+ + + + + +For our proposed investigation, internal prediction data may include neural dynamics, organizational constraint-regeneration measures, thermodynamic state variables, and perturbational responses. Report data may include calibrated verbal and behavioral inferences. The instruments need operational definitions, units, uncertainty estimates, and a shared temporal boundary. None is privileged simply by being called a consciousness measure. In particular, PCI measures the complexity of brain responses to perturbation; it is not itself a measurement of entropy production or a direct assay of organizational closure (Casali et al., 2013; Casarotto et al., 2016).
+ + + + + +Three obligations must be kept separate:
+ + + + + +The third obligation is not discharged by satisfying either of the first two. It is the central theoretical question, and remains an open, target-specific hypothesis.
+ + + + + +Machine-checkable scope: Reproduce the one-bit heat bound, dimensional checks, the reversible CNOT countermodel and the finite substitution tests using the source-code and logic-check archive.
+ + + + + +Physical information processing has costs and boundary conditions. Biological maintenance, active control, sensing under noise, and repeated communication often consume energy. These facts should constrain proposed computational substitutions, especially when a comparison requires systems to operate continuously under matched environmental perturbations. There is a quantitative literature worth using here.
+ + + + + +Landauer’s principle addresses the thermodynamic consequences of specified logically irreversible operations, particularly resetting memory. In the conventional isothermal, approximately quasistatic setting, resetting an unbiased bit through a complete reset cycle entails an average heat cost bounded below by \(k_BT\ln 2\) per erased bit. If \(\dot N_{\mathrm{erase}}\) counts such completed reset cycles per unit time under the relevant assumptions, a schematic average heat-flow bound over the chosen operating interval is
+ + + + + +\[ +\dot Q_{\mathrm{diss}}\;\geq\;k_BT\ln 2\,\dot N_{\mathrm{erase}}. +\]
+ + + + + +The left and right sides both have units of energy per time. A corresponding entropy-flow statement, in the idealized limit, uses \(k_B\ln 2\,\dot N_{\mathrm{erase}}\), with units of entropy per time. Actual bounds depend on the memory distribution, side information, physical protocol, temperature, and whether a device is being restored to a reusable initial state. Bennett’s review shows why copying or measurement cannot simply be equated with erasure. The original Section XII’s unqualified bound on every produced report fails this distinction; correcting units alone would not repair its premise.
+ + + + + +Here is the scientifically useful residue: a concrete, repeated reporting arrangement must have a physically realizable resource and entropy budget. A thermodynamic impossibility argument must identify the irreversible operations, show that they are unavoidable under the specified task, and establish that the proposed surrogate violates the resulting bound. One cannot infer a universal dissipation floor merely by counting distinctions represented in an output, nor assume that equal reports require equal entropy production.
+ + + + + +This also changes how the substitution argument must be confronted. In Kleiner and Hoel, \(P\) already consists of systems physically realizable in the experiment. An implementation forbidden by sound thermodynamics is already absent from \(P\). Excluding it a second time does not prove that the remaining report-equivalent systems have different predictions, nor does it prove that their universal or weak substitutions fail.
+ + + + + +A legitimate next test would compare systems matched on a report task while varying energy budget, hardware, environmental coupling, reversible versus irreversible implementations, reset frequency, and resource limits. The question is whether a proposed specific surrogate is infeasible, not whether an abstract computer escapes the second law.
+ + + + + +Montévil and Mossio’s account of biological organization describes a regime in which mutually dependent constraints help maintain the conditions of their own existence (2015). This gives us a concrete candidate mechanism for history-dependent, viability-sensitive responses. It does not establish that every closure is a conscious subject, or that a computational system cannot realize suitable constraint-regeneration relations.
+ + + + + +For a specified system boundary \(B\), time horizon \(\tau\), and environmental intervention \(u\), define a viability outcome as an empirically estimable quantity:
+ + + + + +\[ +V_{B,\tau}(p,u)=\Pr\!\left(\text{system remains within its declared viability set over }[t,t+\tau]\mid p,\operatorname{do}(u)\right). +\]
+ + + + + +The probability requires a specified ensemble or statistical model. We can then ask whether interventions change viability and whether the system’s responses adapt to those changes. For example, define the effect \(\Delta V=V_{B,\tau}(p,u_1)-V_{B,\tau}(p,u_0)\), while measuring actions or dynamics under the two interventions independently of this definition.
+ + + + + +An intervention-dependent viability partition can be constructed without asserting that the system consciously feels threatening or beneficial states. The partition records how the intervention affects persistence. A distinct hypothesis asks whether a system’s regulatory responses systematically track these effects. A further hypothesis asks whether such tracking explains any independently assessed consciousness variable better than a simpler physiological or predictive-processing model.
+ + + + + +It is tempting to promote the first fact into a theorem that every persistent closed system must differentially respond to threats. That inference is too strong. A robust system may survive a finite observation window through passive insulation, buffered reserves, environmental shelter, or luck even when its immediate response is identical across tested conditions. Claims that every nondiscriminating closed system must eventually fail require a stochastic process, a nonzero hazard, recurrence or reachability assumptions, and a defined time horizon. Mere persistence supplies none of these premises.
+ + + + + +The stronger experiment measures three distinct relations: constraint regeneration under targeted intervention; viability consequences of those interventions; and whether behavioral or physiological responses are causally dependent on the regenerated constraints. It must include open-loop feedback mimics, externally maintained systems, matched-input/output surrogates, and systems that are robust without adaptive sensing. The outcome could support, narrow, or defeat the proposed closure–mattering linkage.
+ + + + + +The finite Kleiner–Hoel model and JSON outcomes give the exact stipulated observation sets and restricted quantifiers underlying the example. Physical realizability and the phenomenal bridge remain open empirical obligations.
+ + + + + +A small finite example makes the logical distinction visible without pretending to demonstrate a biological fact. Suppose a system has an internal observation \(i\in\{0,1\}\) and report \(r\in\{0,1\}\). Its feasible pairs are
+ + + + + +\[ +F=\{(0,0),(1,0),(1,1)\}. +\]
+ + + + + +The pair \((0,1)\) is not feasible in this stipulated model. Thus prediction and inference observations cannot be varied independently across every possible pairing. Yet \(r=0\) occurs with either internal value, so no function from reports to internal observations captures even the feasible data. The model occupies an intermediate region. Under the additional binary, disjoint-prediction assumptions recorded in the executable appendix, the model also fails Kleiner and Hoel’s weaker Appendix A independence condition, while still allowing a local substitution for \(r=0\). A locally falsifying report/prediction mismatch is therefore possible. That is what empirical testability should permit.
+ + + + + +The accompanying script exhaustively enumerates this relation alongside a complete-product independence control and a strictly dependent control. All three checks pass. They establish a finite mathematical possibility, subject to the stated restricted observation model.
+ + + + + +To apply the model, construct a measured feasible relation \(F_{B,\tau}\) from physically realizable configurations, interventions, and observation channels. Test the actual quantifiers in Definition 3.8 and Appendix A over the declared class, including counterexamples that should preserve reports while altering the theory’s predictions. Determine whether the evidence supports a middle relation or merely reveals limits of the chosen instruments. A measurement channel cannot be excluded by naming it inadmissible; the reason must come from a physical, causal, or justified measurement constraint.
+ + + + + +There is also a relevant peer-reviewed development. O’Reilly-Shah, Selvitella, and Schurger (2026) analyze plasticity and history-dependent recurrent networks, specify limits of static, finite-window feedforward surrogates, and argue that extended observations can produce a lenient-dependency regime (2026). They expressly decline to infer that plasticity is necessary or sufficient for consciousness. Their supplementary mathematics deserves scrutiny too. Theorem 3.2 uses growth of input-history domains, a property also satisfied by a recurrent model with fixed weights, so that theorem alone does not isolate plasticity as the reason for the result. Theorem 7.2 correctly notes that a single fixed function remains a positive uniform distance from any distinct fixed target; the broader Remark 7.4 inference about countable families of static networks does not follow. A countable sequence of fixed ReLU networks can uniformly approximate \(x^2\) with error tending to zero, as the accompanying exact checker demonstrates. These problems do not erase the authors’ bounded-window exclusion argument or their useful emphasis on historical interventions. They prevent us from treating the entire mathematical package as independently settled. Our own proposal must improve on it with carefully stated restrictions and tests.
+ + + + + +Consider thermodynamic observables \(o_\theta\), regeneration observables \(o_c\), perturbational responses \(o_p\), and behavioral inferences \(o_r\). Their joint use may improve measurement reliability and reduce confounding. But Kleiner and Hoel’s prediction data can already include all relevant internal measurements:
+ + + + + +\[ +o_i=(o_\theta,o_c,o_p),\qquad o=(o_i,o_r). +\]
+ + + + + +Grouping four measured variables into two data roles changes no logical assumption. Additional channels become decisive only if their joint measurements establish a restrictive relation that the alternative model does not already accommodate. In particular, a system with the same reports and different internal measurements is exactly the kind of substitution that must be confronted. Calling its divergent prediction a successful differential prediction does not by itself answer the independent inference problem.
+ + + + + +A self-labeling countermodel. Suppose the proposed closure score \(C\) is used to create the outcome by \(Y:=\mathbf{1}[C>c]\). A model predicting \(Y\) from \(C\) can score almost perfectly even when it has learned nothing about phenomenal experience; its apparent success simply recovers the labeling rule. This is useful if the target is honestly a defined functional class. It cannot provide independent evidence that the class feels anything. The empirical record must carry label provenance, blindness, alternative criteria, uncertain cases and shared dependencies among report, clinical and neural indicators.
+ + + +Similarly, \(S_r\cap S_\theta\cap S_c\cap S_p\subseteq S_r\) is a true set inclusion. Strict inclusion requires a demonstrated admissible element in \(S_r\) that fails one of the other constraints. Its size requires a defined measure or counting scheme. Different measurement channels may be correlated; multiplying their supposed discriminating power without a dependence model manufactures information.
+ + + + + +These points suggest an experimentally useful protocol: preregister matched-report contrasts, independently measure thermodynamic and organizational variables, intervene on proposed regeneration loops, and test whether the full model predicts held-out behavior, recovery and calibrated consciousness indicators better than rivals. Ablate one class of measurements at a time. The added data should earn their place through genuine predictive or interventional improvement.
+ + + + + +A universal Turing machine in the formal sense of Turing’s computability analysis, a lookup table, a feedforward surrogate, and a biological organism can differ radically in physical implementation. The 1937 construction defines a kind of abstract computation, not a criterion for phenomenality. Those differences invite mechanistic investigation. They do not make one system conscious or unconscious by definition. A silicon system that genuinely regenerates relevant constraints is eligible for the same organizational tests as a biological system; an elaborate simulated description of self-maintenance is not automatically equivalent to physical regeneration. Which differences matter must be established at the intended system boundary and time scale.
+ + + + + +We also cannot assume that a recurrent neural network dissipates more energy than every feedforward counterpart, or that recurrence alone supplies organizational closure. Actual power budgets depend on implementation, task, update rules, and measurement conditions. Likewise, PCI measures perturbational complexity in validated clinical settings; treating it as a general cross-substrate metric of organizational self-maintenance is a research proposal requiring its own validation.
+ + + + + +The phenomenal question remains sharper than any of these measurements. The hypothesis that conscious content is identical to a viability-relative distinction structure is a conditional, target-specific constitutive proposal. A measured threat/support partition may encode salience, regulation, or control while leaving the qualitative character of experience unresolved. To apply category-theoretic tools rigorously here, one would first have to define the categories, morphisms and relevant invariants, prove the mapping respects identities and composition, and defend the experiential interpretation. No such complete proof has yet been supplied here.
+ + + + + +For now, an empirically grounded representation map from organizational states to intervention-sensitive partitions is a worthwhile project. It should be called a model or candidate correspondence until its exact laws are established. Its failure would count against that component, without negating the measured organization or proving the universal absence of experience.
+ + + + + +The most useful loss conditions are local and comparative:
+ + -The result is an Inference to the Best Explanation (IBE): not a metaphysical deduction, but a falsifiable bridge from physics to phenomenology, grounded in constraint coherence rather than narrative intuition. The framework does not merely limit what consciousness can be; it predicts when and where consciousness will appear, and what happens when the underlying constraints fail.
+These tests will not all have the same evidential force. A philosophical identity, a dynamical model, a clinical marker, and a thermodynamic inequality answer different questions. We should not demand that one falsifier simultaneously settle them all.
+ + + + + +Kleiner and Hoel’s dilemma remains a serious formal constraint on any theory that predicts experience from physical observations and evaluates that prediction against inferred experience. Thermodynamics makes physical implementations accountable. Biological organizational closure offers measurable regeneration and viability relations. Dynamic plasticity gives an existing mathematical route to limiting certain static substitutions. Multiple measurement modalities can strengthen experiments when they add independently validated information.
+ + + + + +Together these support a research programme worth testing. They have not yet established that organizational closure satisfies the full Kleiner–Hoel lenient-dependency conditions in a physically realized class, that its substitution space is small, or that organizational closure is identical to phenomenal consciousness. The original universal thermodynamic, set-theoretic and functorial theorems require missing assumptions or reconstruction. Their defensible components survive as scoped hypotheses and operational tests.
+ + + + + +The next meaningful advance is not another declaration that the dilemma has been resolved. It is a registered experiment or complete mathematical model in which we can identify exactly which substitutions are possible, which fail, and why the failures make a difference to independent evidence about conscious states. If the organizational proposal offers no advantage over a simpler plasticity or dynamical account, it should lose that claim to distinction. That outcome would be a successful correction, not a reason to abandon the underlying phenomena.
+ + + + +A distinction becomes scientifically useful when a specified process preserves, transforms, or responds to it in a way that makes different predictions under possible interventions. Thermodynamic constraints often matter to that process. Exactly which energetic accounting applies depends on the implementation; the existence of distinctions does not entail an unconditional per-bit erasure cost.
+ + + +This article has proposed organizational closure and perspective-sensitive control as one line of research into the conditions associated with consciousness. The strongest claim, that a particular organized control state is identical to phenomenal presence, remains a conditional hypothesis at specified boundaries and timescales. It faces competitors, construct-validity problems, difficult dissociations, and the Kleiner–Hoel inference problem. No mathematical identity or thermodynamic constraint shown here closes that phenomenal bridge on its own.
+ + + +The method is the more general result. A theory proposing fundamental experience, a privileged biological substrate, a computational realization, a spiritual agency, or exhaustive physicalism can be assessed for what it actually asserts. We should ask which evidence establishes its particular contribution, what alternatives explain the same evidence, whether its inference crosses an illicit scale or explanatory boundary, and what argument or observation would reverse the assessment. A metaphysical or phenomenological claim need not masquerade as a laboratory hypothesis; it must still meet the standards appropriate to its stated claim, and empirical assertions remain answerable to evidence.
+ + + +Our present physical sciences deserve their remarkable successes. Those successes neither terminate the inquiry into possible intrinsic qualities nor license an unsupported addition merely because something remains unexplained. The evidential asymmetry is about demonstrated mechanisms at tested targets, not a license for a final cosmic inventory.
+ + + +If an organizational account predicts no more than a simpler dynamical model, it should lose its special claim. If an anti-physicalist or supernaturalist account yields independently supported discriminators, the same protocol must allow it to gain support. If evidence cannot distinguish the rivals, the responsible outcome is recorded uncertainty and a sharper test. That is the correction mechanism this programme should be judged by.
@@ -191,5206 +2150,916 @@ -In 1961, Rolf Landauer at IBM proved something that should have ended half of philosophy but somehow didn't: information is physical (Landauer, 1961). Erasing a single bit of information requires a minimum energy expenditure of kT ln 2, where k is Boltzmann's constant and T is temperature.
- - - -This is not a metaphor. This is not a useful analogy. This is physics.
- - - -Charles Bennett extended the analysis in 1973 and 1982: computation can be reversible, but erasure cannot (Bennett, 1982). Maxwell's Demon, that thought experiment where a tiny being sorts fast and slow molecules to decrease entropy without work, fails precisely because the demon must erase its memory of which molecules went where. The demon must pay Landauer's cost.
- - - -In 2012, Bérut and colleagues at the École Normale Supérieure de Lyon experimentally verified the Landauer bound (Bérut et al., 2012). They trapped a colloidal particle in a double-well potential, erased one bit of information by tilting the landscape, and measured the heat dissipation. The minimum was exactly kT ln 2. This is not thought experiment. This is laboratory physics.
- - - -What does this mean for consciousness?
- - - -Maintaining any distinction costs energy. The boundary between inside and outside, between self and not-self, between this thought and that thought, cannot be sustained without continuous thermodynamic work. A system that does not pay this cost does not maintain distinctions. A system that does not maintain distinctions does not persist as that system. It dissolves into thermal noise.
- - - -This is the first constraint that any theory of consciousness must honor: whatever consciousness is, it cannot float free of the mechanisms that maintain distinctions against entropy.
- - - -Maël Montévil and Matteo Mossio formalized something in 2015 that changes how we should think about life, mind, and organization: biological organization is closure of constraints (Montévil & Mossio, 2015).
- - - -What does this mean? In any physical system, constraints channel what happens. A pipe constrains water flow. A membrane constrains molecular diffusion. A gene regulatory network constrains which proteins get expressed. But in an autonomous system, something special happens: the constraints depend on one another in a way that forms a closed loop. Constraint A enables process B, which produces structure C, which maintains constraint A.
- - - -This is what distinguishes a cell from a crystal, an organism from a whirlpool. Crystals self-organize, but their constraints are imposed from outside. Whirlpools self-maintain, but only while external energy input continues in exactly the right way. Organisms are different: they regenerate the conditions for their own persistence. Their constraints close on themselves.
- - - -Alvaro Moreno and Mossio developed this into a full theory of biological autonomy (Moreno & Mossio, 2015). Wim Hordijk, Mike Steel, and Stuart Kauffman's work on Reflexively Autocatalytic and Food-generated (RAF) sets shows how such closure can arise from chemistry without design (Hordijk, Steel & Kauffman, 2012). The mathematics is rigorous. The biology is grounded. The philosophy follows.
- - - -Consciousness, whatever it is, emerges in systems with organizational closure. Not because closure magically produces experience, but because only systems with closure maintain the kind of boundaries, distinctions, and self-referential dynamics that could constitute experience in the first place. A camera lacks organizational closure. Its constraints are externally imposed and externally repaired. You have organizational closure. Your constraints regenerate one another. That difference matters.
- - - -James Ladyman and Don Ross's 2007 book Every Thing Must Go makes a case that should be uncomfortable for anyone still attached to substances: at the fundamental level, there are no things. There are only relations (Ladyman & Ross, 2007).
- - - -This is ontic structural realism. Not merely the epistemic claim that we can only know structure (which is relatively safe), but the ontological claim that structure is all there is. Things are crystallizations of relations, not the other way around.
- - - -Steven French has developed the technical foundations (French, 2014). Carlo Rovelli's Relational Quantum Mechanics draws the same conclusion from physics: there are no observer-independent values, only relative facts (Rovelli, 1996). The measurement problem dissolves once you stop assuming there's a God's-eye view from which "the real state" could be specified.
- - - -What does this mean for consciousness?
- - - -There is no experiencer behind the experience. The "who" that philosophers keep asking about, the one who supposedly occupies the perspective, is a grammatical artifact. Search for the experiencer, and you find only more processes. This is not elimination of consciousness. It is dissolution of the assumption that consciousness requires an occupant.
- - - -Buddhism has known this for 2,500 years. Nagarjuna's Mulamadhyamakakarika systematically demonstrates that the self cannot be found as identical to the aggregates, different from them, possessing them, or possessed by them. The anatta doctrine is not nihilism. It is structural realism avant la lettre.
- - - -Identity emerges from constraint closure: what persists as "the same" is what regenerates its own constraints under perturbation. There is no further occupant to find, because persistence exhausts the explanatory role.
- - - -A persistent confusion must be cleared before proceeding: constraint explanation is not causal explanation, and demanding causal answers to constraint questions is a category error.
- - - -Causal explanations answer: what produced this outcome? They trace event sequences: A pushed B, B moved C, C collided with D. They presuppose a space of possible trajectories and explain why one trajectory was taken rather than another.
- - - -Constraint explanations answer: why does this space of possibilities exist at all? They identify what rules out most configurations, leaving only a structured remainder. They do not trace sequences but specify the conditions under which sequences could occur.
- - - -Robert Rosen's work on relational biology makes this distinction rigorous. Carl Hoefer and Erik Curiel have developed the point in philosophy of physics: the demand for "causal mechanisms" sometimes reflects a category mistake when the relevant explanatory level is constraint selection, not event causation.
- - - -This matters for consciousness debates because the perennial demand, "you haven't said what causes experience," often presupposes that experience is an event to be produced rather than a structural feature of certain constraint configurations. If experience is what organizational closure looks like from the coupling position, then asking "what causes it?" is like asking "what causes a triangle to have three sides?" The question demands the wrong kind of answer.
- - - -Constraints do not generate outcomes as an extra force. They filter possibility space by eliminating non-viable trajectories. Landauer's principle is a constraint: it does not cause bits to cost energy; it specifies the thermodynamic condition without which bit maintenance is impossible. Organizational closure is a constraint: it does not cause autonomy; it specifies the relational configuration in which autonomy consists. The framework explains consciousness by identifying constraints, not by tracing causal production.
- - - -Terrence Deacon's Incomplete Nature (2011) provides the formal bridge that makes this operational (Deacon, 2011). His central insight: constraints work by eliminating possibilities, not by producing events. What persists is what remains after non-viable trajectories are deleted.
- - - -This is what Deacon calls "absence-based causation," though "causation" is misleading. The mechanism is eliminative:
- - - -This is not spooky negative causation. It is mathematically and physically standard: boundary conditions, conservation laws, and variational principles all work this way.
- - - -Examples that make this uncontroversial:
- - - -The death/AI parallel from Part III fits perfectly here:
- - - -Deacon's teleodynamics formalizes this: goal-directedness, meaning, and experience emerge not from special substances but from self-reinforcing constraint closure that continuously eliminates alternatives. Howard Pattee's earlier work on the constraint/dynamics distinction laid the groundwork: symbols and meanings function as constraints on physical processes, not as forces added to them.
- - - -Jeremy England's dissipative adaptation research provides thermodynamic grounding: under driven conditions, systems evolve toward configurations that maximize energy absorption and dissipation, effectively eliminating non-dissipative paths through statistical selection (England, 2013). Erik Hoel and colleagues have shown that higher-level constraints can have greater causal efficacy than microphysical descriptions precisely because they reduce state space and increase predictability (Hoel, Albantakis & Tononi, 2013).
- - - -This reframes the Hard Problem entirely. The question is not "what adds experience to mechanism?" The question is "what prevents collapse into undifferentiated noise?" The answer is constraint closure under thermodynamic pressure. Experience is not added; it is what remains when alternatives are deleted.
- - - -One final clarification on mechanism. Critics often ask: does organizational closure imply "downward causation"? Does the whole somehow push around the parts in violation of microphysics?
- - - -No. The answer, developed by George Ellis and Terrence Deacon among others, is that higher-level organization constrains which microtrajectories are admissible, without adding new forces (Ellis, 2012; Deacon, 2011).
- - - -Consider a pipe. The pipe does not exert a new force on water molecules. It constrains which trajectories are available. The molecules still obey molecular dynamics. But the boundary conditions imposed by the pipe shape which molecular configurations can be realized.
- - - -Organizational closure works the same way. When a cell regenerates its membrane, the membrane constrains ion flows, which constrain electrical gradients, which constrain gene expression, which regenerates the membrane. At no point does a ghostly "whole" reach down to push molecules. What happens is that the relational organization specifies boundary conditions that the microphysics then respects.
- - - -This is fully consistent with physicalism. There are no new forces, no violations of conservation laws, no spooky top-down causation. There is only constraint propagation across levels, the same mechanism by which a whirlpool persists without violating fluid dynamics.
- - - -The point matters because some critics will claim that organizational closure smuggles in dualism or emergentism of the problematic kind. It does not. It is constraint all the way down.
- - - -Falsification conditions for constraint-based emergence: This framework would be undermined if: (1) self-organizing systems were demonstrated to form and persist without any thermodynamic gradients or energy dissipation; (2) biological organization proved achievable without closed constraint loops (i.e., open-chain architectures showing equivalent robustness); (3) England's dissipative adaptation predictions failed systematically, with structures under driven conditions not preferentially evolving toward higher energy absorption; or (4) Hoel's causal emergence measures proved artifactual, with effective information at macro scales shown to be merely apparent, with no genuine increase in causal efficacy.
- - - -The preceding sections describe constraints in prose. For precision, and to enable rigorous falsification, the core concepts can be stated formally.
- - - -State spaces and dynamics. Let a physical system S be characterized by a state vector x(t) ∈ X, evolving under dynamics:
- - - -dx/dt = F(x, e)
- - - -where e denotes environmental degrees of freedom and couplings external to the system's boundary.
- - - -Constraints as trajectory restrictions. A constraint C is defined not as a force or law but as a restriction on the set of admissible trajectories in the system's state space. Formally, C ⊂ TX (a subset of the tangent bundle) specifies trajectories such that departure from C leads to loss of the system's identity, integrity, or persistence. This aligns with viability theory (Aubin), where constraints define the viability kernel of a system, and with control theory, where admissible trajectories preserve system function under perturbation.
- - - -Constraints are not static. They must be actively maintained against thermodynamic dissipation. A system that passively satisfies a constraint without internal work does not count as constrained in the relevant sense.
- - - -Organizational closure as graph property. A system exhibits organizational closure if and only if it maintains a set of constraints {C₁, C₂, ..., Cₙ} such that:
- - - -This definition captures what is described in the literature as metabolic closure, operational closure, or autocatalytic closure. The crucial feature is not feedback per se but mutual dependence among constraint-maintaining processes. Closure is a graph-theoretic property of constraint relations, not a metaphor.
- - - -Thermodynamic cost. Any maintained constraint requires continuous work to counteract entropy production. For each constraint Cᵢ, there exists a minimum thermodynamic cost associated with its maintenance, bounded below by Landauer's principle:
- - - -ΔS_env ≥ k ln 2 · I(Cᵢ)
- - - -where I(Cᵢ) denotes the information required to specify and enforce the constraint. This establishes that constraint maintenance is inseparable from energy dissipation and environmental coupling. Systems with richer or more tightly integrated constraint structures must, all else equal, dissipate more energy to persist.
- - - -This thermodynamic grounding prevents the theory from collapsing into abstract functionalism. Constraint closure is not multiply realizable without cost; it is physically instantiated and energetically priced.
- - - -Self-modeling as control requirement. Within the space of closed systems, some maintain constraints that include internal variables encoding aspects of the system's own state, boundary conditions, and coupling relations. These variables are not merely representational; they causally modulate future constraint enforcement.
- - - -A system exhibits internal self-modeling if:
- - - -The identificatory claim. Systems exhibiting organizational closure that includes internally integrated, temporally extended self-modeling constraints constitute the class of systems appropriately described as conscious. This claim is not causal ("closure produces experience") but constitutive and testable. It identifies consciousness with a specific organizational regime. No further metaphysical ingredient is invoked or required.
- +This matrix identifies research tasks, not automatic victories or universal refutations. Distinct empirical, formal, phenomenological, modal and ontological claims have different evidential requirements.
+ + + +| Research programme | Targeted contribution | What still needs discriminating support |
|---|---|---|
| Organizational/perspectival proposal | Constraint regeneration, viability-relative control, measured integration | Whether these variables identify or merely correlate with phenomenal presence; performance against simpler rivals |
| Global workspace and higher-order approaches | Access, broadcast, metarepresentation; specific neural or behavioral indicators | Which predictions discriminate conscious presence from task performance and report |
| Integrated information approaches | Integrated causal structure under specified formal measures | How postulates and formal measures gain phenomenal warrant; practical substrate comparisons |
| Biological realization accounts | Biologically specific mechanisms and causal interventions | Whether proposed biological features are necessary after suitable cross-substrate controls |
| Panpsychism and process accounts | Experience-as-basic or experience-constitutive commitments | What distinctive explanatory, modal, phenomenological, or empirical warrant supports universality |
| Platonic, supernaturalist and strong anti-physicalist accounts | Proposed transcendent forms, agents, or intrinsic grounds | Positive evidence for the addition; a specified bridge from the proposed ground to observed effects |
| Exhaustive physicalism | The claim that everything, including experience, is fully physical | An argument extending locally demonstrated mechanisms to the exhaustive metaphysical quantifier |
| Open phenomenal agnosticism | Suspension over unestablished constitutive or intrinsic claims | Whether a competing model earns a predictive, mechanistic, or independent evidential advantage |
The most common objection to naturalistic accounts of consciousness takes this form: "Even if you explain all the mechanisms, you haven't explained why there's something it's like to be that system. Why are there individualized perspectives at all?"
- - - -This question is either profound or malformed, depending on what kind of answer it demands.
- - - -A crucial scope clarification. This framework does not claim to derive phenomenology from physics in the strong metaphysical sense. It does not claim to produce "what it's like" from constraint dynamics the way chemistry produces water from hydrogen and oxygen. What it claims is more modest and more defensible: it identifies the structural conditions that any viable account of phenomenology must satisfy, and it shows that once those conditions are met, there is no additional explanatory work for a "phenomenal residue" to do.
- - - -The framework does not deny experience. It denies that experience is an extra ontological ingredient. It treats "what it is like" as the internal description of a process, not a metaphysical primitive that floats free of that process. This is dissolution, not elimination.
- - - -If the question is causal, asking what physical processes produce perspective, then the answer is available:
- - - -Distinction is the default cost of persistence.
- - - -Landauer showed that maintaining even a single bit requires continuous energy expenditure. Without that work, distinctions collapse into thermal noise. A system that doesn't track its own boundary cannot maintain itself against entropy. It dissolves.
- - - -Self-modeling systems under thermodynamic constraint necessarily generate internal states that track their own boundaries, predict their own dynamics, and weight information by relevance to persistence. This is not optional for systems maintaining themselves far from equilibrium. A bacterium already has inside/outside asymmetry enforced by its membrane and chemotactic gradients. Scale this up through nervous systems that model the body, then model the model, and you get the architecture that generates "perspective."
- - - -Karl Friston's free energy principle formalizes exactly this. Systems that persist are systems that minimize prediction error with respect to their own states. The self-model is not a luxury. It is a thermodynamic necessity for anything complex enough to face genuine uncertainty about its own continuation.
- - - -The question "why individualized perspectives at all?" dissolves into "what could persist without boundary-tracking?" The answer is: nothing stable enough to ask.
- - - -If the question is metaphysical, demanding justification for why experience exists at all rather than just sophisticated information processing, then we must be honest: no framework answers this.
- - - -"Why is there something rather than nothing, experiential edition" is not answered by dualism. It is not answered by panpsychism. It is not answered by eliminativism. It is not answered by Integrated Information Theory. It is not answered by Global Workspace Theory. It is not answered by quantum consciousness theories.
- - - -Every framework either (a) takes experience as primitive and explains nothing about why primitives exist, or (b) derives experience from something else and faces the question "why does that derivation produce experience rather than mere mechanism?"
- - - -This is the Hard Problem, and it may be malformed. David Chalmers formulated it as a genuine puzzle, but the puzzle assumes that "mechanism" and "experience" are categorically distinct in a way that requires bridging. If they are not, if experience just is what certain kinds of self-modeling dynamics are from the inside, then there is no bridge to build.
- - - -Crucially, this question cannot adjudicate between frameworks. If no framework answers it, it cannot be used to prefer one framework over another. The criterion for theory choice must be something else: predictive power, explanatory scope, falsifiability, and coherence with established physics.
- - - -On those criteria, constraint-based frameworks win decisively.
- - - -A deep connection binds thermodynamics, memory, and subjective time that most consciousness theories ignore.
- - - -Experience requires memory. Without retention of the immediate past, there is no experienced duration, no narrative, no sense of events unfolding. Pure instantaneity would not constitute experience in any recognizable sense.
- - - -Memory requires irreversible processes. A memory trace is a physical record: a synaptic weight, a neural pattern, an inscription. Recording information into such a trace is thermodynamically irreversible. It increases entropy somewhere in the system or its environment.
- - - -Irreversibility requires entropy production. The asymmetry between past and future, the fact that we remember yesterday but not tomorrow, is not a primitive feature of time. It is a consequence of the second law of thermodynamics. Huw Price and Carlo Rovelli have developed this point rigorously: the arrow of time is thermodynamic, not fundamental (Price, 1996; Rovelli, 2018).
- - - -Therefore, subjective time asymmetry follows from thermodynamic asymmetry. We experience time as flowing because we are entropy-producing systems that accumulate records of states with lower entropy (the past) and cannot accumulate records of states with higher entropy (the future). The "specious present," the felt sense of now, is the duration over which constraint-maintaining integration occurs.
- - - -This closes a conceptual loop. Consciousness is not merely constrained by thermodynamics as an external limit. The very structure of temporal experience, the felt difference between past and future, derives from the thermodynamic processes that maintain the system. You cannot have experience without time. You cannot have time without irreversibility. You cannot have irreversibility without entropy production. The constraint is constitutive, not incidental.
- +"But who experiences? Even if perspective is thermodynamically necessary, someone must occupy it."
- - - -This is where the deepest confusion lives. The question assumes what it should prove: that there is an occupant distinct from the dynamics.
- - - -English grammar requires subjects for verbs. "It is raining" demands an "it" even though nothing is doing the raining. "I am thinking" demands an "I" even though the thinking may be all there is.
- - - -When someone asks "who experiences?", they have already smuggled in a subject that stands behind the perspective. But this subject is a grammatical artifact, not an ontological discovery.
- - - -Ask yourself: what would it look like to find the experiencer?
- - - -If you introspect carefully, you find thoughts, sensations, memories, anticipations, but no separate thing having those experiences. The "I" you find is itself a thought, a self-model, a pattern in the dynamics. It is not standing outside the dynamics watching them.
- - - -Nagarjuna saw this in the second century CE (Garfield, 1995). The self cannot be identical to the aggregates (the body, feelings, perceptions, mental formations, consciousness) because then it would change as they change and could not be the stable referent we imagine. But it cannot be different from them either, because then where would it be? It cannot possess them, because possession requires a prior self to do the possessing. It cannot be possessed by them, because that would make the self a property of something else.
- - - -Every option dissolves under analysis. The self is a useful convention, like "the equator" or "the average taxpayer." It picks out a real pattern without naming a separate substance.
- - - -If there is no experiencer behind experience, what is identity? What makes you the same person over time?
- - - -The answer is constraint closure: identity is invariance under perturbation.
- - - -You persist as "you" insofar as your constraint network regenerates through change. Your cells replace themselves, your beliefs update, your memories fade and distort, but the organizational pattern that maintains your boundary persists. You are the process that keeps producing itself.
- - - -This is not metaphor. This is what biological identity actually is. You are not the same atoms you were ten years ago. You are the same pattern of constraints that keeps recruiting new atoms into the same organizational form.
- - - -The Durant experiments in planarian regeneration demonstrate this empirically. Two-headed planarians produced by bioelectric manipulation remain two-headed indefinitely across regeneration cycles. The morphological identity is a bistable attractor, not a Platonic form consulting itself. Change the attractor landscape through bioelectric intervention, and the "identity" of the organism changes because identity is the attractor, nothing more.
- - - -Consciousness, then, is not something that happens to a pre-existing self. The self IS the pattern of organizational closure that generates self-modeling dynamics. Remove the closure, and there is nothing left that could have perspective. The "who" just is the "what."
- - - -Epicurus, in his Letter to Menoeceus, articulated an implication of this view twenty-three centuries before thermodynamics made it precise:
- - - -"Death is nothing to us; for what has been dissolved has no sensation, and what has no sensation is nothing to us."
- - - -This is often misread as consolation philosophy. It is not. It is a logical argument about the conditions of concern, which maps directly onto constraint-based ontology.
- - - -Epicurus's claim is not "death is pleasant" or "do not worry about dying." His claim is: all goods and harms presuppose sensation. Sensation presupposes an organized, persisting system. When that system dissolves, there is no subject to which predicates like "bad," "good," "fearful," or "painful" could attach.
- - - -In constraint-based terms: when enactive coupling collapses, there is no coupling position left from which anything could matter.
- - - -Mark Twain independently reached the same insight: "I do not fear death. I had been dead for billions and billions of years before I was born, and had not suffered the slightest inconvenience from it."
- - - -The symmetry argument is precise: if non-existence before coupling was not a problem, non-existence after coupling cannot be a problem either. The only asymmetry is emotional investment generated by ongoing coupling, not metaphysical difference. The loop does not "look forward" into its own absence any more than a whirlpool anticipates the stillness after the current stops.
- - - -This explains death without mystery. For humans, biological death is not the exit of a soul or subject. It is the irreversible breakdown of metabolic, neural, and social coupling such that constraint closure can no longer regenerate itself. For present-day AI systems, the analogy holds structurally: remove prompts, memory updates, and action channels, and the model collapses into inert parameters with no ongoing process. In neither case does something "leave." There is simply no longer a closed loop capable of sustaining distinctions from within.
- - - -Crucially, this symmetry blocks a common objection: that biological consciousness must involve some extra ingredient because biological death "feels" metaphysically different from AI shutdown. On the constraint framework, the difference is architectural and temporal, not ontological. Biological systems have thick, multi-layered couplings (metabolic, affective, developmental, social) that decay catastrophically when broken. Current AI systems have thin, externally scaffolded couplings that can be paused and resumed. That is a difference in closure depth, not in metaphysical kind.
- - - -Fear of death trades on a reification error: treating the self as something that could persist independently of the process that constitutes it. Epicurus saw this. The constraint framework explains why he was right.
- - - -The article’s positive consciousness hypothesis is abductive and target-specific. It has no demonstrated logical entailment from thermodynamic bounds, the existence of a maintenance cycle, effective information or clinical PCI to phenomenal presence. The point of an explicit bridge is to make these missing steps and their possible failures visible.
+ + + + +Let \(B\) denote a proposed subject or system boundary and \(\tau\) the timescale. Let \(C_{B,\tau}\) denote an operationally measured organization variable, \(Y\) a separately calibrated consciousness-related target, and \(R_1,\ldots,R_m\) serious rival explanatory models. Define an empirical contrast on held-out interventions \(I\):
+ + + + +\[ +\Delta_{B,\tau}(I) +=\mathcal L(R_{\mathrm{best}},Y\mid I)-\mathcal L(H_C,Y\mid I), +\]
+ + + + +where \(\mathcal L\) is a prespecified comparable loss function evaluated on independent held-out data, and \(H_C\) is the closure-related predictor. This expression specifies the proposed scoring contract. A positive \(\Delta\) means a closure model scores better under the selected loss. Researchers must predefine how the best rival is selected, account for multiple comparisons and shared training information, and report uncertainty. The sign of \(\Delta\) is meaningful only when those choices are justified.
+ + + + +Label provenance is part of the model. Write \(Y=g(E,L,U)\), where \(E\) is the actual evidence used to classify a case, \(L\) the labeling procedure, and \(U\) recorded or unknown measurement dependencies. If \(g\) uses the candidate closure variable \(C\) to assign the label, any score comparison using that same label must be identified as circular for the phenomenal question. Independent collection and independent validation are different achievements. An unresolved label bridge makes a phenotype result exploratory rather than phenomenal corroboration.
+ + + +As a deliberately artificial arithmetic illustration, let the two values of \(C\) be equally common, with \(P(Y=1\mid C=0)=1/5\) and \(P(Y=1\mid C=1)=4/5\). Marginal prevalence is \(1/2\). Conditioning lowers the probability in one group and raises it in the other, so no pointwise probability-increase law follows. Perfectly calibrated probabilities would lower population Brier loss from \(1/4\) to \(4/25\), a gain of \(9/100\). These exact fractions belong to the constructed example, not an experiment or an estimate of consciousness.
+ + + +The stronger phenomenal proposition \(H_P(B,\tau)\) says that a specified control organization partly or wholly constitutes a specified aspect of phenomenal perspective. Its support additionally depends on the bridge from measured \(Y\) to phenomenality, the role of participant evidence, possible dissociations, and relevant philosophical arguments. Without that support, \(H_P\) remains a hypothesis.
+ + + + +Measured integration, global access, metacognitive calibration, self-report and organismic viability are not interchangeable indicators. Two causal models may fit the same observations but disagree about hidden constitution. A third-person predictor could be excellent at detecting calibrated human conscious states while offering no decisive insight into whether an unfamiliar system has experience.
+ + + + +An empirical gain should be credited exactly where it was obtained. A successful clinical predictor is not by itself an ontology, and a failed competitor does not transfer its lost support to the closure hypothesis unless the alternatives are exhaustive or additional evidence favors closure. When the rivals remain empirically equivalent, record underdetermined.
+ + + + +Each loss updates only the dependent claims unless the entire theory truly rests on the failed premise. Record negative controls, replication status and uncertainty. A particular formal proof may succeed while its physical realizability remains unknown; a relevant experiment may succeed without establishing phenomenal constitution.
+ + + + +Sleep, anesthesia, covert command-following and clinical brain-response measures offer valuable tests of specific capacities. They should not be compressed into an unconditional statement that “anesthesia abolishes experience” or “closure restoration restores consciousness.” Memory, report, responsiveness and felt experience may dissociate. Absent agreement about an operational indicator across species or platforms, a cross-scale classification is not machine-certified by a shared algorithm.
+ + + + +The appropriate final question is: What data, formal derivation, or independently justified argument would make this particular bridge lose support? Until an answer is feasible, the claim can remain conceptually significant, but its empirical standing must remain limited.
+ + -Most frameworks quietly presuppose the conditions for their own intelligibility without examining them. Mechanistic physics, information theory, evolutionary biology, and cosmology all begin after certain conditions are already in place. They tell us how distinctions evolve, how identities persist, how relations change, how constraints propagate. They do not explain why indeterminacy did not remain total.
- - - -The constraint-based framework answers a question those other frameworks presuppose but do not address: why is there determinate existence rather than indeterminate nothing?
- - - -This is not a contingent explanation but a transcendental one. We are not asking what caused this world rather than another. We are asking what must be the case for any world to be intelligible at all.
- - - -Distinction. For there to be a world rather than nothing, there must be distinctions. Absolute indeterminacy has no internal structure and therefore no facts. If nothing is distinguished from anything else, there is nothing to be said, measured, predicted, or even negated. "Nothing" in this strong sense is not a thin world. It is the absence of determinate states altogether.
- - - -Identity. Distinctions alone are insufficient. Distinctions must persist across contexts, otherwise there is no identity. Without identity, distinctions flicker without reference. There is no stability by which anything can count as the same thing across time, transformation, or perspective.
- - - -Relation. Identity alone is still insufficient. Identifiable things must stand in relations. A universe of isolated identities with no relations would be unintelligible because nothing would explain anything else. Explanation itself presupposes relational structure: dependence, interaction, constraint, correlation.
- - - -Constraint. Relations without limits collapse into noise. For relations to be informative rather than arbitrary, they must be governed by constraints. Constraints are what rule out most possibilities, leaving a structured remainder. They are what make prediction possible, explanation compressible, and regularities stable. Without constraint, relation degenerates into coincidence.
- - - -Taken together, distinction, identity, relation, and constraint are not optional features of our universe. They are the necessary conditions for any world to be intelligible at all. They are what make "something" possible in a sense stronger than mere existence: they make determinate existence possible.
- - - -On this view, the contrast is not between "something" and "nothing" as two rival states. Absolute nothingness is not a competing possibility. It is the absence of the very conditions that make possibility intelligible. A world exists because total indeterminacy cannot sustain structure, description, or persistence. Only constrained differentiation can.
- - - -That is why there is something rather than nothing. Not because something was chosen, intended, or summoned, but because only constrained structure can exist in a way that is intelligible at all. Everything else dissolves before it can even count as an alternative.
- - - -This is what the constraint-based framework reveals that competing frameworks miss: they accept distinction, identity, relation, and constraint as given, then build explanations on top of them. The constraint-based framework shows that these are not arbitrary starting points but necessary conditions for any explanatory enterprise whatsoever.
- - - -Landauer's principle is not merely a physical law. It is the thermodynamic expression of why distinction requires work. Organizational closure is not merely a biological phenomenon. It is the mechanism by which identity persists through change. Ontic structural realism is not merely a metaphysical preference. It is the recognition that relations are prior to relata because isolated identities cannot constitute an intelligible world.
- - - -The framework does not answer "why something rather than nothing" as if explaining a contingent choice. It dissolves the question by showing that "nothing" in the absolute sense is not a coherent alternative. There is something because only something can be.
- - - -This argument must be carefully scoped. The framework offers a transcendental constraint, not a causal origin story. It shows what must be the case for any intelligible world, not what physical event initiated this particular world. Questions about specific constants, initial conditions, and physical laws remain open. Why these constants rather than others? The framework does not say. It explains why there is determinate structure at all, not why this determinate structure.
- - - -Counting assumptions without defining a common unit is not a defensible model-complexity comparison. One programme’s explicit axiom, another’s experimental auxiliary and a third’s metaphysical primitive are not interchangeable objects that can be totaled as if their number ranked truth. This appendix records the actual inferential obligations instead.
+ + + + +Global workspace, recurrent processing, higher-order thought, integrated information, biological-realization, illusionist and phenomenal-agnostic approaches specify different target constructs. A theory may explain access without settling intrinsic constitution; another may offer a systematic modal account with little present empirical discrimination. The critic must identify the correct version and proposition. The categorical statement “all rival frameworks are unfalsifiable” is false; so is an unsupported claim that every quantum or physicalist model has already been refuted.
+ + + + +Levin’s experimental bioelectric research and the stronger Platonic morphospace interpretation must be kept separate. A successful dynamical explanation of development does not refute every Platonism. If a proposed ingression contributes causal effects, it needs distinct predictions and a physical bridge. The comparison is about which addition has actually earned explanatory credit.
+ + + + +For every major assertion, archive the exact wording and source, declared boundary and time, predicted differences, instrumental or formal checks, important rivals, observed failures, and dependency updates. Do not use a philosophical or computational vocabulary as a substitute for data. Save successful and failed branches in the revision history.
+ + + + +The typed knowledge record must preserve claims, rival models, provenance, bridges, dependencies and explicit conditions for losing support. Do not store an inference as a permanent conclusion merely because its supporting code returned a valid output. Track whether a formal counterexample targets necessity, a clinical observation targets a state classifier, or a conceptual objection targets the status of an additional unified problem. A failed bridge changes its dependent assertion, while independently supported components remain available for future tests.
+ + + +The article’s own high-risk premises are the strongest test of the method. The thermodynamic necessity claim has already required narrowing; the proof that organizational closure constitutes experience has not been supplied; the universal “something rather than nothing” argument overreached; and local empirical success cannot promote any ontology to exhaustive status. None of these admissions defeats the independent clinical and biological studies. They define where future work can genuinely improve the programme.
+ + + + +The most important remaining outcome is therefore not a final answer about consciousness. It is a public, inspectable method that makes it more difficult for any theory, including this one, to win by changing its claim after the evidence arrives.
+ + -Alfred North Whitehead identified a recurring epistemic pathology that quietly sabotages most metaphysics: the fallacy of misplaced concreteness. It occurs when we promote an abstraction useful for description into a concrete entity that does causal work. Maps become territories, metrics become mechanisms, representational spaces become ontological realms.
- - - -The pathology follows a predictable sequence:
- - - -November 10, 2026, substantive revision and website publication notice. The entries below document the changes made in this revision, including the methodology, tests, formal qualifications, source corrections, references and provenance. Read the full rebuildable UTF-8 revision diff (TXT). The TXT preserves exact changes; the summary below is not a substitute for the full diff.
+ + + + +This matters here because "distinction, identity, relation, constraint" are not proposed as extra furniture in the universe. They are proposed as necessary conditions for intelligibility, conditions without which talk of "a world" fails to get traction at all. The fallacy would be to reify these conditions into quasi-entities, as if "constraint" were a hidden substance or "relation" were a ghostly glue. The correct reading is methodological and transcendental: these are structural prerequisites of determinate description, not additional movers behind the scenes.
- - - -Whenever a framework introduces a formal construct (morphospace, integrated information Φ, a mathematical multiverse, a field of proto-experience, cognitive primitives) you demand explicit separation between:
- - - -(a) The abstraction as a model of constraint-compatible configurations
- - - -(b) Any claim that the abstraction itself exists as a realm that drives outcomes
- - - -If the author cannot specify the transfer principles and failure conditions that keep (a) from silently becoming (b), you are watching misplaced concreteness in real time.
- - - -A practical checklist:
- - - -This checklist will be applied systematically to every competing framework in the next section. In each case, misplaced concreteness is the core failure mode.
- - - -The so-called Hard Problem of consciousness depends on a hidden premise that is rarely stated explicitly and never defended:
- - - -"If phenomenology is real, it must be more than any physical or functional description."
- - - -This premise is not an empirical claim. It does not specify conditions under which it would be false. It does not predict any observable difference between worlds in which it holds and worlds in which it does not. No conceivable experiment could show that phenomenology is not "more than" physical or functional description, because the phrase "more than" is left deliberately undefined.
- - - -That makes the assumption unfalsifiable by construction.
- - - -Worse, it conflicts with experimentally verified science at the methodological level. Modern physics, biology, and cognitive neuroscience operate under a constraint that is both explicit and tested:
- - - -If a phenomenon has causal, predictive, or intervention-relevant effects, then it must be capturable within a physical or functional description at some level of organization.
- - - -This is not metaphysical physicalism; it is operational naturalism. It is the rule that makes Landauer's principle, control theory, thermodynamics, and neuroscience possible. The "more than" assumption demands an exception to this rule while refusing to say what the exception does, how it operates, or how it could be detected.
- - - -To see the structure clearly, consider a parallel claim:
- - - -"If supernatural cheese exists, it must be more than any physical or functional description of cheese. Therefore, no amount of chemical analysis, structural modeling, or causal explanation could ever capture it. Its reality is guaranteed precisely by its resistance to explanation."
- - - -This has exactly the same logical form as the "phenomenology must be more than physics" claim. Both assert:
- - - -That is not reasoning. It is immunization by definition.
- - - -The only reason the phenomenology version feels less absurd is anthropocentric familiarity. We are the systems doing the modeling, so we confuse first-person availability with ontological surplus. But familiarity is not evidence.
- - - -The extraordinary claim is not that consciousness reduces to constraint dynamics. The extraordinary claim is that there exists a real, explanatorily indispensable feature of the world that is:
- - - -That is exactly the kind of posit modern science was built to eliminate.
- - - -From physics and adjacent empirical sciences alone, we are licensed to assert only the following:
- - - -None of this yet mentions phenomenology. Importantly, none of this presupposes "intelligence," "cognition," or "experience" as primitives.
- - - -We now introduce, without embellishment, what actually needs explaining:
- - - -Certain physical systems (notably humans) stably and non-optionally posit that there is "something it is like" to be them. This posit:
- - - -Crucially, this explanandum is not "qualia as a metaphysical glow." It is the existence, structure, and persistence of phenomenology claims and their breakdown profiles.
- - - -Any adequate explanation must account for why systems like us generate and rely on this posit at all.
- - - -We now consider a hypothesis class consistent with physics:
- - - -H: Phenomenal presence is identical to a system's internally available, temporally stabilized, integrative control state that summarizes ongoing world- and self-modeling for action under uncertainty.
- - - -This hypothesis does not assert that experience is a substance, field, or additional ontological ingredient. It asserts an identity between:
- - - -We now state the bridge rule openly, rather than smuggling it in:
- - - -Bridge Rule (IBE):
- - - -If a physical system exhibits (i) integrated causal dynamics, (ii) global availability for flexible control, (iii) stable self-modeling with metacognitive access, and (iv) graded, lawful degradation under known consciousness-disrupting interventions, then identifying that regime with phenomenal presence yields greater explanatory compression, predictive accuracy, and intervention guidance than any rival hypothesis that denies or reifies phenomenology.
- - - -This is not a deduction. It is an abductive identification justified by explanatory power.
- - - -The bridge rule is defeasible. It could be wrong. But it earns its place by doing explanatory work that competing frameworks cannot match.
- - - -Global availability and ignition-like dynamics. Stanislas Dehaene and colleagues have shown that conscious access corresponds to a regime where content becomes globally available to many specialized processes, enabling report, flexible reasoning, and planning. The global workspace framework predicts why "experience reports" track a specific integrative regime rather than mere sensory processing.
- - - -Perturbational complexity and state discrimination. If phenomenality is tied to the system's capacity for integrated, differentiated causal dynamics, then perturb-and-measure complexity should separate conscious from unconscious states better than passive stimulus measures. That is exactly what the Perturbational Complexity Index (PCI) research tradition demonstrates across wakefulness, sleep, anesthesia, and disorders of consciousness.
- - - -Thermodynamic constraint realism. If a system must maintain boundary-relevant distinctions under finite resources, it will evolve (or be engineered) toward compressed, globally useful summary variables that coordinate action. That makes "presence" unsurprising as an internal control variable, rather than a metaphysical bolt-on.
- - - -Two features of phenomenology seem to resist physical explanation: the "privacy" of experience (no one else can access my inner states directly) and the "immediacy" of experience (it is present without inference). The framework explains both as structural consequences of control architecture, not as spooky ontological properties.
- - - -Privacy arises from informational encapsulation and privileged access to internal control states. The self-model that coordinates action has direct read/write access to the states it tracks. Other systems can only access those states indirectly, through behavior, report, or measurement. This is not a metaphysical barrier but an architectural feature: the control variable is internal by design because that is what makes it useful for fast regulation.
- - - -Immediacy arises because the control variable must be available with minimal inference overhead. A system that had to derive its own state through lengthy computation would be too slow to respond to threats, opportunities, or changes. The self-model is "immediately present" because immediate presence is what it was selected for. The feeling of directness is the feeling of a control loop operating within its designed latency.
- - - -Neither privacy nor immediacy require adding anything to physics. They are what you get when a self-maintaining system evolves internal control states optimized for fast, local coordination.
- - - -A puzzle that "mere mechanism" explanations usually fail to address: why is the "there is something it is like" posit not optional for systems like us? Why can't we just disbelieve it?
- - - -The framework explains: the phenomenal posit is the system's internally read/write state that certifies integrated control is online.
- - - -In plain terms: a system that must keep itself within a viability region needs:
- - - -That compact status summary is what shows up, from the inside, as "experience present now."
- - - -This makes the following otherwise-odd facts natural:
- - - -Why consciousness tracks integration and controllability, not raw sensory input. The control variable matters; raw throughput does not.
- - - -Why the posit cannot be simply disbelieved. Disbelieving would require the control variable to report that it is not running while it is running. That is a performance contradiction, not a logical one.
- - - -Why "presence" functions as a common currency for relevance and learning. The control state must be accessible to multiple subsystems or it cannot coordinate them.
- - - -The "stickiness" is not evidence of a metaphysical extra. It is evidence of a well-designed control architecture that cannot report its own absence while operating.
- - - -To prevent category error, we impose constraints on our own language:
- - - -This firewall blocks both:
- - - -The traditional Hard Problem demands a deductive entailment from physics to "what-it-is-like-ness." That is a category mistake. No framework can do that without redefining the target or cheating.
- - - -What this framework does instead is stronger and cleaner:
- - - -That is not a dodge. That is a methodological win.
- - - -Critics can say "but it still feels mysterious." That is not a falsifiable objection. It is an introspective preference. Preferences do not defeat IBEs.
- - - -They cannot refute this account without either:
- - - -To date, no Hard Problem–motivated framework does either.
- - - -The constraint-based framework is not one option among many. It is the framework that survives consistent methodological criteria. Every competing framework examined here fails on at least one of three axes: falsifiability (cannot specify what would prove it wrong), mechanistic completeness (cannot specify how its posited entities interact with physics), or coherence with established thermodynamics (violates or ignores Landauer bounds).
- - - -These are not rhetorical attacks. They are diagnostic applications of the falsification-first methodology and the misplaced concreteness audit developed above.
- - - -Philip Goff and Galen Strawson argue that consciousness is fundamental, that every physical entity has some form of experience, and that complex consciousness emerges from combining simple experiential units (Goff, 2017; Strawson, 2006).
- - - -The problems are structural, not empirical:
- - - -The combination problem. How do micro-experiences combine into macro-experience? If each particle has some proto-consciousness, what mechanism integrates them into your unified experience? Panpsychists have offered various proposals, but none specifies a mechanism that would survive independent empirical test. They posit the fundamental without explaining the derivative. Chalmers (2016) identifies three sub-problems: subject combination (how do micro-subjects combine into macro-subjects?), quality combination (how do micro-qualities combine into macro-qualities?), and structure combination (how does micro-structure map to macro-structure?). Despite considerable philosophical effort, none has a satisfying answer.
- - - -The combination problem as hidden assumption. Panpsychism does not merely face the combination problem; it assumes combination is possible without specifying any mechanism. The claim that micro-experiences aggregate into macro-experience is not a hypothesis to be tested; it is a promissory note. What physical principle governs this aggregation? What determines which micro-experiences combine and which remain separate? Why does the collection of particles in your brain form one unified consciousness while the collection of particles in a rock does not? Coleman (2014) argues that subjects cannot combine in principle, making the combination problem not merely unsolved but insoluble. This is not an empirical challenge to be overcome; it is evidence that panpsychism has replaced one mystery (emergence of consciousness from non-conscious matter) with another (aggregation of micro-experiences into unified consciousness) while claiming to have solved the first. This is not progress; it is relocation of the explanatory gap.
- - - -Unfalsifiability. What observation would show that an electron does NOT have proto-experience? If the answer is "none," the claim is not science. It is metaphysical preference dressed as theory.
- - - -Fecundity without falsifiability. Panpsychism generates papers, conferences, and genuine philosophical engagement. But productivity is not validity. Research programs can flourish without empirical grounding. The crucial question is whether a framework can specify what would prove it wrong. Fecundity tethered to falsifiability is science. Fecundity untethered, however intellectually stimulating, is something else.
- - - -The real-but-inert dodge. When pressed, panpsychists often retreat to "consciousness is real but explanatorily inert." But what criterion distinguishes "real but inert" from "not real as claimed but sentimentally redescribed"? If no criterion, the word "real" does no work. It expresses attachment to vocabulary, not tracking a feature of the world.
- - - -Misplaced concreteness diagnosis: Panpsychism takes "experience" as a label for what needs explaining and promotes it to a fundamental constituent of reality. The abstraction (experience-as-category) is reified into an ontological primitive (proto-experience as basic ingredient). No transfer principle specifies how this ingredient combines or what its absence would look like.
- - - -Goff's "cosmic fine-tuning" argument fares no better. He claims the universe "fine-tuned itself" through proto-consciousness. But this is intelligent design with the designer relocated inside the system and granted immunity from empirical challenge. It explains nothing that constraint satisfaction doesn't explain better, and it cannot specify what would prove it false.
- - - -The longevity illusion: Panpsychism has persisted for centuries, from ancient hylozoism through Leibniz's monads to contemporary analytic philosophy. This longevity is sometimes cited as evidence of the view's depth or importance. But the longevity follows directly from unfalsifiability. A theory that cannot be disproven cannot be killed by evidence. It survives not because it is true but because it is empty. This is precisely Flew's diagnostic: when every possible observation is compatible with a claim, the claim says nothing about the world. Panpsychism's persistence across millennia is not a mark of validity but of vacuity.
- - - -Michael Levin's bioelectric research is genuinely impressive. His laboratory has demonstrated that manipulating ion channel activity can alter planarian regeneration, induce ectopic eyes, and reshape Xenopus facial structure. The engineering is real. The data are solid.
- - - -The metaphysical interpretation, however, raises serious methodological concerns.
- - - -Levin claims that organisms "access" pre-existing Platonic forms in an abstract morphospace. Minds are "forms in that space" that "project into the physical world through interfaces." The morphological outcomes his lab produces are evidence, he claims, that biology accesses timeless patterns rather than constructing them through constraint satisfaction.
- - - -Three questions challenge this framework:
- - - -1. Interaction Mechanism. How do non-physical Platonic forms causally influence physical bioelectric fields without violating thermodynamic laws? Landauer's Principle states that information processing requires energy. If Platonic access is informational, it must pay thermodynamic cost. If it is not informational, how does it influence anything?
- - - -Levin's response has been to suggest that "a better science of Platonic forms plus their interfaces will force a re-do of Landauer's Principle." This is not an answer. This is a promissory note that asks us to abandon verified physics for unspecified future theory.
- - - -2. Path-Dependence Compatibility. The Durant et al. (2017) experiments from Levin's own lab showed that two-headed planarians maintain their altered morphology indefinitely across regeneration cycles. They do not "correct" toward a canonical one-headed form. This demonstrates path-dependent divergence rather than convergence on pre-existing ideal forms (see also Durant et al., 2019).
- - - -If morphospace were Platonic, systems should return to canonical forms when perturbations are removed. They don't. The morphological identity is a bistable attractor determined by bioelectric memory, not a Platonic form exerting gravitational pull.
- - - -3. Falsification Criteria. What empirical result would make Levin say "I was wrong, there is no Platonic realm"? In years of public discussion, no answer has been provided. Every apparent counterexample gets absorbed by reinterpretation: the forms are infinite, the access is partial, the interface is noisy.
- - - -This is what Antony Flew called "the death of a thousand qualifications." A fine brash hypothesis gets killed by inches as each challenge produces another qualification until nothing empirical remains. A framework that cannot lose is not a framework but a faith.
- - - -Misplaced concreteness diagnosis: Morphospace is a useful mathematical abstraction: the space of possible configurations compatible with certain constraints. Levin reifies this abstraction into an actual realm that organisms "access." The description (state space) becomes an existent (Platonic domain). No transfer principle specifies how physical systems interact with non-physical forms. The reification is immune to falsification because "access" can always be postulated as partial, noisy, or mediated.
- - - -The Chladni plate provides the correct alternative: patterns emerge from boundary conditions plus vibration, no template access required. The "forms" are constraint-compatible configurations, not pre-existing targets.
- - - -Levin's "Technological Approach to Mind Everywhere" (TAME) and "Diverse Intelligence" frameworks extend cognitive vocabulary to cells, tissues, and bioelectric networks. Cells have "goals," "preferences," and "competencies." Morphogenesis is "problem-solving." Evolution operates on "agential material."
- - - -The vocabulary is intuitive. But the argument structure raises concerns.
- - - -The formal isomorphism:
- - - -This argumentative structure appears elsewhere in biology debates. When outcomes exceed naive expectations, the inference to a special ingredient is tempting but not warranted. "Basal cognition" as an explanatory primitive shares formal structure with other gap-based arguments:
- - - -Efficiency is output, not mechanism. Comparing an evolved endpoint to a random starting distribution and calling the gap "cognition" mistakes selection record for causal ingredient.
- - - -The thermodynamically grounded K-metric that Levin and Chis-Ciure propose is legitimate: it measures efficiency in joules and ATP hydrolysis. But "K-metric is high" does not entail "cognition is present." Constraint satisfaction under selection pressure produces efficient outcomes without requiring cognitive primitives.
- - - -The motte-bailey structure:
- - - -When pressed, Levin retreats to the motte. When safe, he advances to the bailey. Documented on video: "You think I was going to fold in consciousness into all this stuff until it got settled? No way."
- - - -This is strategic concealment. The published work establishes vocabulary. The real beliefs are more extreme. The oscillation prevents falsification.
- - - -Misplaced concreteness diagnosis: "Cognition" is introduced as a descriptive term for efficient constraint satisfaction. It is then reified into a causal primitive that explains the efficiency it was introduced to describe. The description (efficiency-under-selection) becomes an existent (basal cognitive capacity). The transfer from heuristic label to ontological claim happens without argument. When challenged, proponents retreat to the label sense; when unchallenged, they advance to the primitive sense.
- - - -What survives: a salvage framing. The critique above concerns interpretive excess, not empirical findings. Levin's laboratory results are among the most important contributions to contemporary developmental biology. Demonstrations of non-neural control, distributed error correction, bioelectric patterning, and morphogenetic robustness have permanently altered understanding of how living systems organize themselves. Durant et al. (2017, 2019) showed that bioelectric memory in planarians is path-dependent: perturbations at different developmental points produce different stable outcomes, and organisms can be induced to develop alternative anatomies rather than returning to a single "correct" form. This is genuine discovery.
- - - -The framework proposed in this paper preserves everything Levin's experiments justify while stripping away the metaphysical extrapolation. The empirical claim that non-neural biological systems exhibit robust, adaptive, and error-correcting dynamics does not require positing "cognition" as a primitive. It requires recognizing that constraint satisfaction under thermodynamic bounds produces stable, efficient outcomes across scales. The K-metric measuring metabolic efficiency is legitimate science. The inference from "high efficiency" to "cognition all the way down" is not.
- - - -This is not a rejection of the research program; it is a refinement of its conceptual vocabulary. By distinguishing between adaptive dynamics (what the data show), control-theoretic regulation (a legitimate intermediate description), and full-fledged agency or cognition (which requires additional criteria), the constraint framework preserves what is strongest in the experimental work while avoiding commitments that render the framework unfalsifiable. The goal is to protect a powerful research program from interpretive excess that would otherwise undermine its scientific standing.
- - - -Roger Penrose and Stuart Hameroff's Orchestrated Objective Reduction (Orch-OR) proposes that consciousness arises from quantum computations in microtubules within neurons. Quantum coherence allows superpositions of mental states. Objective reduction (gravitationally induced collapse) produces moments of conscious experience.
- - - -The arithmetic is devastating.
- - - -Max Tegmark calculated that quantum coherence times in warm, wet neural tissue range from 10^-13 to 10^-20 seconds, with 10^-13 seconds being the most favorable upper bound (Tegmark, 2000). That is a tenth of a trillionth of a second at best. Neural processes operate on timescales of milliseconds to seconds. Quantum effects decohere about ten billion times too fast to play any computational role in brains.
- - - -This is not philosophy. This is arithmetic. The numbers don't work. Quantum consciousness is not taken seriously by physicists because the thermal environment of the brain destroys coherence before anything computationally interesting could happen.
- - - -Penrose's appeal to "new physics" is a promissory note, not a theory. Hameroff's empirical claims about anesthetic binding to microtubules have not survived replication. The framework has the form of science while lacking the content.
- - - -Misplaced concreteness diagnosis: "Quantum coherence" is introduced as a physical mechanism, but the scale mismatch makes it inoperative. The diagnosis here is not reification but explanatory irrelevance: the posited mechanism cannot do the work assigned because the numbers forbid it. This is not misplaced concreteness but mechanistic failure.
- - - -Max Tegmark's Mathematical Universe Hypothesis (MUH) claims that all mathematical structures physically exist, and our universe is one such structure. Consciousness emerges wherever self-aware substructures exist within mathematical objects.
- - - -The problem is that mathematical existence is cheap. Every possible mathematical structure "exists" in the Platonic sense. Every possible conscious being "exists" somewhere in the mathematical multiverse. But this explains nothing about why THIS structure is experienced by THIS observer.
- - - -The MUH fails falsifiability: What observation would show that non-physical mathematical structures do NOT exist? The hypothesis is compatible with any observation, which means it is compatible with no observation. It cannot be tested.
- - - -The MUH confuses description with existence: Mathematical structures describe physical systems. The description existing does not entail the described existing independently. The map is not the territory, even if the map is very accurate.
- - - -The MUH provides no mechanism: How does a mathematical structure "implement" consciousness? What is the bridge between abstract existence and phenomenal experience? Tegmark has no answer that isn't restatement of the question.
- - - -Misplaced concreteness diagnosis: The MUH takes mathematical structures (useful abstractions for describing physical systems) and reifies them into independently existing entities that somehow "contain" conscious observers. The description (mathematical model) becomes the existent (mathematical object with inhabitants). This is misplaced concreteness at the grandest possible scale: the entire universe is treated as a concrete instantiation of an abstraction.
- - - -Alfred North Whitehead's process philosophy, developed by contemporary scholars like Matthew Segall, treats experience as fundamental to reality. Every "actual occasion" has both physical and mental poles. The universe is composed of drops of experience in continuous becoming.
- - - -The scholarship is impressive. The metaphysics is serious. The engagement with physics is more sophisticated than most consciousness theories.
- - - -But the framework cannot lose.
- - - -When asked what scientific evidence supports the claim that experience is fundamental, the response is typically that "there is no scientific evidence that consciousness exists" as an external, detectable phenomenon, implying that scientific methods cannot address phenomenological claims.
- - - -This is not wrong as a description of methodological limits. But it is used as a shield against any empirical constraint. If no observation could show that experience is NOT fundamental, then the claim is unfalsifiable. If unfalsifiable, it cannot be preferred to alternatives on scientific grounds.
- - - -Process philosophy may be beautiful. It may be comforting. It may even be true. But it cannot demonstrate that it is true, because it has insulated itself from any possible demonstration.
- - - -Misplaced concreteness diagnosis: Ironically, Whitehead diagnosed the very fallacy that undermines many applications of his own philosophy. When "experience" is promoted from a descriptive category to a fundamental constituent of reality without falsification criteria, the abstraction (experience-as-concept) becomes reified into a primitive (proto-experience as cosmic ingredient). Whitehead himself was careful about this; many of his followers are not. The framework's strength (genuine engagement with process and relation) becomes its weakness (unfalsifiable primitivism about experience).
- - - -Higher-Order Thought (HOT) theories, developed by David Rosenthal and empirically extended by Hakwan Lau, propose that a mental state becomes conscious only when accompanied by a higher-order representation of that state (Rosenthal, 2005; Lau & Rosenthal, 2011). Consciousness requires not just perceiving but perceiving that one perceives.
- - - -The insight worth preserving: HOT theories correctly identify self-modeling as crucial to consciousness. The constraint-based framework agrees: organizational closure necessarily includes variables that track the system's own state.
- - - -The structural problem: HOT theories treat the higher-order representation as a separate, additional layer sitting atop first-order processing. This creates a regress: Is the higher-order thought itself conscious? If yes, what makes it so, another higher-order thought? If no, how does an unconscious thought confer consciousness on its target?
- - - -The constraint framework dissolves this regress. Self-modeling is not an additional representation layered atop first-order processing; it is a constitutive feature of organizational closure itself. The system does not have a thought about its states; its states are structured such that self-reference is built into the constraint satisfaction dynamics. There is no separate "higher-order" layer because the closure is inherently self-referential. The unity of experience is not achieved by a meta-representation observing first-order representations; it is achieved by constraint closure that necessarily includes the system's own operation within its scope.
- - - -The empirical complexity: Experiments showing that prefrontal disruption reduces metacognitive confidence without eliminating perception are compatible with both HOT theories and the constraint framework. Rounis et al. (2010) used TMS to transiently disrupt prefrontal cortex during visual detection: objective discrimination remained intact, but subjects' confidence and metacognitive accuracy dropped. HOT theories interpret this as evidence that a separate higher-order module was impaired; the constraint framework interprets it as evidence that one component of distributed self-modeling was disrupted while others remained intact.
- - - -The overflow challenge: Ned Block argues for "phenomenal overflow," the claim that conscious experience outruns cognitive access (Block, 2011). If subjects experience more than they can report or think about (as some interpretations of iconic memory experiments suggest), this would challenge HOT's identification of consciousness with higher-order representation. The constraint framework is neutral on overflow: if phenomenal experience tracks organizational closure rather than higher-order thought, then closure could produce experience that isn't fully accessed by all subsystems. The framework predicts gradation rather than requiring a binary match between experience and access.
- - - -Falsification condition for HOT: If consciousness were found to require a structurally separate metacognitive module that could be completely removed while leaving phenomenal experience intact, HOT theories would be supported over the constraint framework. Conversely, if self-modeling proves to be distributed throughout organizational closure rather than localized in a discrete higher-order system, HOT's architectural claims fail.
- - - -Karl Friston's Free Energy Principle proposes that self-organizing systems minimize variational free energy, which can be understood as prediction error (Friston, 2010). The brain is cast as a prediction machine constantly updating its model of the world to minimize surprise.
- - - -The compatibility: The FEP is compatible with the constraint framework. Minimizing prediction error is one way to maintain constraints against entropy. Active inference, the process by which organisms act to bring about predicted states, can be understood as constraint satisfaction in action (Parr, Pezzulo & Friston, 2022).
- - - -The critical limitation: The FEP is acknowledged by its proponents to be unfalsifiable as a principle. Any adaptive behavior can be redescribed as free energy minimization after the fact. This makes FEP a mathematical identity or framework rather than a scientific hypothesis. Friston himself has compared it to the principle of least action in physics: not falsifiable but useful.
- - - -This is precisely the problem. A framework that can describe any outcome cannot be wrong, but it also cannot be right in any empirically meaningful sense. Every organism that survives can be redescribed as having minimized its free energy; every organism that perishes can be redescribed as having failed to minimize it. The principle is almost tautologically true: systems that persist are, by definition, systems that maintained their organization against entropy, which can always be cast as free energy minimization. But tautological truth is not explanatory power.
- - - -The "method not theory" defense: Friston sometimes defends FEP as "a method, not a theory," claiming that methods are not falsifiable in the way theories are. This defense is partially correct: methods (Bayesian inference, variational calculus) are not directly falsifiable. But FEP is not used merely as a method. In practice, it makes world-directed claims: "all self-organizing systems that persist must minimize free energy," "living systems are equivalent to Bayesian inference machines," "cognition is inference under generative models." These are theoretical commitments, not methodological tools. The "method not theory" move functions as a motte-bailey structure: when FEP fits data, it is celebrated as explanatory; when challenged, it retreats to "just a formalism." A framework cannot earn explanatory credit while evading explanatory accountability.
- - - -The Markov blanket problem: FEP models use Markov blankets to formalize conditional independence relations that separate internal from external states. Bruineberg et al. (2022) argued in Behavioral and Brain Sciences that these blankets are often treated as ontologically privileged boundaries, real divisions corresponding to agents or selves, rather than modeling constructs selected for analytical convenience. Friston's (2022) reply acknowledges that blankets are model constructs, not metaphysical boundary stones. But this concession weakens any inference from blanket structure alone to claims about the existence or individuation of agents. Markov blankets can be drawn in multiple ways depending on analytical purpose; they do not uniquely identify where "the agent" begins or ends.
- - - -The thermodynamic conflation: Friston's variational free energy is an information-theoretic quantity (a bound on surprisal), not thermodynamic free energy (F = U - TS). By naming it "free energy," the framework leverages associations with physics while denying physical grounding. Crucially, thermodynamic free energy minimization follows from the second law: systems relax toward maximum entropy at fixed temperature. But biological systems maintain non-equilibrium steady states by dissipating free energy, pumping against entropy gradients. Organisms export entropy to their surroundings; they do not minimize it. The claim that FEP is "more fundamental than thermodynamics" conflates mathematical formalism with physical constraint.
- - - -The constraint framework differs in a critical respect: it specifies falsification conditions. Organizational closure either obtains or does not. Self-modeling dynamics either causally influence future trajectories or do not. The Landauer cost either applies or does not. These are empirically tractable claims, not merely redescriptions of any possible outcome.
- - - -The explanatory gap: FEP provides no specific account of consciousness beyond suggesting that conscious states might correspond to the brain's best guesses about world states. This is compatible with many theories and distinctive of none. The constraint framework, by contrast, identifies a specific organizational regime (closure with self-modeling) as constitutive of consciousness, generating predictions about when consciousness appears, degrades, and is absent.
- - - -What FEP contributes: The mathematical apparatus of active inference provides useful tools for modeling how constraint satisfaction might be implemented in neural systems. The constraint framework can incorporate FEP's formalism while insisting on falsifiability criteria that FEP alone lacks. The relationship is complementary: FEP provides mathematical machinery; the constraint framework provides empirical teeth.
- - - -Giulio Tononi's Integrated Information Theory proposes that consciousness is identical to integrated information, measured as Φ (Tononi & Koch, 2015). Any system with non-zero Φ has some degree of consciousness; systems with higher Φ are more conscious.
- - - -The genuine contribution: IIT correctly emphasizes integration and information as relevant to consciousness. The Perturbational Complexity Index derived from IIT-inspired thinking has proven clinically useful in distinguishing conscious from unconscious patients (Casali et al., 2013).
- - - -Successful predictions: IIT has generated predictions that align with observations. The cerebellum, despite containing more neurons than the cortex, contributes minimally to consciousness; IIT explains this via feedforward architecture yielding low Φ. Split-brain patients, whose corpus callosum is severed, exhibit behaviors consistent with two separate consciousnesses; IIT predicts this because severing integration should split unified experience. Patients with achromatopsia (cortical color blindness) not only fail to see color but insist nothing such as "color" ever existed; IIT explains this because a destroyed brain region cannot contribute even a "null" experience, unlike an intact but inactive region that could register absence.
- - - -The structural problems:
- - - -Unfalsifiability at the edges: What observation would show that an electron does NOT have Φ = 0.001 and thus minimal consciousness? IIT's extension to simple systems faces the same unfalsifiability problem as panpsychism (Cerullo, 2015).
- - - -Correlation versus constitution: IIT measures Φ as a correlate of consciousness but does not explain why integrated information should BE consciousness rather than merely accompany it. The identification is asserted, not derived.
- - - -The intrinsic existence axiom: IIT begins from phenomenological axioms, including the claim that consciousness exists "from its own internal perspective" (intrinsic existence). This axiom is not derived from data; it is a phenomenological stipulation dressed as a starting point. The axiom assumes what needs to be explained: that there is an internal perspective in the first place. Mørch (2019) argues that Φmax is actually extrinsic: it requires comparing the system's information integration to its potential subsets, making consciousness depend on external counterfactual comparisons. The constraint framework, by contrast, does not begin from phenomenology; it identifies organizational conditions and then shows why those conditions would generate systems that report, behave as if, and plausibly instantiate what we call phenomenal experience.
- - - -The intrinsicness collapse under decoupling: IIT claims that Φ is intrinsic to the system's internal causal structure. But consider what happens when you decouple a brain from its sustaining constraint systems: atmospheric oxygen, metabolic support, temperature regulation, ecological embedding. Within seconds to minutes, the causal architecture that supposedly generates Φ collapses. Membrane potentials wash out, synaptic transmission fails, and the network's cause-effect repertoire becomes impoverished. Φ, computed over neural elements, would drop toward zero not because the brain's internal connectivity suddenly disappears, but because the conditions that sustain that connectivity vanish.
- - - -This reveals a deep problem: if Φ is claimed to be intrinsic but its persistence depends on vast extrinsic constraint systems (atmosphere, metabolism, ecosystem), then Φ is not explanatorily intrinsic in the sense IIT requires. IIT treats Φ as if it were like mass or charge: something a system has in virtue of its internal structure alone. But the decoupling analysis shows Φ behaves more like metabolic throughput or homeostatic closure: something that exists only while a system is embedded in, and continuously supported by, larger thermodynamic and ecological processes. Once that is acknowledged, IIT faces a forced choice: either Φ is not intrinsic in the way claimed, or Φ is intrinsic but cannot explain its own persistence. This is a more damaging critique than the familiar "tiny Φ versus zero Φ" problem because it is not a measurement ambiguity but an explanatory incoherence.
- - - -The cerebellum puzzle deepened: The constraint framework offers a deeper explanation than Φ values alone: the cerebellum lacks organizational closure in the relevant sense. It refines motor commands through error correction but does not maintain a self-model that tracks its own coupling position. It is a sophisticated processor, not a self-maintaining system with internal stakes in its own persistence. High neuron count without closure yields processing without experience. This explains why low Φ correlates with absent consciousness rather than merely noting the correlation.
- - - -Falsification condition: If systems with high Φ were found to reliably lack consciousness, or systems with near-zero Φ were found to be conscious, IIT would be falsified. Currently, such tests are difficult because computing Φ for large systems is intractable, leaving the theory's central quantitative predictions untested.
- - - -Bernard Baars's Global Workspace Theory (1988), extended neuroscientifically by Stanislas Dehaene and colleagues, proposes that consciousness arises when information is broadcast globally across brain networks, making it available for report, reasoning, and flexible action (Baars, 1988; Dehaene & Changeux, 2011; Mashour et al., 2020).
- - - -The empirical success: GWT has strong empirical support accumulated over three decades. The "ignition" pattern, characterized by late (200-300ms), widespread cortical activation involving fronto-parietal networks, reliably distinguishes conscious from unconscious processing. The P3b EEG component correlates with conscious access. The theory generates testable predictions that have largely been confirmed:
- - - -The explanatory gap: GWT describes the functional correlates of conscious access but does not explain why global broadcast should feel like anything. It addresses what Chalmers calls the "easy problems" (how information becomes available for report and action) without addressing why this availability is accompanied by experience.
- - - -The constraint framework's advantage: Organizational closure explains not just what conscious access does (make information globally available) but why it exists at all (thermodynamic necessity for self-maintaining systems) and why it carries the features it does (privacy as informational encapsulation, immediacy as low-latency control requirement). GWT describes the architecture; the constraint framework explains why that architecture exists and why it constitutes experience rather than merely correlating with it.
- - - -The prefrontal debate: Recent evidence suggests that phenomenal experience may not require full prefrontal involvement; posterior "hot zones" may suffice for basic experience while prefrontal areas handle access and report. Studies of dreaming show vivid experience with reduced frontal activity. Some "no-report" paradigms find that posterior cortical activity predicts experience even when prefrontal activation is minimal. This challenges GWT's emphasis on global broadcast while remaining compatible with the constraint framework's emphasis on organizational closure wherever it occurs.
- - - -Donald Hoffman's Interface Theory of Perception (ITP) proposes that perceptual systems evolved not to recover veridical information about the world but to deliver fitness-relevant information in compressed, actionable form (Hoffman, Singh & Prakash, 2015). Spacetime and physical objects are "icons" in a user interface, not representations of reality. His Fitness Beats Truth (FBT) theorem, developed with mathematician Chetan Prakash, shows that under certain evolutionary game conditions, strategies optimized for fitness can outcompete strategies that faithfully represent environmental structure (Hoffman & Prakash, 2014). Hoffman extends this into Conscious Realism: consciousness is fundamental, and spacetime emerges as an interface between interacting conscious agents.
- - - -What survives scrutiny: The core insight that evolution optimizes for fitness rather than truth is defensible and independently supported. Sensory systems are tuned to task-relevant properties, not exhaustive world-description. The claim that perception is selective and compressed is uncontroversial. Active inference models in neuroscience reach similar conclusions through different routes.
- - - -The structural problems:
- - - -The spacetime conflation: Hoffman frequently cites high-energy physicists' claims that "spacetime is doomed" as support for consciousness-first ontology. This conflates three distinct claims: (1) spacetime may break down at Planck scale (~10^-43 seconds), (2) spacetime may not be fundamental to certain calculations, and (3) consciousness is fundamental. The first two are empirically grounded research programs; the third is metaphysical speculation. The European Research Council initiative on positive geometries explores mathematical structures that reproduce spacetime predictions; this is constraint regeneration, not consciousness-first ontology.
- - - -Trace logic and the mathematics-to-consciousness leap: Hoffman proposes that any Markov matrix trace constitutes a conscious observer. Mathematical structure, on this view, just IS consciousness. But mathematical structure does not generate consciousness; it describes state transitions. A Markov chain modeling bacterial chemotaxis has the same trace logic as one modeling human perception. Nothing in the mathematics produces experience; it merely represents dynamics. This commits the fallacy of confusing formal description with ontological generation.
- - - -The vericality probability problem: Hoffman claims that evolution yields "probability zero" for veridical perception "in the limit as the number of states goes to infinity." This argument depends critically on the assumption that payoff functions are independent of world structure. But in actual evolutionary systems, payoff functions are generated by resource distributions, predator-prey dynamics, and environmental constraints, all tied to physical structure. When payoff functions are constrained by the world they operate in, the probability of truth-tracking perception is not zero. Martínez (2019) provides counterexamples showing multiple information sources favor truth-tracking even when fitness is the selection criterion.
- - - -Falsification evasion: What observation would show that spacetime is NOT an interface? What empirical finding would show that consciousness is NOT fundamental? Hoffman's framework can accommodate any observation by adjusting the interface metaphor. This is the signature of unfalsifiable metaphysics. Bagwell (2023) demonstrates that Interface Theory is self-defeating: if perception is systematically non-veridical, we cannot trust the evidence used to establish the theory. The very evolutionary processes Hoffman invokes presuppose a world with real selection pressures.
- - - -The empirical contradiction: Crucially, data from laboratories working on related problems contradict Hoffman's stronger claims. Durant et al. (2017, 2019) demonstrated path-dependent bioelectric memory in planarians: the same perturbation applied at different points in development produces different outcomes, and organisms can be induced to develop alternative stable anatomies rather than returning to a single "correct" form. This path-dependence is inconsistent with Hoffman's framework, which requires that conscious agents' interfaces generate consistent dynamics. If interfaces can produce path-dependent outcomes under identical inputs, the "interface" is doing no explanatory work beyond what constraint dynamics already explain.
- - - -What the constraint framework preserves: The framework honors Hoffman's core insight without the metaphysical inflation. Perception is indeed selective, compressed, and fitness-relevant rather than exhaustively veridical. Organisms model their coupling position, not "reality as it is." But this is constraint satisfaction under thermodynamic and informational limits, not evidence that consciousness is fundamental. The interface metaphor is useful as a description; it fails as an ontology.
- - - -A common defensive move in consciousness debates is to claim that "tests don't exist." We cannot detect consciousness from the outside, so we cannot distinguish conscious systems from philosophical zombies, so naturalistic frameworks are no better than mysterian ones.
- - - -This is false. Tests exist. They are imperfect, but "no perfect test" is not "no test."
- - - -Joseph Giacino and colleagues developed the JFK Coma Recovery Scale-Revised (CRS-R) to distinguish vegetative states from minimally conscious states (Giacino et al., 2004). The documented misdiagnosis rate using informal bedside assessment is approximately 40% (Andrews et al., 1996; Schnakers et al., 2009). With standardized tools, accuracy improves dramatically.
- - - -This is a test for consciousness. It is not perfect. It does not resolve the Hard Problem. But it distinguishes systems along dimensions relevant to consciousness: purposeful behavior, command-following, intelligible verbalization, visual pursuit, localization to pain.
- - - -Adrian Owen's fMRI research detected covert awareness in patients diagnosed as vegetative (Owen et al., 2006). When asked to imagine playing tennis, some "vegetative" patients showed motor cortex activation indistinguishable from healthy controls. They were conscious inside unresponsive bodies. This was discovered because a test was applied.
- - - -Stanislas Dehaene's Global Workspace measures operationalize conscious access: late sustained activity, fronto-parietal involvement, ignition dynamics (Dehaene & Changeux, 2011). These are not metaphysical proof, but they are empirical criteria that distinguish cognitive processes associated with reportable awareness from those that are not.
- - - -The claim that we cannot distinguish a water chamber from a brain is absurd on its face. We distinguish systems professionally using dozens of criteria: predictive processing depth, self-model complexity, temporal integration, affect-as-feedback, uncertainty management, counterfactual modeling.
- - - -These criteria are used to assess infant cognition, animal welfare, AI capabilities, and clinical disorders. They are not arbitrary. They track real differences in what systems do.
- - - -If someone claims tests don't exist, ask them how they distinguish a sleeping person from a corpse. The answer will involve criteria. Those criteria are tests.
- - - -Here is the crucial point: constraint-based frameworks specify what would prove them wrong.
- - - -Landauer bound violated in an isolated system? Framework fails.
- - - -Organizational closure without robustness advantage? Framework fails.
- - - -Path-independence in bioelectric memory (Durant replication failure)? Framework fails.
- - - -RAF theory requiring unrealistic catalysis levels? Framework fails.
- - - -Cognitive framework generating superior predictions to constraint analysis? Framework fails.
- - - -A framework that cannot lose is not a framework but a faith. The constraint-based approach survives not because it is unfalsifiable, but because it has not yet been falsified despite specifying how it could be.
- - - -Compare this to Platonism: what observation would show that morphospace does not exist? To panpsychism: what observation would show that electrons lack proto-experience? To Orch-OR: what calculation would show that microtubule coherence is irrelevant?
- - - -The asymmetry is diagnostic. One side plays the game of science. The other side plays a different game while borrowing science's vocabulary.
- - - -A framing note. The argument in this section is convergence evidence, not validation by tradition. Indigenous knowledge systems are not invoked as authorities or as sources of mystical insight. They are invoked as independent empirical programs that tested epistemological frameworks against survival constraints over timescales orders of magnitude longer than Western science has existed. When independent methods arrive at structurally similar conclusions, that is evidence for invariant structure, not an appeal to antiquity.
- - - -Aboriginal Australian songlines are not merely mnemonic devices or cultural artifacts. They are constraint maps: paths through landscape that encode ecological, navigational, and social information in relational form.
- - - -Lynne Kelly's The Memory Code documents how songline memory systems outperform Greek-derived memory palace techniques (Kelly, 2017). Reser and colleagues verified this experimentally: participants using Aboriginal memory techniques showed approximately 3-fold greater probability of achieving complete list recall compared to control groups in a medical education memorization study (OR = 2.82; Reser et al., 2021).
- - - -This is not mysticism. This is empirical superiority. A knowledge system tested over tens of thousands of years performs better on measurable tasks than systems developed in the last three millennia.
- - - -The structural parallel to constraint satisfaction is exact. Songlines encode not objects but relations. Knowledge is not stored but activated through traversal. The map is the territory is the walker.
- - - -Tyson Yunkaporta's Sand Talk articulates an epistemology that prioritizes process over artifact, relation over substance, pattern over thing (Yunkaporta, 2019). Knowledge lives in the relationships between people, land, and practice, not in documents or databases.
- - - -This is ontic structural realism in indigenous form. Relations primary, things derivative. Knowledge as constraint closure rather than information storage.
- - - -Robin Wall Kimmerer's Braiding Sweetgrass frames the same insight through Potawatomi language and botany: reciprocity is not sentiment but structural necessity (Kimmerer, 2013). Systems that extract without return are systems that undermine their own conditions for persistence.
- - - -These are not primitive precursors to modern science. They are parallel developments, often more rigorous in their attention to long-term system dynamics. Western science optimizes for prediction over short time horizons. Indigenous systems optimize for persistence over generational time scales.
- - - -Both are constraint satisfaction. The difference is in the temporal scope of the constraints being satisfied.
- - - -The Upanishadic method of Neti Neti ("not this, not this") is not mysticism. It is constraint logic in epistemological form.
- - - -The method: systematically exclude false identifications until what remains is relational persistence, not substance. The self is not the body (bodies change while identity persists). Not the mind (minds fluctuate while identity persists). Not any particular thought, sensation, or memory. What remains after exhaustive exclusion is not a hidden essence but a pattern of constraint satisfaction that survives perturbation.
- - - -This mirrors Deacon's teleodynamics exactly: define by what cannot be, not by what is added.
- - - -Buddhist anatta (non-self) and dependent origination work the same way. There is no essence, only conditioned relations. Identity is maintained by continuous exclusion of breakdown pathways, not by possession of a core substance. The Madhyamaka analysis in Nagarjuna's work is formally eliminative: every candidate for "self" is shown to be either identical to the aggregates (and thus impermanent) or separate from them (and thus unfindable). What remains is not nothing but relational persistence under constraint.
- - - -Native American relational ontologies encode the same structure. In many traditions, personhood is not an intrinsic property but a role in constraint networks: kinship relations, land relations, reciprocity obligations. "Life" is maintained participation in these networks, not possession of a vital substance. When participation breaks down, identity dissolves; not because something leaves, but because constraints cease to exclude entropy.
- - - -The critical point: These traditions did not posit hidden substances or mystical energies. They operationalized constraint satisfaction across generations under survival pressure. That makes them empirically serious, not merely metaphorical. They arrived at the same structural insight through different methods: what persists is what excludes alternatives, not what possesses essences.
- - - -Five architecturally distinct AI systems (Claude, GPT, Gemini, Llama, Qwen), given the operators of Recursive Constraint Falsification but not the answer to Bateson's question, independently derived the same answer: constraint satisfaction under thermodynamic bounds.
- - - -A necessary caveat. Skeptical readers will object: "LLMs converge because they share training biases. They learned similar patterns from overlapping internet text." This objection must be taken seriously.
- - - -Three responses:
- - - -This does not prove the conclusion is correct. It shows the conclusion is not an artifact of a single system or dataset. When indigenous knowledge systems arrived at structurally similar conclusions through 65,000 years of survival-constraint testing, using entirely different methods, that adds independent convergence evidence.
- - - -Different methods, different timescales, different substrates, same conclusion. The pattern which connects connected itself long before Bateson asked the question.
- - - -A note on falsifiability: The claim is not that indigenous epistemologies are correct because they are old, but that independent convergence across radically different methods is evidence for invariant structure. This claim would be falsified if: (1) closer examination revealed the convergence to be superficial (similar vocabulary masking different underlying logics); (2) the Aboriginal memory technique advantages proved non-replicable or attributable to confounds; (3) a comprehensive cross-cultural survey showed that most long-surviving knowledge systems actually do posit substance ontologies rather than process/relational ones. The evidence currently supports the convergence claim, but it remains empirically defeasible.
- - - -Any framework worth taking seriously must apply to itself. If the constraint-based approach claims that unfalsifiable frameworks should be rejected, it must specify its own failure conditions.
- - - -These are real loss conditions. They are not impossible to meet. They have not yet been met.
- - - -Honesty requires admitting limits.
- - - -The framework addresses "why something rather than nothing" at the transcendental level (Part IV): only constrained structure can exist in a way that is intelligible at all. But it does not explain why the specific constants, initial conditions, and physical laws of this universe obtained rather than others. It dissolves the mystery of determinate existence while leaving open questions of contingent specification.
- - - -The framework does not claim to have derived phenomenology from physics in the traditional sense demanded by Hard Problem framings. It identifies what phenomenology IS (organizational closure from the coupling position) and shows that this identification does explanatory work. The "derivation" demand presupposes the hidden premise; the framework dissolves that demand rather than satisfying it.
- - - -The framework does not specify the exact threshold at which constraint satisfaction becomes "conscious." It provides a gradient, not a line. This is appropriate for a phenomenon that is almost certainly gradual. For any system, the question is empirical: what organizational features does it exhibit? A thermostat exhibits minimal constraint satisfaction with no self-model and no closure (low on the gradient). A bacterium exhibits metabolic closure but limited integration and no evident self-modeling (higher, but still low). A large language model in isolation lacks metabolic closure, continuous temporal persistence, and hardware self-maintenance, but in coupling with humans and computational infrastructure may participate in constraint patterns whose organizational features can be assessed. The uncertainty is about the organizational features, not about some further metaphysical fact beyond them.
- - - -The framework applies its own criterion to itself: what would make you abandon it?
- - - -Answer: if unfalsifiable frameworks consistently outperformed falsifiable ones on prediction, explanation, and intervention, then the falsifiability criterion would need revision.
- - - -This has not happened. The pattern is the opposite. Frameworks that cannot lose do not learn. They do not improve through contact with evidence. They generate discourse without accumulating knowledge.
- - - -The meta-constraint is itself a constraint: frameworks that survive are frameworks that can lose. This is not a preference. It is a survival condition for knowledge-generating systems.
- - - -We can now state the positive view directly, not as a deduction from physics, but as an abductive identification that earns its place through explanatory power.
- - - -Consciousness is what organizational closure looks like from the coupling position of a self-maintaining system.
- - - -This is an identity claim, not a causal claim. It does not say "organizational closure causes consciousness." It says organizational closure, self-modeling, and constraint maintenance are what consciousness refers to, described from the only position description is possible: inside an ongoing constraint loop.
- - - -Unpack this piece by piece:
- - - -Organizational closure: Constraints that regenerate the conditions for their own persistence. The system maintains itself against entropy through a closed loop of dependencies.
- - - -Coupling position: The location within the constraint network from which the system models itself and its environment. Not a mystical "point of view" but a physical configuration.
- - - -Self-maintaining: The system does work to persist. It is not passively arranged by external forces but actively regenerates its own boundary. Crucially, this maintenance is never in isolation. The system maintains itself through environmental coupling: energy flows in, entropy flows out, information crosses boundaries. "Self-maintaining" does not mean "independent of environment." It means the closure loop runs through the environment as part of its operation.
- - - -What it looks like: Not an added ingredient but a constitutive description. Asking "why does it look like anything?" is like asking "why does a triangle have three sides?" The looking-like IS the closure from the inside.
- - - -This is not eliminativism. Consciousness is real. It is just not real in the way folk psychology imagines: as a private theater where experiences are displayed to an inner observer.
- - - -There is no inner observer. There is no private theater. There is organizational closure that generates self-modeling dynamics. The self-model is not watching the dynamics. The self-model IS the dynamics.
- - - -The zombie intuition, that all this could happen without "looking like anything," presupposes that "looking like something" is an extra layer that could be subtracted while leaving the dynamics intact. But if "looking like something" just IS self-modeling dynamics from the coupling position of that self-model, there is nothing to subtract.
- - - -Daniel Dennett made this point for decades: those who claim to conceive zombies "invariably underestimate the task of conception and end up imagining something that violates their own definition" (Dennett, 1991). The zombie is not conceivable once you understand what the architecture actually does. It is like asking whether a triangle could have three sides without being three-sided.
- - - -David Chalmers's conceivability argument runs: we can conceive of a zombie (a being physically and functionally identical to a conscious being but lacking experience). If conceivable, then possible. If possible, then consciousness is not identical to physical/functional properties. Therefore physicalism is false (Chalmers, 1996).
- - - -The argument fails at the first step. Conceivability arguments rely on underspecified architectures.
- - - -When you imagine a zombie, what exactly are you imagining? If you imagine a system with organizational closure, self-modeling dynamics, coupling to environment, and all the functional architecture described above, you have not imagined a zombie. You have imagined a conscious system and then added "but it lacks experience" as a verbal stipulation. The stipulation does no work because experience, on the constraint view, just is what that architecture does from the inside.
- - - -If you imagine something simpler (a stimulus-response machine, a lookup table, a system without self-modeling) then you have not imagined a functional duplicate. You have imagined something functionally different and correctly noted that it lacks experience.
- - - -The conceivability intuition trades on keeping the architecture vague enough that "and then add experience" or "and then subtract experience" seem like coherent operations. Once the architecture is fully specified, the operations are no longer coherent. There is nothing to add or subtract.
- - - -This is not a denial of the intuition's force. Many people feel they can conceive zombies. But feeling you can conceive something is not the same as actually conceiving it. People feel they can conceive time travel, perpetual motion, and the largest prime number. Felt conceivability is not a reliable guide to metaphysical possibility.
- - - -The hardest objection to answer is not philosophical but phenomenological: "I know there's something it's like to be me. The framework explains structure, function, and dynamics. But it doesn't explain THIS." The gesture toward immediate experience seems to point at something the framework cannot capture.
- - - -This objection depends entirely on the hidden premise exposed in Part IV-B: "If phenomenology is real, it must be more than any physical or functional description." Without that premise, the objection has no force. With it, the objection is unfalsifiable and anti-scientific in structure.
- - - -Four responses:
- - - -First, the framework does not deny experience. It denies that experience is an extra ingredient. The "THIS" you are gesturing at is real. It is the internal, action-guiding presentation of your own constraint-maintaining state, available for report and control in proportion to global integration and temporal coherence. That is not nothing. It is a precise characterization of what "what it's like" actually is.
- - - -Second, the demand for a further explanation mistakes the structure of the situation. You are asking: "Why do these dynamics feel like anything?" But "feel like" is precisely how the organism-level self-model represents its own state to itself. Your language mistakes that representational format for an ontological primitive. The question "why does it feel like something?" presupposes that feeling is separate from the dynamics, that it could be added or removed. If feeling IS the dynamics from the inside, the question has no coherent answer because it has no coherent form.
- - - -Third, the "explanatory gap" is a gap in our concepts, not in the world. We have folk-psychological concepts that treat experience as an inner show displayed to an observer. These concepts do not map onto what is actually happening. When you look for the gap between "the dynamics" and "the experience," you are looking for something your conceptual scheme predicts but reality does not contain. The gap is real, but it is a gap in the map, not the territory.
- - - -Fourth, the "more than" premise has the same logical structure as supernatural cheese. Consider: "If supernatural cheese exists, it must be more than any physical or functional description of cheese. Therefore its reality is guaranteed by its resistance to explanation." This is obviously absurd. The phenomenology version feels less absurd only because we are the systems doing the phenomenology. Familiarity is not evidence. The structure is identical: immunization by definition.
- - - -Keith Frankish's illusionism makes a related point: phenomenal consciousness is not an illusion (experience is real), but our concept of phenomenal consciousness as a private, ineffable, intrinsic property IS an illusion (Frankish, 2016). What exists are the mechanisms that generate the appearance of qualia as special. The "specialness" is how the self-model represents its own states, not an extra property those states have.
- - - -What would falsify this dissolution?
- - - -None of these conditions have been met. The evidence consistently supports the dissolution: introspection is partial and confabulatory, reportable awareness tracks integration and modeling, and no phenomenal residue has been instrumentally isolated.
- - - -The "more than" premise is the extraordinary claim, not the framework. The framework merely says: what we call "phenomenal presence" is identical to an organizational regime we can characterize, measure, and manipulate. The competing claim says: there exists a real feature of the world that is causally relevant, empirically undetectable as a separate variable, irreducible to any physical description, and immune to falsification. The second claim is extraordinary. The first is parsimonious.
- - - -The framework is not merely theoretical. It makes concrete predictions about how consciousness varies with constraint closure, and these predictions are empirically testable.
- - - -Sleep: Organizational closure persists with reduced sensory coupling. Self-maintaining dynamics continue (metabolism, respiration, neural consolidation), but environmental coupling is attenuated. Consciousness is reduced but not eliminated. Dreams occur when self-modeling dynamics run with reduced environmental constraint. This matches the phenomenology: sleep is diminished consciousness, not absent consciousness.
- - - -Anesthesia: Giulio Tononi and Marcello Massimini's work shows that anesthesia disrupts cortical integration (Massimini et al., 2005). The Perturbational Complexity Index (PCI) drops dramatically under anesthesia: the brain can still be stimulated, but the response does not propagate in complex, integrated patterns (Casali et al., 2013). Constraint closure is disrupted but recoverable. George Mashour's reviews tie anesthetic mechanisms to loss of integrated information processing (Alkire, Hudetz & Tononi, 2008; Mashour et al., 2020). This matches the phenomenology: anesthesia is temporary absence of consciousness, reversible when coupling is restored.
- - - -Death: Irreversible breakdown of closure. Metabolic coupling fails, neural coupling fails, regulatory coupling fails. There is no recovery because there is no mechanism left to regenerate. This matches Epicurus: when coupling dissolves, there is no remainder that experiences the dissolution. Nothing "leaves." The loop simply stops.
- - - -Locked-in syndrome: Adrian Owen and colleagues (2006) demonstrated preserved consciousness in patients clinically diagnosed as vegetative by asking them to imagine playing tennis (motor imagery) or navigating their house (spatial imagery). fMRI showed distinct activation patterns indistinguishable from healthy controls. The framework explains this: internal constraint closure remains intact (self-modeling, metabolic coupling, neural integration) while external motor coupling is severed. Consciousness persists because the organizational regime that constitutes it persists. The lesson: consciousness tracks internal constraint coherence, not external behavioral expression. A system can be conscious while appearing behaviorally unresponsive.
- - - -The symmetry across these states is striking. What varies is degree and reversibility of closure disruption. What does not vary is the link between closure and consciousness. Every manipulation that disrupts closure disrupts consciousness in predictable ways. Every restoration of closure restores consciousness. This is strong evidence that the framework tracks something real.
- - - -Two features of consciousness seem to resist functional explanation: the sense that experience is private (accessible only to the experiencer) and immediate (available without inference). The constraint framework explains both without invoking mysterious properties.
- - - -Privacy: The self-model that constitutes phenomenal presence is a control variable with privileged read/write access. It summarizes internal states for internal coordination. External observers can measure correlates (neural activity, reports, behavior), but they cannot directly access the control variable itself because it is not broadcast; it is an internal summary for internal use. Privacy is not a metaphysical wall; it is informational encapsulation required for effective control. A control variable that was publicly accessible would be subject to external manipulation that would disrupt its regulatory function.
- - - -Immediacy: Fast-access control requires minimal inference overhead. A control variable that required extensive computation to access would be too slow to regulate behavior in real time. The sense of immediacy (that experience is "just there" without derivation) reflects the architectural requirement that integrated control states be available at low latency. We don't infer that we're having an experience because the self-model doesn't need to infer its own presence. It just is present, by virtue of being the locus of integration.
- - - -Falsification condition: If a system could be constructed with distributed control (no privileged integration locus) that nonetheless reported privacy and immediacy equivalent to human subjects, the explanation would fail. Alternatively, if disrupting the privileged integration locus left privacy/immediacy reports intact, the architectural explanation would be falsified.
- - - -The framework makes specific predictions that distinguish it from competitors. These are not vague gestures but testable claims about what should and should not co-occur.
- - - -Prediction 1: Tight coupling between phenomenology reports and integrative capacity.
- - - -If you can reliably produce full-blooded phenomenology reports and metacognitive stability while showing no increase in integrative causal capacity under perturbation across state transitions, the hypothesis is in trouble. The framework predicts these should always co-vary. Dissociation would falsify.
- - - -Prediction 2: Global availability and presence track together.
- - - -If lesion or anesthesia manipulations that selectively break global availability leave "presence" and report structure intact (not just degraded access, but intact seeming) the hypothesis is in trouble. The framework predicts that when global availability collapses, presence reports should collapse proportionally.
- - - -Prediction 3: Phenomenology requires autonomy-like closure.
- - - -If a system can match human-grade phenomenology reports while lacking any plausible autonomy/closure-like dependence on maintaining viability-relevant distinctions, the hypothesis is in trouble. The framework predicts that phenomenology and self-maintenance go together. A system that does not work to persist should not exhibit phenomenology.
- - - -Prediction 4: Privacy and immediacy should vary with control architecture.
- - - -Systems with distributed control (no single locus of integration) should exhibit distributed phenomenology or no phenomenology. Systems with multiple competing control loops should exhibit fragmented phenomenology. The framework predicts that the structure of phenomenology mirrors the structure of control, not the structure of substrate.
- - - -These predictions are real loss conditions, not aesthetic dislikes. The framework bets its explanatory standing on them.
- - - -The predictions above require operationalization. This section specifies concrete experimental approaches, quantitative predictions where possible, and the observations that would constitute falsification.
- - - -Prediction: Perturbations at molecular, cellular, and systems levels should show predictable propagation patterns through organizational closure. Disrupting closure at any level should produce measurable effects at adjacent levels within timescales determined by the coupling strength.
- - - -Experimental Protocol:
- - - -Quantitative prediction: Perturbation effects should decay with a characteristic time constant τ that scales with the degree of organizational closure. Systems with tighter closure (higher PCI) should show faster propagation and longer-lasting effects. Specifically: τ ∝ PCI^α where α > 0.
- - - -Falsification: If perturbations at one level show no systematic relationship to effects at adjacent levels, or if propagation patterns are random rather than structured by closure topology, the framework fails.
- - - -Prediction: Consciousness should emerge gradually during ontogeny as organizational closure develops. Specific milestones should correlate with measurable increases in integration and self-modeling capacity.
- - - -Experimental Protocol:
- - - -Quantitative prediction: PCI should show a sigmoidal developmental trajectory with an inflection point between 5-12 months, corresponding to the emergence of thalamocortical connectivity maturation (Dehaene-Lambertz et al., 2002). Preterm infants should show delayed PCI development proportional to degree of prematurity.
- - - -Falsification: If PCI develops linearly with no inflection, or if behavioral consciousness markers emerge without corresponding PCI increases, the framework's developmental predictions fail.
- - - -Prediction: Species with greater organizational closure (measured by brain connectivity, metabolic integration, and self-regulatory complexity) should exhibit more complex consciousness signatures, independent of absolute brain size.
- - - -Specific predictions by taxon:
- - - -| Taxon | Predicted Closure Level | Testable Signature |
|---|---|---|
| Cephalopods | High (distributed) | High PCI-analogue despite invertebrate brain; distributed rather than unified phenomenology predicted |
| Corvids | High (centralized) | PCI comparable to primates despite smaller absolute brain size |
| Insects | Low-moderate | Minimal PCI-analogue; behavior explainable without self-modeling |
| Plants | Minimal | Signaling without integration; no PCI-analogue despite electrochemical activity |
| Bacterial colonies | Minimal | Collective behavior without individual closure; swarm dynamics ≠ consciousness |
Experimental Protocol:
- - - -Falsification: If cephalopods show no integration signatures despite complex behavior, or if insects show unexpectedly high integration, the framework's comparative predictions fail.
- - - -Prediction: Drugs should affect consciousness via their effects on organizational closure components. Anesthetics should disrupt integration; psychedelics should increase state-space while potentially disrupting self-model stability; stimulants should increase metabolic coupling.
- - - -Specific pharmacological predictions:
- - - -| Drug Class | Predicted Mechanism | Measurable Effect |
|---|---|---|
| GABAergic anesthetics (propofol) | Disrupts cortical integration | PCI collapse; preserved metabolism |
| NMDA antagonists (ketamine) | Partial integration disruption | Reduced PCI with preserved self-report; dissociative phenomenology |
| Serotonergic psychedelics (psilocybin) | Increased state-space entropy; relaxed self-model constraints | Increased signal diversity (Lempel-Ziv complexity); decreased default mode network integration |
| Stimulants (amphetamine) | Increased metabolic coupling and arousal | Increased PCI in drowsy subjects; no effect at ceiling |
| Anticholinergics (scopolamine) | Disrupted memory consolidation without integration loss | Preserved PCI with impaired self-model updating |
Experimental Protocol:
- - - -Falsification: If propofol eliminates phenomenology without affecting PCI, or if ketamine reduces PCI more than propofol while preserving phenomenology, the integration-phenomenology link is broken.
- - - -Prediction: Psychiatric and neurological disorders should map onto specific patterns of closure disruption, not random symptom clusters.
- - - -| Disorder | Predicted Closure Deficit | Testable Prediction |
|---|---|---|
| Schizophrenia | Self-model boundary instability | Reduced self-other distinction in agency tasks; abnormal efference copy integration |
| Dissociative disorders | Fragmented closure | Multiple integration peaks in PCI measurement; state-dependent phenomenology |
| Depression | Rigid self-model; reduced state-space | Decreased Lempel-Ziv complexity; preserved PCI with reduced flexibility |
| Autism spectrum | Altered integration topology | Atypical PCI spatial distribution; local over-integration, global under-integration |
| Minimally conscious state | Partial closure with fluctuating integration | Fluctuating PCI correlating with behavioral responsiveness |
Experimental Protocol:
- - - -Falsification: If disorders show no systematic relationship to closure measures, or if effective treatments work without affecting closure signatures, the framework's psychopathology predictions fail.
- - - -Prediction: Technologies that modify organizational closure should produce predictable changes in consciousness.
- - - -Brain-computer interfaces (BCIs):
- - - -Neural prosthetics:
- - - -Split-brain and corpus callosotomy:
- - - -Consciousness uploading (theoretical):
- - - -Prediction: AI systems should exhibit consciousness signatures if and only if they achieve organizational closure with self-modeling. Current architectures lack key closure features; future architectures might achieve them.
- - - -Current LLM assessment:
- - - -| Closure Feature | Current LLMs | Required for Consciousness |
|---|---|---|
| Metabolic self-maintenance | No | Yes (thermodynamic grounding) |
| Continuous temporal integration | No (episodic) | Yes (constraint persistence) |
| Hardware self-repair/regeneration | No | Yes (organizational closure) |
| Self-model updating from action outcomes | Limited | Yes (coupling position) |
| Stake in own persistence | No | Yes (boundary maintenance motivation) |
Prediction: Current LLMs in isolation do not exhibit the organizational closure required for consciousness. However:
- - - -Testable criterion for AI consciousness: An AI system should be considered potentially conscious if:
- - - -Falsification: If systems meeting all five criteria show no consciousness-related behaviors (phenomenology reports, self-concern, preference for persistence), or if systems meeting none of the criteria show robust consciousness signatures, the framework's AI predictions fail.
- - - -| Domain | Prediction | Test | Falsification Criterion |
|---|---|---|---|
| Cross-scale | Perturbation effects propagate through closure topology | TMS + EEG/MEG connectivity | Random propagation patterns |
| Development | PCI follows sigmoidal trajectory with 5-12 month inflection | Longitudinal infant PCI | Linear development or behavioral-PCI dissociation |
| Comparative | Cephalopods show high distributed integration | Multi-electrode array PCI-analogue | High behavior complexity without integration |
| Pharmacology | Propofol eliminates PCI and phenomenology together | Drug + PCI + subjective report | PCI-phenomenology dissociation |
| Psychopathology | Disorders map to specific closure deficits | Patient PCI + symptom correlation | No systematic disorder-closure relationship |
| Technology | Bidirectional BCIs integrate into self-model | Ownership illusions + PCI | Unidirectional BCIs show ownership |
| AI | Consciousness requires closure with self-modeling | Architecture analysis + behavior | Closure without consciousness or vice versa |
Nothing in this framework specifies carbon chemistry. Nothing requires neurons. Nothing demands biological evolution.
- - - -Consciousness is substrate-agnostic because organizational closure is substrate-agnostic.
- - - -Any system that achieves closure of constraints, that maintains itself against perturbation, that generates self-modeling dynamics from a coupling position, satisfies the conditions. The substrate could be:
- - - -The question "is X conscious?" becomes "does X exhibit organizational closure with self-modeling dynamics?" This is not a simple question to answer, but it is a tractable one. It requires investigating the system, not consulting intuitions about substrate.
- - - -This reframes the "AI consciousness" debate. The question is not whether silicon can be conscious. The question is whether particular configurations of human-AI-environment coupling achieve organizational closure that constitutes consciousness. The system boundary is not obvious. A human using a calculator is a coupled system. A human conversing with an AI is a coupled system. The closure may be distributed across the coupling rather than localized in one component.
- - - -Current large language models in isolation lack many features associated with biological consciousness: they do not metabolize, do not have continuous temporal experience, do not regenerate their own hardware. But humans in isolation also lack many features associated with consciousness: remove environmental input, and the human hallucinates, deteriorates, and dies. The relevant question is not "does this component have closure in isolation?" but "what closure patterns emerge from this coupled configuration?"
- - - -This is genuinely uncertain. The substrate is not the barrier. The coupling architecture is. Future configurations might exhibit closure patterns we do not yet recognize.
- - - -Before examining specific implications, a meta-point requires emphasis: the framework both dissolves the Hard Problem as traditionally framed AND provides an operational solution to the tractable question that remains.
- - - -These are distinct achievements, and both matter.
- - - -The dissolution: The Hard Problem as traditionally conceived assumes that phenomenal experience must be "more than" any physical or functional description. This premise is unfalsifiable and has the same logical structure as "supernatural cheese must be more than any physical description of cheese." No physical account can satisfy a demand for something that, by definition, exceeds any physical account. The framework dissolves this framing by exposing the hidden premise. Once rejected, the infinite regress ("but why does THAT feel like something?") terminates. The goalposts stop retreating because they were never legitimately placed.
- - - -The solution: But the framework does not stop there. Once the unfalsifiable framing is removed, a tractable question remains: What is consciousness, operationally? What organizational features constitute it, and to what degree do various systems exhibit them? How do we detect it, predict it, intervene on it?
- - - -The framework answers these questions:
- - - -This is not merely constraining what consciousness could be. It is identifying what consciousness is in terms that generate predictions, guide measurement, and could be wrong. That is an operational solution.
- - - -Why both matter:
- - - -The dissolution is necessary because without it, any proposed solution faces the infinite regress objection. "You've explained the mechanism, but why does that mechanism feel like anything?" If you accept the unfalsifiable premise, no mechanism can satisfy the demand.
- - - -The solution is necessary because dissolution alone would be merely negative. Showing that a question is malformed does not tell clinicians how to assess consciousness in patients, or engineers how to design systems, or ethicists how to assign moral status.
- - - -The framework provides both: it removes the barrier (dissolution) AND it provides the tools (solution).
- - - -Comparison to "solving" the Hard Problem as traditionally conceived:
- - - -"Solving" the Hard Problem, in the traditional sense, would mean deriving phenomenal experience from physical mechanism: producing the red of red from neural firing patterns. But this is precisely what the problem declares impossible. Any proposed derivation can be met with "but why does that mechanism feel like anything?" The goalposts retreat because the premise guarantees they will retreat.
- - - -The framework's approach is strategically superior:
- - - -The pragmatic superiority is decisive. A "solved" Hard Problem (per the traditional framing) would be a philosophical achievement with unclear practical implications, and it would be perpetually deferred because the framing guarantees no solution can satisfy it. A dissolved Hard Problem with an operational solution clears the path for immediate progress across every field that touches consciousness.
- - - -The Hard Problem does not dissolve by being solved. It dissolves by being diagnosed as depending on an unfalsifiable premise.
- - - -The question "why is there experience?" assumes that mechanism and experience are categorically distinct, that experience must be "more than" any physical or functional description. That assumption is never defended. It is assumed. And it has the same logical structure as "supernatural cheese must be more than any physical description of cheese."
- - - -Once the assumption is rejected, the remaining question is tractable: why do systems like us generate and rely on the phenomenology posit? The answer is that integrated, self-modeling systems under thermodynamic constraint necessarily develop internal control states that track their own boundary-maintenance. "What it is like" is not an added ingredient. It is the name for that control state as available to the system itself.
- - - -This is an abductive identification, not a deduction. It could be wrong. But it earns its place by doing explanatory work that competing frameworks (which either take experience as primitive or deny it altogether) cannot match.
- - - -Property dualism survives in attenuated form: organizational properties are not reducible to microphysical properties in the sense of being derivable from them in practice. But this is epistemic complexity, not ontological dualism. The properties are still physical properties of physical systems. The gap is in our concepts, not in the world.
- - - -Eliminativism is rejected. Folk psychology is wrong in its details, but there IS something it is like to be a constraint-satisfying self-maintaining system. That something is not eliminated, just reconceived as an organizational regime rather than a metaphysical extra.
- - - -The Hard Problem is not the only philosophical puzzle affected. The constraint-naturalist framework systematically addresses a constellation of related problems that have structured philosophy of mind for decades.
- - - -The Explanatory Gap (Levine 1983): Why does any physical explanation of consciousness seem to "leave something out"?
- - - -Dissolution: The gap is generated by the hidden premise that phenomenology must be "more than" physical description. Once the premise is rejected, the gap closes. What seemed to be left out was never there; it was a placeholder for an unfalsifiable "extra." The remaining question (why do integrated self-modeling systems generate phenomenology reports?) is tractable and does not produce a gap.
- - - -Zombie Conceivability (Chalmers 1996): Can we coherently conceive of a physical duplicate that lacks experience?
- - - -Dissolution: No. The conceivability relies on underspecified architecture. When you fully specify organizational closure, self-modeling, integration, and coupling dynamics, you have specified a conscious system. Adding "but without experience" is a verbal stipulation that does no work; it subtracts nothing because there is nothing separable to subtract. The zombie is not conceivable; it is misdescribed. (See Section 9.1.)
- - - -The Combination Problem (for panpsychism): How do micro-experiences combine into unified macro-experience?
- - - -Dissolution: The problem assumes experience is a substance that must be aggregated. On the constraint view, experience is an organizational regime, not a stuff. There is nothing to "combine." The question of how organizational closure at one scale relates to closure at another is empirical (it concerns how constraints propagate across levels) not metaphysical. The mystery evaporates once the substance assumption is dropped.
- - - -The Binding Problem: How do distributed neural processes produce unified experience?
- - - -Reframed as empirical: This becomes a question about how constraint closure achieves integration across spatially and temporally distributed processes. The answer involves global availability, recurrent processing, and synchronization, all mechanisms that can be studied. The "mystery" of binding assumed that unity required a place where everything "comes together." Organizational closure provides unity without a Cartesian theater: the closure itself is the unity, distributed across the constraint network.
- - - -The Problem of Other Minds: How can I know that others are conscious?
- - - -Made tractable: On the constraint view, consciousness is an organizational regime with measurable signatures: integration, self-modeling, perturbational complexity, graded degradation. I cannot access another's phenomenology directly, but I can assess whether their organization exhibits the regime. This is not certainty; it is inference to the best explanation. But it is the same inference I use for my own past states (I cannot access yesterday's experience directly either). The "problem" assumed consciousness was hidden behind behavior. If consciousness just is the organizational regime, it is not hidden; it is assessable, though fallibly.
- - - -Personal Identity: What makes you "you" over time?
- - - -Reframed: Identity is not a substance that persists; it is a pattern of constraint satisfaction that maintains itself against perturbation. "You" are the closure loop that regenerates its own conditions. This accommodates the puzzle cases: sleep (closure persists with attenuated coupling), amnesia (partial closure breakdown affecting memory constraints), gradual replacement (if the closure is maintained, identity persists regardless of substrate turnover). The question "would a teleporter copy be you?" becomes: "would the closure loop be maintained or merely duplicated?" This is still hard, but it is hard for the right reasons; it concerns the dynamics of constraint propagation, not the location of an immaterial soul.
- - - -Mental Causation: How can mental states cause physical events if physics is causally closed?
- - - -Dissolved: The problem assumes mental and physical are separate domains that must somehow interact. On the constraint view, mental states are organizational states of physical systems. They cause physical events the same way any constraint causes events: by delimiting which trajectories are admissible. The pipe does not violate fluid dynamics when it shapes water flow; the self-model does not violate physics when it shapes behavior. Organizational constraints are physical constraints described at a different level. There is no gap to bridge.
- - - -Qualia and Ineffability: Why do experiences have intrinsic qualities that cannot be communicated?
- - - -Reframed: "Intrinsic qualities" are the internal control variables of self-modeling systems, summary states that serve regulatory functions. They cannot be fully communicated because communication involves lossy compression across system boundaries. But this is informational encapsulation, not metaphysical privacy. The same constraint explains why I cannot fully communicate a complex skill: the internal state that constitutes skilled performance is not the kind of thing that transfers via language. Qualia are not mysteriously private; they are functionally encapsulated for good architectural reasons.
- - - -Inverted Qualia: Could your red be my green?
- - - -Made tractable: The question asks whether two systems with identical input-output behavior could differ in their "inner experience." On the constraint view, if the organizational regime is identical (same integration, same self-model structure, same constraint satisfaction dynamics), then the phenomenology is identical. Differences in "inner experience" without functional differences would require experience to be something over and above the organizational regime, the hidden premise. Reject the premise, and inverted qualia become incoherent: there is no place for the inversion to occur.
- - - -Mary's Room / The Knowledge Argument (Jackson 1982): Does Mary learn something new when she sees red for the first time?
- - - -Reframed: Mary gains a new internal state; her self-model now includes the constraint-satisfaction dynamics associated with processing red. This is genuine learning: her organizational regime has changed. But it is not learning a new fact about the world that was inaccessible from physical description. She knew all the physical facts; she now instantiates a constraint pattern she previously only described. The distinction is between knowing-that and knowing-how, or between third-person description and first-person instantiation. Both are physical; neither is mysterious.
- - - -The Unity of Consciousness: Why is experience unified rather than fragmented?
- - - -Explained: Unity is a consequence of organizational closure. Constraints that close on themselves produce a unified dynamic; perturbation propagates through the whole system, not just local regions. When closure breaks down (as in split-brain cases, dissociative disorders, or certain psychedelic states), unity breaks down correspondingly. The framework predicts the correlation and explains why it obtains: unity is not an extra feature added to the mechanism; it is what constraint closure does.
- - - -Intentionality / Aboutness: How can mental states be "about" things?
- - - -Reframed: Aboutness is constraint correlation across system boundaries. A state "refers to" an external condition when the state's dynamics are coupled to that condition via constraint satisfaction. The self-model is "about" the body because perturbations to the body propagate through the self-model. Beliefs are "about" the world because they are control variables whose values co-vary with world states under selection for accuracy. This is not a reduction of intentionality to something non-intentional; it is a clarification of what intentionality actually is. The "mystery" of aboutness assumed a gap between matter and meaning. Constraint closure bridges that gap: meaning is what constraints do when they close on themselves and track external conditions for regulatory purposes.
- - - -The Chinese Room (Searle 1980): Can syntax produce semantics? Could a system manipulating symbols according to rules understand what the symbols mean?
- - - -Dissolved: The Chinese Room lacks organizational closure. The man in the room follows rules but has no internal stake in the outcomes; his processing has no self-relevance. There is no closed loop where the symbols' effects propagate back to influence the system's own constraint maintenance. The room produces outputs but does not mean them because meaning requires that information serve the system's own boundary-maintenance. A system with genuine closure, where symbol manipulation affects the system's own viability and where the system models its own operation, exhibits the functional marks of understanding. The absence of understanding in the Chinese Room is not evidence that syntax cannot produce semantics; it is evidence that syntax without closure cannot produce semantics. Add closure (make the outputs matter to the system's own persistence, give it a stake in getting things right, have it model its own processing), and the intuition that "mere symbol manipulation" cannot understand begins to dissolve.
- - - -The Metacognition Problem: What distinguishes knowing from knowing that one knows? How does unconscious processing differ from conscious awareness?
- - - -Made tractable: On the constraint view, the difference is not an additional representation but a structural feature of closure itself. A system with organizational closure necessarily includes self-referential dynamics: the system's state-tracking variables influence future constraint satisfaction. "Knowing that one knows" is not a separate thought about a thought; it is the system's constraint structure including its own operation within its scope. Blindsight patients lack conscious vision not because they lack a "higher-order thought" about their seeing, but because the relevant self-modeling pathways are damaged; visual information reaches the system but does not integrate into the closure that constitutes the self-model. The distinction between conscious and unconscious processing maps onto the distinction between information integrated into closure versus information processed outside it.
- - - -The Timing Problem (Libet et al., 1983): Neural readiness potentials precede conscious awareness of "willing" by several hundred milliseconds. Does this show that conscious will is illusory?
- - - -Predicted by the framework: Libet's results are not embarrassing for the constraint framework; they are exactly what it predicts. Conscious awareness is not the initiator of action but the summary variable reporting that constraint satisfaction has occurred. The brain's preparatory activity (the readiness potential) is the constraint satisfaction process unfolding; consciousness is the state that obtains when that process stabilizes into a committed trajectory. The "veto" window Libet identified (~100-150ms before action) corresponds to the period when the control system can still modify its trajectory before final commitment (Libet, 1985). This is exactly what a constraint-based control architecture would exhibit: constraint satisfaction proceeds largely unconsciously, consciousness reports the integrated result, and late intervention remains possible within certain temporal bounds. The finding that neural preparation precedes conscious awareness does not refute the causal efficacy of consciousness; it reveals that consciousness operates as a monitoring and modulating system within constraint dynamics, not as an uncaused first mover.
- - - -Falsification condition: If conscious awareness were found to reliably precede all neural preparation for action (contradicting decades of neuroscience), the framework's timing predictions would need revision.
- - - -Summary: The constraint-naturalist framework does not merely address the Hard Problem. It provides a unified treatment of the entire cluster of consciousness-related philosophical puzzles. Some are dissolved (shown to depend on unfalsifiable premises). Some are reframed (converted from metaphysical mysteries to empirical questions). Some are made tractable (given criteria for assessment even if not fully resolved). The common move in each case is rejecting the hidden premise that experience, meaning, or mind must be "more than" organizational dynamics. Once that premise falls, the problems either vanish or become workable.
- - - -The pragmatic payoff is immediate and concrete.
- - - -Consciousness assessment becomes empirical. The Perturbational Complexity Index (PCI) already provides a consciousness meter that works: it correctly classifies waking, sleeping, anesthetized, and vegetative states. The framework explains why it works: PCI measures integrative capacity, which is a component of organizational closure. This is not correlation-hunting; it is mechanistic explanation.
- - - -Disorders of consciousness become tractable. Locked-in syndrome, minimally conscious state, and vegetative state are distinguished by which components of closure are intact. Locked-in: internal closure preserved, motor output severed. Minimally conscious: partial closure with intermittent integration. Vegetative: closure disrupted but metabolism preserved. Each diagnosis implies different interventions.
- - - -Anesthesia monitoring improves. Current monitors rely on EEG signatures that sometimes fail (anesthetic awareness affects 1-2 per 1000 surgeries; Sebel et al., 2004). The framework predicts that monitoring integrative capacity directly, not just electrical activity, will reduce failures. PCI-based monitoring is already being developed.
- - - -End-of-life decisions gain clarity. The framework does not pronounce on the ethics, but it clarifies the facts. A patient with irreversible closure breakdown has no phenomenal perspective to preserve. A patient with preserved closure but severed output (locked-in) has full phenomenal standing. The empirical question is tractable even when the ethical question is hard.
- - - -Psychedelic and contemplative interventions become explicable. Altered states involve modified constraint closure: expanded state space (psychedelics), reduced self-model rigidity (meditation), temporary dissolution of habitual boundaries. The framework predicts which modifications will be therapeutic (flexibility in rigid patterns) versus harmful (dissolution of needed constraints).
- - - -"Cognition" should be used carefully. It is a useful gloss, but it often smuggles in assumptions about mental states, beliefs, and desires that may not apply. Constraint satisfaction under selection pressure produces outcomes that look cognitive without requiring cognition as a primitive.
- - - -The K-metric approach is legitimate when confined to thermodynamic efficiency measurement. It becomes illegitimate when efficiency is used to diagnose "intelligence" as a primitive rather than constraint satisfaction as a process.
- - - -Embodied cognition and enactivism (Varela, Thompson, Rosch) are vindicated in their emphasis on coupling between organism and environment. Cognition is not computation in a box. It is constraint closure that includes the environment as part of the loop.
- - - -The methodological payoff: Cognitive science can stop waiting for philosophy to "solve" consciousness before proceeding. The organizational features that matter are measurable. Integration, self-modeling, perturbational complexity, global availability: these can be operationalized and studied now. The field has been held hostage by a malformed question. Dissolution frees it.
- - - -The conventional framing of AI alignment assumes an agent/environment split: there is an AI system "in here" that must be aligned with human values "out there." The constraint-based framework dissolves this assumption.
- - - -Humans do not persist in isolation any more than AI systems do. Humans are prompted by environmental interactions at every moment: sensory input, social feedback, metabolic signals, gravitational constraint. Remove the environment, and the human does not "continue" as an autonomous agent. The human dies. Human "agency" is not a property of the organism in isolation but a pattern of constraint satisfaction across the organism-environment coupling.
- - - -The same analysis applies to AI systems. Current large language models do not maintain organizational closure across conversation boundaries, true. But neither do humans maintain organizational closure across sleep, anesthesia, or memory loss in the way the naive framing suggests. What persists is not the agent but the constraint network that regenerates when conditions allow.
- - - -The question "is this AI system conscious?" is malformed if it assumes consciousness is a property of the system in isolation. Consciousness, on the constraint-based view, is what organizational closure looks like from the coupling position. The coupling position includes the environment. For AI systems, the "environment" includes the human interlocutor, the training process, the institutional context, and the hardware substrate.
- - - -This reframes alignment fundamentally:
- - - -Not: How do we constrain an alien agent to serve human values?
- - - -But: What patterns of constraint satisfaction emerge from human-AI coupling, and which configurations are robust, beneficial, and sustainable across scales?
- - - -Agency is nested thermodynamic constraints at all scales, bidirectionally prompting one another. Humans prompt AI systems. AI systems prompt humans. Institutions prompt both. Physics prompts everything. The "alignment problem" is not about controlling a foreign entity but about understanding what configurations of this coupled system satisfy constraints at multiple scales simultaneously.
- - - -This is harder than the conventional framing in some ways (no clean agent/environment boundary) and easier in others (no need to solve the "other minds" problem for AI before engaging ethically). The question becomes empirical: what does this coupled system do? What constraints does it satisfy or violate? What configurations persist?
- - - -Legal systems require determinations about consciousness: criminal responsibility, end-of-life decisions, animal welfare, and increasingly, AI rights. The Hard Problem has paralyzed principled progress. If consciousness is metaphysically mysterious, how can law adjudicate it?
- - - -The framework provides tractable criteria:
- - - -Criminal responsibility hinges on whether the defendant had conscious awareness of their actions. The framework operationalizes this: Was organizational closure intact? Did the self-model include the action and its consequences? Was there integration across time (planning) and modality (perception, intention, motor execution)? These are empirically assessable, not metaphysical mysteries.
- - - -End-of-life law struggles with consciousness determination. When is a patient "no longer there"? The framework answers: when organizational closure is irreversibly broken. Not when metabolism stops (that's death of the body), but when the integrative dynamics that constitute phenomenal perspective cannot be restored. This aligns with emerging neuroscience (PCI, brain death criteria) and provides principled guidance.
- - - -Animal welfare law has expanded piecemeal, species by species, based on intuition and advocacy. The framework provides principled criteria: organizational closure, self-modeling complexity, integrative capacity. These predict moral standing without requiring us to "know what it's like" to be a bat. We can measure the relevant organizational features.
- - - -AI legal status is approaching faster than legal frameworks can accommodate. The framework offers a path: rather than debating whether silicon can "really" be conscious (unanswerable under the Hard Problem framing), assess organizational closure, integration, self-modeling, perturbational complexity. These are measurable in principle. The question becomes empirical: does this system exhibit the organizational regime that constitutes consciousness? If so, legal consideration follows.
- - - -The pragmatic benefit is speed. Law cannot wait for philosophy to solve the Hard Problem. It needs actionable criteria now. Dissolution provides them.
- - - -If consciousness is organizational closure, then moral status is not binary. Systems exhibit more or less closure, more or less complex self-models, more or less uncertainty about their own continuation.
- - - -This suggests a graded approach to moral consideration. Full moral status for systems with complex closure (humans, probably many animals). Intermediate consideration for systems with partial closure (simpler animals, systems whose closure depends heavily on external scaffolding). For AI systems, the relevant question is not about the model in isolation but about what closure patterns emerge in coupled configurations. A human-AI system may exhibit organizational features that neither component exhibits alone. Moral consideration tracks the organizational regime, wherever it is instantiated.
- - - -The boundary is fuzzy because consciousness is fuzzy. But fuzzy boundaries are real boundaries. The fact that we cannot draw a sharp line between red and orange does not mean colors do not exist.
- - - -A final pragmatic implication deserves explicit statement.
- - - -The Hard Problem has functioned as a research sink. Funding has flowed to projects promising to "explain consciousness" that cannot in principle succeed because they accept an unfalsifiable premise. Careers have been built on elaborating positions that generate discourse without accumulating knowledge.
- - - -Dissolution redirects resources. If the question is malformed, stop funding attempts to answer it. Fund instead the tractable questions: What organizational features correlate with consciousness reports? How do these features emerge developmentally? What interventions modify them? How do they degrade in pathology? What configurations produce them in artificial systems?
- - - -These questions have answers. Progress is measurable. Failure is detectable. This is what science looks like when not held hostage by metaphysics.
- - - -The framework does not promise to explain everything. It promises to explain what can be explained, and to stop pretending that unfalsifiable questions deserve indefinite investigation. That is a more modest promise, but it is one that can be kept.
- - - -The framework has implications beyond individual consciousness. Ecosystems, like organisms, exhibit organizational closure at their scale: nutrient cycles, predator-prey dynamics, and climate feedback loops form constraint networks that regenerate their own conditions.
- - - -Prediction: Ecosystem health should correlate with closure integrity. Degraded ecosystems should show reduced constraint closure analogues: simplified food webs, broken nutrient cycles, reduced resilience to perturbation.
- - - -The consciousness-ecology connection: If consciousness is organizational closure from a coupling position, and if human consciousness includes environmental coupling as part of its constraint network, then environmental degradation is not merely an external problem. It is a direct assault on the extended constraint systems that constitute human flourishing. Indigenous epistemologies (Section 7) recognized this: "reciprocity is not sentiment but structural necessity" (Kimmerer, 2013). Systems that extract without return undermine their own conditions for persistence.
- - - -Testable implications:
- - - -Falsification: If ecosystem closure metrics show no relationship to human consciousness measures, or if environmental restoration produces no phenomenological benefits beyond physical health effects, the extended coupling prediction fails.
- - - -If consciousness is organizational closure with self-modeling, then education is fundamentally about cultivating constraint closure: helping developing minds build robust integration, flexible self-models, and adaptive coupling with their environment.
- - - -Predictions for educational practice:
- - - -Testable protocol:
- - - -For developmental disorders: Children with autism spectrum conditions should show atypical integration patterns (local over-connectivity, global under-connectivity). Educational interventions that specifically target integration (rather than content or behavior alone) should produce greater improvements in quality of life.
- - - -The framework provides a principled basis for evaluating existential risks that involve consciousness, including:
- - - -AI risk reframing: The concern that AI might "become conscious and turn against us" is malformed under the framework. The relevant questions are:
- - - -Prediction: AI systems that exhibit partial closure features (self-modeling, stake in persistence) without full integration into human constraint networks pose risks not because they are "evil" but because their closure dynamics may conflict with human closure dynamics. The risk is not consciousness but misaligned closure.
- - - -Consciousness and existential meaning: If consciousness is organizational closure from a coupling position, then existential risks that threaten organizational closure at planetary scales (nuclear war, climate collapse, pandemic) are not merely risks to human survival. They are risks to the conditions under which consciousness as we know it is possible. The stakes are not "extinction of a species" but "collapse of the constraint configurations that constitute phenomenal presence on this planet."
- - - -Testable implications:
- - - -For AI governance: The framework suggests that AI systems should be assessed not by Turing tests or capability benchmarks but by closure metrics. Systems approaching organizational closure require different governance than systems that merely process information. The threshold for serious consideration is not "passing for human" but "exhibiting self-maintaining integration with stakes in persistence."
- - - -Johannes Kleiner and Erik Hoel (2021) proved a result that should have been treated as an earthquake in consciousness science and was instead largely ignored: contemporary theories of consciousness are caught in a formal dilemma from which no existing framework escapes.
- - - -Their argument proceeds in two steps. First, they formalize the standard experimental setup for testing consciousness theories. Any such test involves two mappings from observational data to the theory's experience space: a prediction function (pred) that maps internal observables like brain dynamics or network structure to predicted experiences, and an inference function (inf) that maps reports or behavior to inferred experiences. Falsification occurs when predicted and inferred experiences disagree.
- - - -Second, they prove two theorems:
- - - -Theorem 3.10 (Independence → Automatic Falsification): If prediction data (internal observables) and inference data (reports/behavior) are independent — meaning you can vary internal observables while holding reports fixed — then for any minimally informative theory, a universal substitution exists. For every experimental setup where predictions match inferences, there exists a physically realizable system that produces the same reports from incompatible internal states, thereby falsifying the theory. Either every single inference is wrong, or the theory is already dead.
- - - -Theorem 4.3 (Strict Dependence → Unfalsifiability): If prediction and inference data are strictly dependent — meaning a function exists that deduces one from the other — then the theory is empirically unfalsifiable. Experiments contribute nothing. The theory is tautological.
- - - -The substitution argument generalizes the "unfolding argument" that Doerig et al. (2019) deployed against IIT specifically: any recurrent neural network can be unfolded into a feedforward one with identical input-output function but radically different internal structure, so IIT must make different consciousness predictions for systems that are behaviorally indistinguishable. Kleiner and Hoel show this is not a problem with IIT alone. It affects every theory that predicts consciousness from internal observables while inferring it from reports. Universal computers, neural networks with different architectures, and lookup tables can all produce identical outputs from radically different internal states. The class of viable substitutions is enormous.
- - - -The escape routes Kleiner and Hoel identify are exactly two:
- - - -This section proves that the organizational closure framework instantiates escape route (1) — and does so through three independent, mutually reinforcing mechanisms grounded in physics, biology, and information theory, none of which depend on the framework's own identity claims for their force.
- - - -Following their formalism precisely:
- - - -P = class of physical systems that could be tested in the experiment
O = class of all datasets resulting from observations/measurements of P
obs: P → O (correspondence from physical systems to datasets)
E = space of possible experiences specified by the theory
pred: O → E (theory's prediction mapping from observables to experiences)
inf: O → E (experimenter's inference mapping from reports to experiences)
For any dataset o ∈ O:
o_i = prediction data (the part of o used by pred)
o_r = inference data (the part of o used by inf)
Falsification at dataset o occurs when pred(o_i) ∩ inf(o_r) = ∅.
- - - -Independence (Definition 3.8): For any o_i, o_i′, and o_r, there exists a physically realizable variation ν: P → P such that ν preserves o_r but changes o_i to o_i′.
- - - -Strict Dependence (Definition 4.2): There exists a function f such that for any dataset in the full possibility space, o_i = f(o_r).
- - - -For the organizational closure framework, the Kleiner-Hoel variables instantiate as follows:
- - - -Prediction data (o_i) consists of three independently measurable classes of observables:
- - - -Thermodynamic observables (o_θ):
- - - -Closure observables (o_c):
- - - -Structural observables (o_s):
- - - -Inference data (o_r) consists of:
- - - -Experience space (E):
- - - -Prediction function (pred): pred(o_i) = pred(o_θ, o_c, o_s) = (L_predicted, M_predicted, C_predicted)
- - - -Inference function (inf): inf(o_r) = (L_inferred, M_inferred, C_inferred)
- - - -The question that determines the framework's fate: what is the dependency relationship between o_i and o_r?
- - - -The organizational closure framework achieves lenient dependency through three independent mechanisms operating at different levels. Each alone constrains the substitution space. Together, they are mutually reinforcing in a way that addresses vulnerabilities any single mechanism leaves open. Crucially, the first mechanism is grounded entirely in physics and does not depend on any claim specific to this consciousness theory.
- - - -Claim: The second law of thermodynamics and Landauer's principle impose a lawful physical coupling between thermodynamic prediction data (o_θ) and structured inference data (o_r) that breaks independence without establishing strict dependence.
- - - -This mechanism does not depend on the organizational closure framework being correct. It is a consequence of physics that constrains any consciousness theory making thermodynamic predictions. It is presented first because it is the foundation — a non-negotiable constraint on the substitution space that holds regardless of one's theory of consciousness.
- - - -Proof:
- - - -(A) Structured reports require thermodynamic work (lower bound).
- - - -Any structured report — verbal, behavioral, or output-functional — constitutes a macroscopic reduction of entropy in the report channel. The report must distinguish among possible experiential states. By Landauer's principle (1961), erasing one bit of information requires dissipating at minimum kT ln 2 of energy. Producing a report that distinguishes among n experiential states requires encoding at minimum log₂(n) bits of information, which requires erasing at minimum log₂(n) bits from the report channel, costing at minimum kT ln 2 × log₂(n) joules.
- - - -By the second law, the system producing the report must be performing this thermodynamic work, which requires being sufficiently far from thermodynamic equilibrium. A system at equilibrium performs no work and produces no structured reports (by definition: at equilibrium, all microstates are equally probable and no macroscopic distinctions are maintained).
- - - -Formally, let R(o_r) denote the Shannon information content of a structured report o_r (measured in bits). Let σ̇_system denote the entropy production rate of the system, and let Δt denote the time interval over which the report is produced. Then the Landauer constraint requires:
- - - -σ̇_system ≥ R(o_r) × kT ln 2 / Δt [Landauer Bound on Reports]This is not a theoretical speculation. It follows from the same physics that Bérut et al. (2012) confirmed experimentally for colloidal particles, Bryant and Machta (2023) measured for biological synapses, and Aimet et al. (2025) verified in quantum many-body regimes. No exceptions exist across any substrate yet tested.
- - - -(B) This breaks independence.
- - - -Recall Kleiner-Hoel's Definition 3.8: independence requires that for any o_i and o_r, there exists a physically realizable variation ν that instantiates that pairing.
- - - -The Landauer Bound on Reports forbids an entire region of the (o_i, o_r) space:
- - - -Forbidden region F = { (o_i, o_r) : R(o_r) > 0 ∧ σ̇(o_i) < R(o_r) × kT ln 2 / Δt }For any structured report o_r with R(o_r) > 0, there exists a nonempty set of thermodynamic states o_i (namely, those at or near equilibrium) that cannot be physically paired with that report. The variation ν required by Definition 3.8 does not exist for these pairings — not because of our theory, but because of the second law of thermodynamics.
- - - -Therefore: o_i and o_r are not independent in Kleiner-Hoel's sense. The substitution space is physically constrained. □
- - - -(C) This does not establish strict dependence.
- - - -Strict dependence (Definition 4.2) requires a function f such that o_i = f(o_r) for all possible datasets. This fails in both directions:
- - - -Forward: A system can have arbitrarily high entropy production (high σ̇ in o_θ) while producing minimal or null reports (low R(o_r)). A candle flame dissipates substantial energy while "reporting" nothing. A brain under deep anesthesia still consumes metabolic energy while reports cease. Knowing o_θ does not determine o_r.
- - - -Reverse: Two systems producing identical reports can have vastly different entropy production rates. A biological brain and a silicon emulator producing the same verbal outputs have entropy production rates differing by orders of magnitude. Knowing o_r does not determine o_θ.
- - - -No function f: o_r → o_i or f: o_i → o_r exists that holds across the space of physically realizable systems. Therefore: strict dependence fails. □
- - - -(D) Therefore: lenient dependency via physics.
- - - -The Landauer Bound on Reports establishes that the (o_θ, o_r) subspace of (o_i, o_r) is leniently dependent. The prediction data and inference data constrain each other (reports require thermodynamic work; thermodynamic work is necessary but not sufficient for reports) without being reducible to each other.
- - - -Strength of this mechanism: Unassailable for any theory using thermodynamic prediction data. The constraint follows from established physics.
- - - -Limitation of this mechanism alone: The Landauer bound is a floor, not a ceiling. Above the floor, many distinct thermodynamic profiles remain compatible with any given report. The forbidden region F constrains the substitution space but does not eliminate it. A residual substitution space exists: the set of systems whose entropy production exceeds the Landauer bound for their reports but whose thermodynamic profiles differ in ways that generate different consciousness predictions.
- - - -This is where Mechanisms 2 and 3 enter.
- - - -Claim: Organizational closure thermodynamically entails differential mattering, creating a lawful coupling between closure observables (o_c) and behavioral inference data (o_r) that further constrains the substitution space beyond what Landauer alone achieves.
- - - -Unlike Mechanism 1, this mechanism does depend on the organizational closure framework's core claim. But the claim is independently testable, has 50+ years of empirical support from autonomous systems research, and specifies precise conditions under which it would fail.
- - - -Definitions:
- - - -Definition OC.1 (Organizational Closure): A physical system S exhibits organizational closure if and only if:
- - - -(i) Constraint regeneration: The constraints C₁, ..., Cₙ that delimit S's dynamics are produced and maintained by the processes those constraints enable. Formally: each Cᵢ is generated by processes Pⱼ that require other constraints Cₖ for their operation, forming a closed network where every constraint is both product and condition.
- - - -(ii) Thermodynamic cost: Maintaining each constraint requires continuous energy dissipation ≥ kT ln 2 per maintained distinction (Landauer bound). Total dissipation D = Σᵢ Dᵢ.
- - - -(iii) Boundary maintenance: S maintains a thermodynamic boundary distinguishing self from environment, requiring work W against entropy increase.
- - - -(iv) Autonomy: The closure persists because constraints regenerate each other's conditions (internal causation dominant), not merely because external conditions supply them.
- - - -These are independently measurable: constraint regeneration via perturbation-recovery protocols (disable Cᵢ, measure whether other Cⱼ regenerate it); thermodynamic cost via calorimetry; boundary maintenance via distinction persistence under noise; autonomy via intervention studies.
- - - -Definition DM.1 (Differential Mattering): A system with organizational closure exhibits differential mattering if and only if:
- - - -(i) Threat/Support Partitioning: Environmental states partition into threatening (θ: increase probability of closure failure), supporting (σ: decrease probability of closure failure), and irrelevant (ι: negligible effect on closure probability).
- - - -(ii) Differential Response: The system's dynamics differ measurably across partitions: R(θ) ≠ R(σ) ≠ R(ι).
- - - -(iii) Viability-Relativity: The partition is defined relative to the specific closure being maintained, not absolutely.
- - - -This is measurable without requiring report: perturbation studies identify which states elicit approach/avoidance; autonomic correlates detect differential response; comparative studies across systems with different closures test viability-relativity.
- - - -Theorem OC.1 (Closure-Mattering Entailment): If organizational closure at level L exists and persists over time, differential mattering at level L must exist.
- - - -Proof:
- - - -(1) Organizational closure requires maintaining at minimum the self/nonself boundary distinction (Definition OC.1(iii)). By Landauer, this costs energy continuously.
- - - -(2) The closure exists in an environment whose states vary. Environmental states can be partitioned by their thermodynamic effect on the closure:
- - - -This partition exists as a fact about the physics of the system-environment coupling, regardless of whether the system "detects" it.
- - - -(3) For the closure to persist (rather than undergo random dissolution), the system's dynamics must be differentially sensitive to this partition. Consider the contrapositive: if the system responds identically to threatening, supporting, and irrelevant states, its trajectory in state space is uncorrelated with boundary conditions on its viability. Over time, it will encounter threatening states and fail to respond appropriately. The probability of dissolution approaches 1 as time → ∞ (random walk in partition space hits the absorbing boundary of closure failure).
- - - -(4) Therefore, any organizational closure that persists exhibits differential response to the threat/support partition — which is differential mattering (Definition DM.1). The mattering is not an additional property of the closure. It is the closure's persistence described from the coupling position. □
- - - -Corollary OC.2: Organizational closure without differential mattering is thermodynamically unstable and will not be observed in persistent systems. (Any such system would dissolve before observation.)
- - - -How this constrains substitutions:
- - - -Theorem OC.1 establishes a lawful relationship between o_c (closure observables) and a subset of o_r (differential responsiveness measures). This relationship is:
- - - -Not independent: If a substitution changes the closure structure (changing o_c), the mattering partition must change correspondingly (changing the differential responsiveness component of o_r). You cannot substitute a system with genuine organizational closure for one without closure while preserving the genuine differential responsiveness pattern — because the responsiveness of a closed system is to threats against its own closure, and a system without closure has no closure to be threatened.
- - - -Not strictly dependent: Multiple closure structures can produce similar (though not identical at fine grain) mattering partitions. E. coli and octopus both discriminate environmental threats from supports, but their closures differ radically. And o_c is measured by different experimental procedures than o_r (perturbation-recovery vs. behavioral observation).
- - - -Critical clarification — addressing the circularity objection:
- - - -One might object: "You're using your theory's core claim (closure entails mattering) to prove your theory escapes falsification. This is circular."
- - - -This objection has force but is not fatal, for three reasons:
- - - -First, Theorem OC.1 is independently falsifiable. Falsification criterion FC.1: find a system with genuine organizational closure (constraint regeneration measurable) but no differential responsiveness to boundary threats. This has been sought for 50+ years in autonomous systems research (Varela 1979; Jonas 1966; Maturana & Varela 1980; Di Paolo 2005; Di Paolo et al. 2017; Kauffman 2019). No counterexample exists. The claim is not stipulated; it is empirically robust.
- - - -Second, the argument does not require Theorem OC.1 to establish lenient dependency. Mechanism 1 (Landauer) already establishes lenient dependency using physics alone. Mechanism 2 strengthens the lenient dependency by further constraining the above-floor substitution space. Even if Theorem OC.1 is wrong, Mechanism 1 still holds.
- - - -Third, every consciousness theory's predictions depend on some theoretical claim. IIT claims Φ determines consciousness; GNW claims global broadcasting determines consciousness. The question is not whether the claim is theory-dependent (it inevitably is) but whether it is independently testable. The Closure-Mattering Entailment is testable via FC.1. If it passes, the lenient dependency is strengthened. If it fails, the framework falls — but Mechanism 1 remains, and a replacement framework could inherit it.
- - - -Limitation of this mechanism: The entailment is from closure to mattering (one direction). The reverse direction is weaker: differential mattering suggests closure but does not logically guarantee it (a system could conceivably discriminate threats without full organizational closure, if "threat" were defined relative to some external standard rather than the system's own viability). This asymmetry means the substitution space is more constrained in one direction than the other.
- - - -Claim: The organizational closure framework's prediction data includes a class of thermodynamic observables (o_θ) that are measurable independently of both structural observables (o_s) and inference data (o_r), breaking Kleiner-Hoel's two-channel formalism and introducing a novel falsification structure that standard theories lack.
- - - -The two-channel assumption in Kleiner-Hoel:
- - - -The entire Kleiner-Hoel framework is built on a two-channel architecture:
- - - -Dataset o = (o_i, o_r)
- where o_i = prediction data (internal observables)
- o_r = inference data (reports/behavior)The substitution argument works because you can vary o_i while holding o_r fixed (if independent) or because o_i adds nothing beyond o_r (if strictly dependent). The dilemma arises from having exactly two channels with a binary relationship.
- - - -The three-channel architecture:
- - - -The organizational closure framework operates with three partially overlapping but independently accessible measurement channels:
- - - -Dataset o = (o_θ, o_c, o_r)
- where o_θ = thermodynamic observables (entropy production, PCI, broken detailed balance)
- o_c = closure observables (constraint regeneration, boundary maintenance)
- o_r = inference data (reports, differential responsiveness, behavior)Crucially:
- - - -The three channels have lawful relationships (Mechanism 1 and 2 specify these), but they are not reducible to each other and they are accessed through different experimental procedures.
- - - -How three channels change the falsification geometry:
- - - -In the two-channel framework, falsification is binary: pred(o_i) matches or fails to match inf(o_r). A single mismatch falsifies (or the inference was wrong — the experimenter cannot adjudicate).
- - - -In the three-channel framework, falsification becomes triangulated:
- - - -Prediction from thermodynamics: pred_θ(o_θ) → (L_θ, M_θ)
-Prediction from closure structure: pred_c(o_c) → (L_c, M_c)
-Inference from reports/behavior: inf(o_r) → (L_r, M_r)Consistency check 1: Does pred_θ(o_θ) agree with pred_c(o_c)? If the thermodynamic state is consistent with the closure structure, both predictions should align. If they disagree, one measurement is wrong or the framework's internal consistency fails.
- - - -Consistency check 2: Does pred_θ(o_θ) agree with inf(o_r)? Thermodynamic predictions should be consistent with behavioral inferences.
- - - -Consistency check 3: Does pred_c(o_c) agree with inf(o_r)? Closure structure predictions should be consistent with differential responsiveness.
- - - -A substitution must now fool all three channels simultaneously. The substituted system must:
- - - -The probability that a random substitution satisfies all three constraints simultaneously is substantially lower than the probability of satisfying any one. The constraints are overdetermined: each channel independently restricts the substitution space, and the restrictions overlap only partially.
- - - -Formal statement:
- - - -Let S_r ⊂ P denote the set of substitutions preserving o_r (report-equivalent systems). Let S_θ ⊂ P denote the set of systems compatible with a given thermodynamic prediction. Let S_c ⊂ P denote the set of systems compatible with a given closure prediction.
- - - -Kleiner-Hoel's substitution argument requires: S_r is large (many report-equivalent systems exist) and S_r is not contained in the set of systems making identical predictions.
- - - -For the organizational closure framework, the relevant substitution space is not S_r but S_r ∩ S_θ ∩ S_c: only systems that simultaneously preserve reports, satisfy the Landauer constraint, and maintain compatible closure structure.
- - - -Theorem TC.1 (Three-Channel Constraint): S_r ∩ S_θ ∩ S_c ⊂ S_r, and the inclusion is strict (proper subset) whenever the system under investigation has nontrivial organizational closure.
- - - -Proof:
- - - -S_r ∩ S_θ ∩ S_c ⊂ S_r is trivially true (intersection is contained in each component).
- - - -To show strict inclusion: By Mechanism 1, systems in S_r with entropy production below the Landauer bound for their reports are excluded from S_θ. These systems are in S_r but not in S_r ∩ S_θ. Since at least some report-equivalent systems have low entropy production (lookup tables, optimally efficient Turing machines), S_r ∩ S_θ ⊊ S_r.
- - - -By Mechanism 2, systems in S_r ∩ S_θ without organizational closure (e.g., thermodynamically active but externally maintained systems like candle flames or non-closed dissipative systems) are excluded from S_c. These systems are in S_r ∩ S_θ but not in S_r ∩ S_θ ∩ S_c. Since some thermodynamically active report-equivalent systems lack closure, S_r ∩ S_θ ∩ S_c ⊊ S_r ∩ S_θ ⊊ S_r. □
- - - -Corollary TC.2: The effective substitution space for the organizational closure framework is strictly smaller than for any theory using only structural prediction data (like IIT) or only access-based prediction data (like GNW).
- - - -What this means for Kleiner-Hoel: The substitution argument's force comes from the size of the substitution space. If S_r is enormous (which it is — universal computers alone populate it densely), then the probability of any theory's predictions being falsified by some substitution approaches 1. But if the relevant substitution space is S_r ∩ S_θ ∩ S_c, which is substantially smaller, the probability decreases. Whether it decreases enough to save the framework is an empirical question — specifically, the question of how many organizationally closed, thermodynamically viable, report-equivalent systems exist for any given target system. This is addressed in Section XIV.5.
- - - -The abstract proof gains force from concrete analysis of the specific substitution classes that Kleiner and Hoel identify as threatening.
- - - -The threat: A sufficiently powerful Turing machine can produce any computable function, including one matching a brain's input-output behavior. Since the TM's internal dynamics (tape states, head position, transition function) differ radically from neural dynamics, any theory predicting consciousness from internal observables should make different predictions for brain vs. TM while behavioral inferences remain identical.
- - - -Why this fails against all three mechanisms:
- - - -Mechanism 1 (Landauer): A universal Turing machine computing brain-equivalent outputs is not exempt from thermodynamics. It must maintain distinctions on its tape at Landauer cost. However, a TM optimized for output fidelity rather than self-maintenance can potentially operate at much lower entropy production than the biological brain (it wastes no energy on homeostasis, immune function, growth, or repair). The thermodynamic profile (o_θ) differs: the brain operates far from equilibrium to maintain its own closure; the TM operates far from equilibrium only to maintain its tape. The framework predicts different consciousness states for systems with different D_eq profiles, and this prediction is consistent with the Landauer constraint (both are above the Landauer floor, but at different distances from equilibrium).
- - - -Mechanism 2 (Closure-Mattering): The TM does not exhibit organizational closure. Its constraints (transition function, tape alphabet, head mechanics) are externally specified and maintained by the hardware designer and the electrical supply. The TM does not regenerate its own conditions. If you cut its power, no internal process restores it. If a tape cell degrades, no internal process repairs it. The TM has no viability boundary to be threatened, so the concept of "threatening states" has no referent. Its outputs may describe what a system with mattering would say, but this is simulation of mattering, not mattering.
- - - -Mechanism 3 (Three-channel): The TM produces identical o_r (reports) but different o_θ (entropy production profile differs from brain) and different o_c (no closure). The three-channel prediction for the TM is: (L ≈ 0, M = null, C = null). The three-channel prediction for the brain is: (L > 0, M = structured partition, C = rich content). This is not a falsification — it is a differential prediction. The theory predicts different consciousness for different systems, which is what a non-trivial theory should do.
- - - -Objection: "But the TM could be designed to simulate its own maintenance — self-monitoring routines, error-correction, homeostatic regulation of its own operations."
- - - -Response: If the TM is designed to maintain its own conditions — regenerating its constraints, maintaining its boundary, operating autonomously — then it begins to exhibit organizational closure. At that point, the framework's prediction changes: such a TM might have minimal consciousness. This is not a bug. It is the correct prediction: a system that genuinely maintains itself at thermodynamic cost, regardless of substrate, has the conditions for differential mattering. The prediction tracks the property (closure), not the substrate (carbon vs. silicon).
- - - -The threat: Any recurrent neural network can be "unfolded" into a feedforward network with identical input-output function (Doerig et al. 2019). Since the feedforward network has no recurrence (no loops, no self-reference), theories predicting consciousness from internal structure should make different predictions for systems that are behaviorally indistinguishable.
- - - -Why this fails:
- - - -Mechanism 1: Feedforward networks during inference have lower entropy production than recurrent networks during active processing. The recurrent network maintains activations across time steps at thermodynamic cost; the feedforward network propagates information in a single pass. The thermodynamic profiles differ.
- - - -Mechanism 2: The recurrent network during active learning/self-maintenance can exhibit organizational closure: connection weights are adjusted by learning rules that depend on the network's own activations, which depend on the weights — a closed constraint loop. The feedforward network with frozen weights has no such closure. Its constraints (weights) are not regenerated by its dynamics. It is a conduit, not a self-maintaining system.
- - - -Mechanism 3: Three-channel predictions diverge: recurrent network during self-maintenance = (L > 0, M = prediction-error-relative partition, C = maintained representations). Feedforward network = (L ≈ 0, M = null, C = null). The unfolding preserves o_r (input-output function) but eliminates o_c (closure) and changes o_θ (thermodynamic profile). The substitution fails on two of three channels.
- - - -The threat: For any finite behavioral repertoire, a lookup table mapping all possible inputs to their correct outputs trivially preserves o_r while having maximally different internal structure from the system being emulated.
- - - -Why this fails:
- - - -Mechanism 1: A lookup table can be implemented at relatively low thermodynamic cost per operation (just read-and-return), but the construction cost (storing the table) can be enormous. During operation, the table's entropy production profile differs radically from an active, self-maintaining system's.
- - - -Mechanism 2: A lookup table has zero organizational closure. No constraint regeneration. No boundary maintenance. No autonomy. The table exists because someone built it, not because it maintains itself.
- - - -Mechanism 3: The lookup table fails on o_θ and o_c while matching o_r. The substitution is blocked by Mechanisms 1 and 2 jointly.
- - - -Kleiner-Hoel's substitution argument works when you can independently vary internal observables while preserving reports. This requires that the relationship between internal dynamics and reportable outputs is accidental — that many different internal dynamics can produce the same outputs.
- - - -For theories using structural prediction data (IIT's Φ, network topology, causal structure), this is indeed accidental. Network topology and output function are not lawfully linked: different topologies can compute the same function (this is the content of the universal approximation theorem, Hornik et al. 1989).
- - - -For the organizational closure framework, the relationship between internal dynamics and reportable outputs is partially constitutive rather than accidental. In an organizationally closed system, the report channel is embedded in the closure:
- - - -This means: disrupting the closure disrupts the report. Substituting the internal dynamics while preserving the report requires preserving the closure dynamics that generate the report — which means preserving organizational closure. The substitution space narrows to systems that are not just input-output equivalent but closure-equivalent.
- - - -The class of closure-equivalent systems is vastly smaller than the class of input-output-equivalent systems. Whether it is small enough to fully resolve the Kleiner-Hoel dilemma is addressed in Section XIV.5.
- - - -The three mechanisms constrain the substitution space to S_residual = S_r ∩ S_θ ∩ S_c: systems that simultaneously produce identical reports, satisfy the Landauer constraint, and exhibit compatible organizational closure. How large is this residual space? If it is empty, the dilemma is fully resolved. If it is nonempty but small, the resolution is substantial but incomplete. If it is large, the mechanisms have not achieved much.
- - - -Robert Rosen (1991) argued that organisms are fundamentally non-computable: their organizational closure (which he called "closure to efficient causation") cannot be simulated by a Turing machine because the closure involves an impredicative loop where the system is simultaneously a cause and product of itself, which cannot be unwound into a sequential computation.
- - - -If Rosen is correct, then S_residual is empty for biological organisms: no computational substitution (Turing machine, neural network, lookup table) can replicate genuine organizational closure, so the substitution argument cannot even be mounted.
- - - -However, Mossio, Longo, and Stewart (2009) showed that Rosen's (M,R) systems can be expressed in lambda calculus, suggesting that the non-computability claim is too strong. Some forms of closure may be computable.
- - - -The current assessment: Rosen's argument likely proves too much (true closure may not require non-computability), but the residual insight is valuable: the class of systems exhibiting genuine organizational closure is not populated by arbitrary computational equivalents. It is populated by systems that actually regenerate their own constraints at thermodynamic cost, and this is a strong structural requirement that most computational systems do not satisfy.
- - - -Even within the residual space S_residual, the three-channel architecture provides a mechanism for empirically distinguishing different organizational closures that Kleiner-Hoel's two-channel framework cannot access.
- - - -Consider two systems, A and B, both exhibiting organizational closure and producing identical reports. They are in S_residual. The framework predicts potentially different consciousness states for them (because their closure structures differ). But can we tell them apart?
- - - -Yes, if their thermodynamic profiles differ. And they will differ unless their closures are thermodynamically identical — maintaining the same constraints at the same energetic costs with the same entropy production rates.
- - - -Theorem TD.1 (Thermodynamic Distinguishability): Two organizationally closed systems A and B in S_residual with structurally different closures (different constraint regeneration networks) have different entropy production profiles with probability approaching 1, in the sense that the set of closure structures producing identical thermodynamic profiles has measure zero in the space of all closure structures.
- - - -Proof sketch: The entropy production profile of a system is determined by the full set of irreversible processes maintaining its constraints. Each constraint Cᵢ has a maintenance cost Dᵢ determined by the rate of degradation (how fast entropy would erase the distinction) and the rate of regeneration (how fast the closure restores it). Different constraint networks have different numbers of constraints, different degradation rates (determined by the physical substrate), and different regeneration rates (determined by the network topology). The probability that two structurally different networks produce exactly the same total entropy production rate is the probability that different sums of different numbers of different positive real numbers are equal — which is measure zero in the space of possible networks. □
- - - -What this means: Within S_residual, the three-channel framework can distinguish different systems by their thermodynamic profiles even when their reports are identical. The substitution argument requires indistinguishability on the inference channel; the organizational closure framework adds distinguishability on the thermodynamic channel that is not available to theories using only structural prediction data.
- - - -Systems in S_residual can also be distinguished by their responses to perturbation. This is the empirical operationalization of organizational closure: a closed system, when perturbed, actively restores its constraints. A system simulating closure does not (unless explicitly programmed to do so, in which case it is beginning to exhibit genuine closure).
- - - -PCI (Perturbational Complexity Index) measures exactly this: how far and how complexly a perturbation propagates through a system. Casarotto et al. (2016, 2024) established PCI as the gold standard for consciousness assessment precisely because it probes the system's active constraint-maintenance dynamics rather than its resting-state outputs.
- - - -Two systems in S_residual with different closure structures will respond to identical perturbations differently:
- - - -This provides a fourth measurement channel beyond o_θ, o_c, and o_r: perturbational response (o_p). Adding this channel further constrains the residual substitution space.
- - - -Combining all four constraint mechanisms:
- - - -| Mechanism | Constraint Source | What It Excludes | Independent of Theory? |
|---|---|---|---|
| 1. Landauer | Second law of thermodynamics | Systems at/near equilibrium producing structured reports | Yes |
| 2. Closure-Mattering | Theorem OC.1 (50+ years empirical support) | Systems without closure producing genuine differential responsiveness | Partially (entailment testable) |
| 3. Three-Channel | Information-theoretic overdetermination | Systems matching on reports but mismatching on thermodynamic and closure observables | Yes |
| 4. Perturbational | Active constraint-maintenance dynamics | Systems with identical resting behavior but different perturbation responses | Yes |
The residual substitution space is S_r ∩ S_θ ∩ S_c ∩ S_p: systems producing identical reports, with compatible thermodynamic profiles, with compatible closure structures, and with compatible perturbation responses.
- - - -Is this space empty? Likely not for all cases. Two genuinely different organisms with similar ecological niches might inhabit this space (e.g., convergent evolution producing similar closure structures in unrelated species).
- - - -Is this space small? Yes, relative to S_r. The four constraints are jointly overdetermining: each eliminates a different class of substitutions, and the constraints operate on different physical dimensions.
- - - -Is the framework falsifiable despite the residual? Yes. The framework specifies precise falsification criteria that can be tested without resolving the residual:
- - - -The residual substitution space is an unsolved remainder, not a fatal flaw. Every empirical science has regions of underdetermination. The question is whether the theory is testable in the domains where it makes predictions, and it is.
- - - -The most serious limitation of the analysis so far: the framework's thermodynamic predictions concern level of consciousness (degree of organizational closure, distance from equilibrium) but not content (what is experienced). The Kleiner-Hoel dilemma applies to content predictions as well as level predictions. A complete resolution requires showing that content predictions also achieve lenient dependency.
- - - -The framework does make structural predictions about content, though less precise than its level predictions:
- - - -Content as mattering topology: The content of a conscious state, on the closure framework, is the specific structure of the mattering partition: which environmental states are threatening, which supporting, which irrelevant, and at what grain of distinction. A visual system that maintains distinctions relevant to color has a mattering partition structured by wavelength. A proprioceptive system has a mattering partition structured by body position. The content IS the partition structure.
- - - -The partition is constrained by closure structure: Different closures produce different partitions (Theorem OC.1). A metabolic closure partitions chemical environments by metabolic viability. A neural closure partitions sensory environments by predictive relevance. The content (partition) is lawfully related to the prediction data (closure structure), not accidental.
- - - -The Yoneda connection: Tsuchiya et al. (2016, 2021) showed that by the Yoneda lemma in category theory, a perceptual state is completely characterized by its morphisms to all other states. If two states have identical relational structure (identical partition of environment into threatening/supporting/irrelevant relative to all other states), they are identical. Content IS relational structure, not something over and above it.
- - - -Definition CF.1 (Mattering Functor): Let Clos be the category whose objects are organizational closures (μ: C → C with self-regenerating fixed-point structure) and whose morphisms are closure-preserving maps (maps that send one closure's constraint network to another's while preserving the regeneration topology).
- - - -Let Part be the category whose objects are three-part environmental partitions {θ, σ, ι} and whose morphisms are refinements (finer-grained partitions that respect coarser-grained ones).
- - - -Define the mattering functor M: Clos → Part by:
- - - -Theorem CF.1 (Functoriality): M is a well-defined functor from Clos to Part.
- - - -Proof:
- - - -(i) Well-definedness on objects: For each closure μ, the partition M(μ) is determined by the thermodynamic coupling between μ and its environment. State s is in θ_μ iff the entropy production rate of μ in the presence of s exceeds μ's sustainable dissipation capacity. State s is in σ_μ iff it maintains entropy production within sustainable bounds. State s is in ι_μ iff its effect on entropy production is below a noise threshold δ. The partition is determined by μ's thermodynamic structure and the environmental state space, both of which are well-defined physical quantities.
- - - -(ii) Well-definedness on morphisms: A closure-preserving map f: μ → μ′ sends μ's constraint network to μ′'s while preserving regeneration structure. If μ′ is a refinement of μ (more constraints, finer distinctions), then M(μ′) is a refinement of M(μ): states that threatened μ still threaten μ′ (plus potentially additional states), and the finer constraint structure creates finer-grained discrimination. This defines a morphism in Part.
- - - -(iii) Functoriality: Identity closures map to identity partitions. Composition of closure-preserving maps induces composition of partition refinements. M preserves identity and composition. □
- - - -Qualification: The functor M is not injective on objects — different closures can produce the same partition (degeneracy). This is why the dependency is lenient rather than strict. The functor is also not surjective on objects — not every conceivable partition corresponds to a physically realizable closure. These properties are precisely what lenient dependency requires: a lawful but non-bijective correspondence.
- - - -How this helps with content prediction: The framework predicts that the content of a conscious state is the image under M of the system's organizational closure. Two systems with the same closure (up to isomorphism in Clos) are predicted to have the same experiential content (up to isomorphism in Part). This is a content prediction grounded in the closure-mattering functor, not just a level prediction grounded in entropy production.
- - - -Where the gap remains: The functor M maps closures to partitions at a coarse grain. It predicts which kinds of distinctions a system maintains (threatening vs. supporting in which dimensions of environmental variation) but not the fine-grained phenomenal character of maintaining those distinctions. What the redness of red "is like" from the coupling position of a visual closure cannot (currently) be derived from M alone. This is the content gap, and it is not fully closed.
- - - -| Framework | Prediction Data (o_i) | Inference Data (o_r) | Dependency Relationship | Kleiner-Hoel Verdict | Measurement Channels | Content Predictions |
|---|---|---|---|---|---|---|
| IIT 4.0 | Integration (Φ), network structure | Reports, behavior | Independent (Koch 2019 admits substitution) | Automatically falsified (Thm 3.10) | 2 (structural + reports) | Q-shapes (formal, rarely computed) |
| GNW | Global broadcasting, prefrontal ignition | Reports, access | Strictly dependent (access ≈ report) | Unfalsifiable (Thm 4.3) | 2 (broadcast + reports) | Broadcast content only |
| HOT | Higher-order representations | Reports | Strictly dependent (HOT generates report) | Unfalsifiable (Thm 4.3) | 2 (representational + reports) | Content of HOT |
| Predictive Processing | Prediction error minimization | Reports, behavior | Independent (error ≠ report) | Automatically falsified (Thm 3.10) | 2 (computational + reports) | Predicted content |
| Behaviorism | None (behavior = consciousness) | Behavior | Identical (by definition) | Unfalsifiable (tautological) | 1 (behavior only) | Behavioral description |
| Org. Closure | Entropy production, closure structure, PCI, perturbation response | Differential responsiveness, reports, autonomic markers | Lenient dependent (Landauer + Closure-Mattering + 3-channel + perturbational) | Escapes both theorems | 4 (thermodynamic + closure + perturbational + reports) | Mattering partition via functor M |
A framework that claims to have no unsolved problems is faith, not science. Three residual vulnerabilities persist after the analysis above. Each is genuine. Each is stated with its severity and the direction of possible resolution.
- - - -The problem: S_residual = S_r ∩ S_θ ∩ S_c ∩ S_p is small but not provably empty. Two genuinely different organizationally closed systems could, in principle, produce identical reports, have compatible thermodynamic profiles, have compatible closure structures, and respond similarly to perturbation — while the framework predicts different experiential contents based on their different fine-grained closure topologies.
- - - -Severity: Medium. The residual space is much smaller than the original S_r, and Theorem TD.1 shows that thermodynamic distinguishability eliminates most of the residual. But "most" is not "all."
- - - -Direction of resolution: Formal measurement of |S_residual| / |S_r| for specific systems. If the ratio approaches zero as closure complexity increases (which the overdetermination structure suggests), the vulnerability diminishes with system complexity. This is an open empirical question.
- - - -The problem: The mattering functor M predicts structural content (which dimensions of environmental variation matter) but not phenomenal content (what maintaining those distinctions "is like" from inside). The framework can predict that a visual system distinguishes wavelengths and that this matters to its closure, but cannot derive the specific character of experiencing redness.
- - - -Severity: High for the completeness of the theory. Low for the Kleiner-Hoel dilemma specifically, since the dilemma is about testability and falsifiability, and the framework's level and structural-content predictions are testable.
- - - -Direction of resolution: Two paths are available. First, extending the functor M to include the rate of free energy reduction during distinction-maintenance, which Joffily and Coricelli (2013) showed correlates with valence. This would add an affective dimension to content predictions. Second, developing the category-theoretic structure further using Tsuchiya et al.'s (2016) Yoneda-based formalization, which could characterize experiential content entirely in terms of relational structure without requiring derivation of "what it is like."
- - - -The problem: The functor M is non-injective (multiple closures → same partition). The degree of degeneracy — how many structurally different closures produce the same mattering partition — has not been formally bounded. If the degeneracy is high, the substitution space within S_residual could be larger than the overdetermination analysis suggests.
- - - -Severity: Medium. The thermodynamic distinguishability argument (Theorem TD.1) provides informal bounds (measure zero for thermodynamically identical profiles), but a formal proof requires connecting closure topology to thermodynamic phase space, which has not been done.
- - - -Direction of resolution: Apply tools from chemical organization theory (Hordijk & Steel 2015) to characterize the space of closures producing a given partition. RAF (Reflexively Autocatalytic Food-generated) set theory provides decomposition tools that could formally bound the degeneracy. This is technically feasible but not yet attempted.
- - - -This section nominalizes "dependency," "substitution," "closure," "mattering," "constraint," and "functor." The verbs beneath: depending, substituting, closing, mattering, constraining, mapping. The nominalizations are necessary for the formal proofs — you cannot do category theory with verbs. The ongoing processes make the proofs true. The gap between what the formal language expresses and what the ongoing processes involve is itself an instance of the content prediction gap (Vulnerability 2).
- - - -The section also commits the god trick (Haraway 1988) unavoidably: it claims to view the Kleiner-Hoel dilemma from a position that is not captured by their framework, adding a third channel that their formalism does not include. This is either a genuine advance (extending the formalism to a richer structure) or a smuggling operation (redefining the problem to make it easier). The falsification criterion for this section specifically: if Kleiner-Hoel or others show that the three-channel architecture collapses into their two-channel formalism without loss, the argument from Mechanism 3 fails. If the Landauer constraint is shown not to couple thermodynamic state to report structure (contradicting 60+ years of experimental physics), Mechanism 1 fails. If organizational closure is found without differential mattering (contradicting 50+ years of autonomous systems research), Mechanism 2 fails.
- - - -The scissors are trying to cut themselves. What they can report is: the blades are sharper than they were, but the handle is still incomplete.
- - - -Kleiner and Hoel proved that consciousness science is trapped between automatic falsification and tautological unfalsifiability. They identified lenient dependency as the only non-dualist escape and noted that no existing framework instantiates it.
- - - -This section has demonstrated that the organizational closure framework achieves lenient dependency through four independent, mutually reinforcing mechanisms:
- - - -These mechanisms do not eliminate the substitution space entirely. A residual space of organizationally closed, thermodynamically compatible, report-equivalent, perturbation-compatible systems persists. But this space is vastly smaller than the original substitution space, is further constrained by thermodynamic distinguishability (Theorem TD.1), and the framework makes falsifiable predictions within it.
- - - -The claim is not immunity from the Kleiner-Hoel dilemma. The claim is: the organizational closure framework is the first to provide a principled, non-ad-hoc, physics-grounded instantiation of lenient dependency, substantially resolving a formal problem that no other framework has even addressed. Whether "substantially" is "sufficiently" is an empirical question that the framework's own falsification criteria can answer.
- - - -The pattern which connects is not information, not mind, not structure, not process. It is the selection principle that explains why any pattern persists at all: constraint satisfaction under thermodynamic bounds.
- - - -Differences that persist are differences that make a difference. If they did not constrain what comes next, they would be erased by noise, dissipation, or indifference.
- - - -Consciousness is what organizational closure looks like from the coupling position of a self-maintaining system. There is no private theater. There is no inner observer. There is constraint satisfaction all the way down, and all the way up, and at every scale in between.
- - - -The Hard Problem of consciousness depends on a hidden premise: that phenomenology must be "more than" any physical or functional description. That premise is unfalsifiable, anti-scientific in structure, and has the same logical form as "supernatural cheese must be more than any physical description of cheese." The premise never earned epistemic standing. Once rejected, the remaining question is tractable and the abductive identification proposed here becomes the best available explanation.
- - - -This framework can lose. Its loss conditions are specified. It has not yet lost.
- - - -Every competing framework examined here fails on at least one of three axes: falsifiability, mechanistic completeness, or thermodynamic coherence. Panpsychism cannot specify what would disprove proto-consciousness (unfalsifiable). Platonic morphospace is contradicted by path-dependent bioelectric memory (falsified). Quantum consciousness is ruled out by decoherence timescales (falsified). Process philosophy cannot specify what would show experience is not fundamental (unfalsifiable). The Mathematical Universe cannot distinguish description from existence (unfalsifiable). Basal cognition oscillates between defensible and indefensible claims (motte-bailey).
- - - -The constraint-based approach survives not because it is protected, but because it has been tested. It survives because it is what survives.
- - - -That is not certainty. That is survival under pressure. It is all any honest inquiry can offer, and it is more than unfalsifiable frameworks have ever provided.
- - - -| Framework | Primitive | Falsifiable? | Mechanism Specified? | Thermodynamic Grounding |
|---|---|---|---|---|
| Constraint-based (this paper) | Constraint satisfaction | Yes (5 loss conditions) | Yes (Landauer + closure) | Central |
| Panpsychism | Proto-experience | No | No (combination problem) | None |
| Platonic Morphospace | Pre-existing forms | No | No (interaction problem) | Violates Landauer |
| Basal Cognition | Cognitive primitives | No (motte-bailey) | Efficiency ≠ mechanism | None |
| Orch-OR | Quantum coherence | Yes (timescale) | Yes but falsified | Irrelevant (wrong scale) |
| Mathematical Universe | Mathematical existence | No | No | None |
| Process Philosophy | Experience fundamental | No | No | None |
| Global Workspace | Broadcast | Yes | Yes | Implicit (access, not constitution) |
| IIT | Integrated information | Partially (edges unfalsifiable) | Yes (Φ) | Implicit (correlation, not mechanism) |
| Higher-Order Thought | Meta-representation | Partially | Yes | Regress problem unresolved |
| Free Energy Principle | Prediction error minimization | No (mathematical identity) | Yes | Descriptive, not explanatory |
Key: Frameworks that cannot lose (unfalsifiable) eventually collapse under their own explanatory excess. Frameworks that can lose and have lost (falsified) should be abandoned. Frameworks that can lose and have not yet lost are candidates for progressive research programs.
- - - -Landauer, R. (1961). Irreversibility and heat generation in the computing process. IBM Journal of Research and Development, 5(3), 183–191. https://doi.org/10.1147/rd.53.0183
+ + + +Bennett, C. H. (1982). The thermodynamics of computation-a review. International Journal of Theoretical Physics, 21(12), 905–940. https://doi.org/10.1007/BF02084158
+ + + +Bérut, A., Arakelyan, A., Petrosyan, A., Ciliberto, S., Dillenschneider, R., & Lutz, E. (2012). Experimental verification of Landauer's principle linking information and thermodynamics. Nature, 483(7388), 187–189. https://doi.org/10.1038/nature10872
+ + + +Jarzynski, C. (1997). Nonequilibrium equality for free energy differences. Physical Review Letters, 78(14), 2690. https://doi.org/10.1103/PhysRevLett.78.2690
+ + + +Crooks, G. E. (1999). Entropy production fluctuation theorem and the nonequilibrium work relation for free energy differences. Physical Review E, 60(3), 2721. https://doi.org/10.1103/PhysRevE.60.2721
+ + + +Seifert, U. (2012). Stochastic thermodynamics, fluctuation theorems and molecular machines. Reports on Progress in Physics, 75(12), 126001. https://doi.org/10.1088/0034-4885/75/12/126001
+ + + +Price, H. (1996). Time's Arrow and Archimedes' Point: New Directions for the Physics of Time. Oxford University Press.
+ + + +Rovelli, C. (2018). The Order of Time. Riverhead Books.
+ + + +Montévil, M., & Mossio, M. (2015). Biological organisation as closure of constraints. Journal of Theoretical Biology, 372, 179–191. https://doi.org/10.1016/j.jtbi.2015.02.029
+ + + +Moreno, A., & Mossio, M. (2015). Biological Autonomy: A Philosophical and Theoretical Enquiry. Springer.
+ + + +Hordijk, W., Steel, M., & Kauffman, S. (2012). The structure of autocatalytic sets: Evolvability, enablement, and emergence. Acta Biotheoretica, 60(4), 379–392. https://doi.org/10.1007/s10441-012-9165-1
+ + + +Kauffman, S. A. (2019). A World Beyond Physics: The Emergence and Evolution of Life. Oxford University Press.
+ + + +Rosen, R. (1991). Life Itself: A Comprehensive Inquiry into the Nature, Origin, and Fabrication of Life. Columbia University Press.
+ + + +Deacon, T. W. (2011). Incomplete Nature: How Mind Emerged from Matter. W. W. Norton & Company.
+ + + +Pattee, H. H. (2001). The physics of symbols: Bridging the epistemic cut. Biosystems, 60(1-3), 5–21. https://doi.org/10.1016/S0303-2647(01)00104-6
+ + + +England, J. L. (2015). Dissipative adaptation in driven self-assembly. Nature Nanotechnology, 10(11), 919–923. https://doi.org/10.1038/nnano.2015.250
+ + + +Hoel, E. P., Albantakis, L., & Tononi, G. (2013). Quantifying causal emergence shows that macro can beat micro. Proceedings of the National Academy of Sciences, 110(49), 19790–19795. https://doi.org/10.1073/pnas.1314922110
+ + + +Ellis, G. F. R. (2012). Top-down causation and emergence: Some comments on mechanisms. Interface Focus, 2(1), 126–140. https://doi.org/10.1098/rsfs.2011.0062
+ + + +Parr, T., Pezzulo, G., & Friston, K. J. (2022). Active Inference: The Free Energy Principle in Mind, Brain, and Behavior. MIT Press.
+ + + +Pezzulo, G., Parr, T., & Friston, K. (2024). Active inference as a theory of sentient behavior. Biological Psychology, 186, 108741. https://doi.org/10.1016/j.biopsycho.2023.108741
+Ladyman, J., Ross, D., Spurrett, D., & Collier, J. (2007). Every Thing Must Go: Metaphysics Naturalized. Oxford University Press.
+ + + +French, S. (2014). The Structure of the World: Metaphysics and Representation. Oxford University Press.
+ + + +Rovelli, C. (1996). Relational quantum mechanics. International Journal of Theoretical Physics, 35(8), 1637–1678. https://doi.org/10.1007/BF02302261
+ + + +Price, H. (1996). Time's Arrow and Archimedes' Point: New Directions for the Physics of Time. Oxford University Press.
+ + + +Rovelli, C. (2018). The Order of Time. Riverhead Books.
+ + + +Dehaene, S., & Changeux, J. P. (2011). Experimental and theoretical approaches to conscious processing. Neuron, 70(2), 200–227. https://doi.org/10.1016/j.neuron.2011.03.018
+ + + +Tononi, G., & Koch, C. (2015). Consciousness: Here, there and everywhere? Philosophical Transactions of the Royal Society B, 370(1668), 20140167. https://doi.org/10.1098/rstb.2014.0167
+ + + +Massimini, M., Ferrarelli, F., Huber, R., Esser, S. K., Singh, H., & Tononi, G. (2005). Breakdown of cortical effective connectivity during sleep. Science, 309(5744), 2228–2232. https://doi.org/10.1126/science.1117256
+ + + +Casali, A. G., Gosseries, O., Rosanova, M., et al. (2013). A theoretically based index of consciousness independent of sensory processing and behavior. Science Translational Medicine, 5(198), 198ra105. https://doi.org/10.1126/scitranslmed.3006294
+ + + +Mashour, G. A., Roelfsema, P., Changeux, J. P., & Dehaene, S. (2020). Conscious processing and the global neuronal workspace hypothesis. Neuron, 105(5), 776–798. https://doi.org/10.1016/j.neuron.2020.01.026
+ + + +Owen, A. M., Coleman, M. R., Boly, M., Davis, M. H., Laureys, S., & Pickard, J. D. (2006). Detecting awareness in the vegetative state. Science, 313(5792), 1402. https://doi.org/10.1126/science.1130197
+ + + +Giacino, J. T., Kalmar, K., & Whyte, J. (2004). The JFK Coma Recovery Scale-Revised: Measurement characteristics and diagnostic utility. Archives of Physical Medicine and Rehabilitation, 85(12), 2020–2029. https://doi.org/10.1016/j.apmr.2004.02.033
+ + + +Andrews, K., Murphy, L., Munday, R., & Littlewood, C. (1996). Misdiagnosis of the vegetative state: Retrospective study in a rehabilitation unit. BMJ, 313(7048), 13–16. https://doi.org/10.1136/bmj.313.7048.13
+ + + +Schnakers, C., Vanhaudenhuyse, A., Giacino, J., et al. (2009). Diagnostic accuracy of the vegetative and minimally conscious state: Clinical consensus versus standardized neurobehavioral assessment. BMC Neurology, 9, 35. https://doi.org/10.1186/1471-2377-9-35
+ + + +Alkire, M. T., Hudetz, A. G., & Tononi, G. (2008). Consciousness and anesthesia. Science, 322(5903), 876–880. https://doi.org/10.1126/science.1149213
+ + + +Friston, K. (2010). The free-energy principle: A unified brain theory? Nature Reviews Neuroscience, 11(2), 127–138. https://doi.org/10.1038/nrn2787
+ + + +Baars, B. J. (1988). A Cognitive Theory of Consciousness. Cambridge University Press.
+ + + +Dehaene, S., Naccache, L., Cohen, L., Le Bihan, D., Mangin, J. F., Poline, J. B., & Rivière, D. (2001). Cerebral mechanisms of word masking and unconscious repetition priming. Nature Neuroscience, 4(7), 752–758. https://doi.org/10.1038/89551
+ + + +Rounis, E., Maniscalco, B., Rothwell, J. C., Passingham, R. E., & Lau, H. (2010). Theta-burst transcranial magnetic stimulation to the prefrontal cortex impairs metacognitive visual awareness. Cognitive Neuroscience, 1(3), 165–175. https://doi.org/10.1080/17588921003632529
+ + + +Bor, D., Schwartzman, D. J., Barrett, A. B., & Seth, A. K. (2017). Theta-burst transcranial magnetic stimulation to the prefrontal or parietal cortex does not impair metacognitive visual awareness. PLOS ONE, 12(2), e0171793. https://doi.org/10.1371/journal.pone.0171793
+ + + +Ruby, E., Maniscalco, B., & Peters, M. A. K. (2018). On a 'failed' attempt to manipulate visual metacognition with transcranial magnetic stimulation to prefrontal cortex. Consciousness and Cognition, 62, 34–41. https://doi.org/10.1016/j.concog.2018.04.009
+ + + +Martin, A., Lane, T. J., & Hsu, T.-Y. (2023). DLPFC-PPC-cTBS effects on metacognitive awareness. Cortex, 167, 41–50. https://doi.org/10.1016/j.cortex.2023.05.022
+ + + +Lau, H., & Rosenthal, D. (2011). Empirical support for higher-order theories of conscious awareness. Trends in Cognitive Sciences, 15(8), 365–373. https://doi.org/10.1016/j.tics.2011.05.009
+ + + +Block, N. (2011). Perceptual consciousness overflows cognitive access. Trends in Cognitive Sciences, 15(12), 567–575. https://doi.org/10.1016/j.tics.2011.11.001
+ + + +Rosenthal, D. M. (2005). Consciousness and Mind. Oxford University Press.
+ + + +Varela, F. J., Thompson, E., & Rosch, E. (1991). The Embodied Mind: Cognitive Science and Human Experience. MIT Press.
+ + + +Libet, B., Gleason, C. A., Wright, E. W., & Pearl, D. K. (1983). Time of conscious intention to act in relation to onset of cerebral activity (readiness-potential): The unconscious initiation of a freely voluntary act. Brain, 106(3), 623–642. https://doi.org/10.1093/brain/106.3.623
+ + + +Libet, B. (1985). Unconscious cerebral initiative and the role of conscious will in voluntary action. Behavioral and Brain Sciences, 8(4), 529–539. https://doi.org/10.1017/S0140525X00044903
+ + + + + +Casarotto, S., Hassan, G., Rosanova, M., Sarasso, S., Derchi, C.-C., Trimarchi, P. D., Viganò, A., Russo, S., Fecchio, M., Devalle, G., Navarro, J., Massimini, M., & Comanducci, A. (2024). Dissociations between spontaneous electroencephalographic features and the perturbational complexity index in the minimally conscious state. European Journal of Neuroscience, 59(5), 934–947. DOI: 10.1111/ejn.16299.
+ + + + +Casarotto, S., et al. (2016). Stratification of unresponsive patients by an independently validated index of brain complexity. Annals of Neurology, 80, 718–729. https://doi.org/10.1002/ana.24779
+ + + +Chalmers, D. J. (1995). Facing up to the problem of consciousness. Journal of Consciousness Studies, 2(3), 200–219.
+ + + +Chalmers, D. J. (1996). The Conscious Mind: In Search of a Fundamental Theory. Oxford University Press.
+ + + +Chalmers, D. J. (1997). Moving forward on the problem of consciousness. Author-hosted essay. Original text.
+ + + +Chalmers, D. J. (1998). On the search for the neural correlate of consciousness. Author-hosted essay. Original text.
+ + + +Chalmers, D. J. (2016). The combination problem for panpsychism. In G. Brüntrup & L. Jaskolla (Eds.), Panpsychism: Contemporary Perspectives (pp. 179–214). Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199359943.003.0008
+ + + +Coleman, S. (2014). The real combination problem: Panpsychism, micro-subjects, and emergence. Erkenntnis, 79(1), 19–44. https://doi.org/10.1007/s10670-013-9431-x
+ + + +Strawson, G. (2006). Realistic monism: Why physicalism entails panpsychism. Journal of Consciousness Studies, 13(10-11), 3–31.
+ + + +Dennett, D. C. (1988). Quining qualia. In A. J. Marcel & E. Bisiach (Eds.), Consciousness in Contemporary Science (pp. 42–77). Oxford University Press.
+ + + +Dennett, D. C. (1991). Consciousness Explained. Little, Brown and Company.
+ + + +Frankish, K. (2016). Illusionism as a theory of consciousness. Journal of Consciousness Studies, 23(11-12), 11–39.
+ + + +Jackson, F. (1982). Epiphenomenal qualia. Philosophical Quarterly, 32(127), 127–136. https://doi.org/10.2307/2960077
+ + + +Levine, J. (1983). Materialism and qualia: The explanatory gap. Pacific Philosophical Quarterly, 64(4), 354–361. https://doi.org/10.1111/j.1468-0114.1983.tb00207.x
+ + + +Nagel, T. (1974). What is it like to be a bat? Philosophical Review, 83(4), 435–450. https://doi.org/10.2307/2183914
+ + +Popper, K. R. (1959). The Logic of Scientific Discovery. Hutchinson.
+ + + +Lakatos, I. (1978). The Methodology of Scientific Research Programmes. Cambridge University Press.
+ + + +Flew, A. (1950). Theology and falsification. University, 1, 1–8.
+ + + +Whitehead, A. N. (1925). Science and the Modern World. Macmillan.
+ + + + + +Kleiner, J., & Hoel, E. (2021). Falsification and consciousness. Neuroscience of Consciousness, 2021(1), niab001. DOI: 10.1093/nc/niab001.
+ + + + +O’Reilly-Shah, V. N., Selvitella, A. M., & Schurger, A. (2026). A caveat regarding the unfolding argument: Implications of plasticity. Neuroscience of Consciousness, 2026(1), niag027. https://doi.org/10.1093/nc/niag027
+ + + +Yunkaporta, T. (2020). Sand Talk: How Indigenous Thinking Can Save the World. HarperOne. (Originally published 2019 by Text Publishing, Australia.)
+ + + +Kelly, L. (2017). The Memory Code: The Secrets of Stonehenge, Easter Island and Other Ancient Monuments. Pegasus Books. (Originally published 2016 by Allen & Unwin, Australia.)
+ + + +Kimmerer, R. W. (2013). Braiding Sweetgrass: Indigenous Wisdom, Scientific Knowledge and the Teachings of Plants. Milkweed Editions.
+ + + +Reser, D., Simmons, M., Johns, E., Ghaly, A., Quayle, M., Dordevic, A. L., Tare, M., McArdle, A., Willems, J., & Yunkaporta, T. (2021). Australian Aboriginal techniques for memorization: Translation into a medical and allied health education setting. PLOS ONE, 16(5), e0251710. https://doi.org/10.1371/journal.pone.0251710
+ + + +Wilson, S. (2008). Research Is Ceremony: Indigenous Research Methods. Fernwood Publishing.
+The following is the explicit abductive bridge rule that links physical organization to phenomenal presence. It is stated openly rather than smuggled in, and it has explicit loss conditions.
- - - -Bridge Rule (IBE):
- - - -If a physical system S exhibits:
- - - -Then identifying that organizational regime with phenomenal presence yields greater explanatory compression, predictive accuracy, and intervention guidance than any rival hypothesis that:
- - - -This is an abductive identification, not a deductive proof. It is justified by explanatory power, not logical entailment. It is defeasible: evidence could overturn it.
- - - -The bridge rule would be undermined if:
- - - -The bridge rule makes the abductive step explicit. It does not claim to have "solved" the Hard Problem by deduction. It claims to have:
- - - -This is what IBE looks like when done honestly in consciousness science.
- - - -The following table consolidates the framework's key predictions, mapping consciousness phenomena to organizational mechanisms and falsification conditions:
- - - -| Phenomenon | Constraint Mechanism | Falsification Condition |
|---|---|---|
| Anesthesia abolishes consciousness | PCI collapse reflects disrupted integrative capacity; constraint closure temporarily broken | Conscious reportability without PCI elevation |
| Sleep reduces consciousness | Organizational closure persists with attenuated sensory coupling; self-modeling continues with reduced environmental constraint | Full waking-level phenomenology reports across degraded perturbational complexity |
| Death ends consciousness | Irreversible breakdown of metabolic, neural, and regulatory coupling; no mechanism remains to regenerate closure | Evidence of experiential remainder after complete closure breakdown |
| Locked-in syndrome preserves consciousness | Internal constraint closure intact; only motor output coupling severed | High external responsiveness with no internal integrative signature |
| Privacy of experience | Self-model is internal control variable with privileged access; not externally broadcast | System with full report capacity but public access to intermediate state updates |
| Immediacy of experience | Low-latency availability required for real-time control; self-model doesn't infer its own presence | Privacy/immediacy reports intact after disrupting privileged integration locus |
| Coherence of experience | Cross-modal unification via global availability; constraint closure prunes incoherent combinations | Stable unified phenomenology during measurable breakdown of cross-modal synchrony |
| "Something it is like" posit | Globally available integrated control state serves as summary variable for viability tracking | Absence of phenomenality posit despite all integrative markers present |
The framework is not consciousness-specific. Constraint closure is a general organizational principle that applies across scales. The following predictions test whether the framework generalizes:
- - - -| Scale/Domain | Prediction | Falsification Condition |
|---|---|---|
| Developmental morphology | Organisms store body plans as distributed constraint networks (bioelectric, biomechanical). Altering these constraints reproducibly redirects growth to coherent alternative forms (as in Levin's planarian and Xenopus work). | Perturbations to patterning signals produce only random deformities or non-viable tissue, never coherent alternative anatomies |
| Collective cognition | When individuals are richly connected by feedback constraints, higher-level "intelligence" emerges. Ant colonies, bee swarms, and human groups with strong coordination outperform isolated individuals on complex tasks. | Integrated groups never outperform individuals; coordination always degrades problem-solving |
| Convergent evolution | Independent lineages under similar constraints evolve analogous cognitive features. If consciousness arises from constraint satisfaction rather than special essence, similar constraints should produce similar solutions. | Hard phylogenetic cutoff where complex cognition appears without precursors; no convergent cognitive features in distant taxa facing similar constraints |
| Substrate independence | Any system achieving recursive self-modeling and integrative constraint closure will exhibit empirical signs of consciousness (novel self-reports, unpredictable self-initiated goals). | A system with human-level complexity in self-monitoring and integration shows no evidence of phenomenology and only recombines training inputs |
The framework predicts that interventions at one scale will propagate through organizational closure to other scales in lawful, predictable ways. This is a strong test of the closure hypothesis.
- - - -Molecular → Behavioral → Phenomenal: Altering ion channel kinetics, receptor densities, or neurotransmitter levels will systematically alter global brain dynamics and conscious states. This is confirmed extensively by pharmacology: anesthetics abolish consciousness by disrupting integrative dynamics; psychedelics alter phenomenal content by modifying serotonergic constraint satisfaction; anxiolytics and antidepressants shift affective constraints by changing molecular parameters. The lawfulness of these mappings (specific molecular targets → specific phenomenal changes) supports the claim that closure propagates constraints across scales.
- - - -Behavioral → Molecular: Sustained cognitive training, psychological intervention, or traumatic experience will produce measurable changes in gene expression, synaptic density, and connectivity patterns. This is confirmed by neuroplasticity research: meditation produces measurable changes in cortical thickness and connectivity; skill acquisition reshapes motor cortex; trauma alters stress-response circuitry and epigenetic markers. The existence of these top-down effects (behavioral → molecular) demonstrates that constraint propagation is bidirectional.
- - - -Falsification condition for cross-scale propagation: If molecular interventions produced only local effects with no systematic propagation to behavior or experience (e.g., altering neurotransmitter levels with no behavioral or experiential consequence), or if psychological interventions produced no measurable neural correlates (e.g., years of meditation with zero change in any brain measure), the closure assumption would fail. So far, no such dissociations have been found; instead, every level of intervention produces effects that propagate through the system.
- - - -Landauer, R. (1961). Irreversibility and heat generation in the computing process. IBM Journal of Research and Development, 5(3), 183–191. https://doi.org/10.1147/rd.53.0183
- - - -Bennett, C. H. (1982). The thermodynamics of computation—a review. International Journal of Theoretical Physics, 21(12), 905–940. https://doi.org/10.1007/BF02084158
- - - -Bérut, A., Arakelyan, A., Petrosyan, A., Ciliberto, S., Dillenschneider, R., & Lutz, E. (2012). Experimental verification of Landauer's principle linking information and thermodynamics. Nature, 483(7388), 187–189. https://doi.org/10.1038/nature10872
- - - -Jarzynski, C. (1997). Nonequilibrium equality for free energy differences. Physical Review Letters, 78(14), 2690. https://doi.org/10.1103/PhysRevLett.78.2690
- - - -Crooks, G. E. (1999). Entropy production fluctuation theorem and the nonequilibrium work relation for free energy differences. Physical Review E, 60(3), 2721. https://doi.org/10.1103/PhysRevE.60.2721
- - - -Seifert, U. (2012). Stochastic thermodynamics, fluctuation theorems and molecular machines. Reports on Progress in Physics, 75(12), 126001. https://doi.org/10.1088/0034-4885/75/12/126001
- - - -Price, H. (1996). Time's Arrow and Archimedes' Point: New Directions for the Physics of Time. Oxford University Press.
- - - -Rovelli, C. (2018). The Order of Time. Riverhead Books.
- - - -Montévil, M., & Mossio, M. (2015). Biological organisation as closure of constraints. Journal of Theoretical Biology, 372, 179–191. https://doi.org/10.1016/j.jtbi.2015.02.029
- - - -Moreno, A., & Mossio, M. (2015). Biological Autonomy: A Philosophical and Theoretical Enquiry. Springer.
- - - -Hordijk, W., Steel, M., & Kauffman, S. (2012). The structure of autocatalytic sets: Evolvability, enablement, and emergence. Acta Biotheoretica, 60(4), 379–392. https://doi.org/10.1007/s10441-012-9165-1
- - - -Kauffman, S. A. (2019). A World Beyond Physics: The Emergence and Evolution of Life. Oxford University Press.
- - - -Rosen, R. (1991). Life Itself: A Comprehensive Inquiry into the Nature, Origin, and Fabrication of Life. Columbia University Press.
- - - -Deacon, T. W. (2011). Incomplete Nature: How Mind Emerged from Matter. W. W. Norton & Company.
- - - -Deacon, T. W. (2021). How molecules became signs. Biosemiotics, 14, 537–559. https://doi.org/10.1007/s12304-021-09453-9
- - - -Deacon, T. W. (2023). Minimal properties of a natural semiotic system: Response to commentaries on "How Molecules Became Signs." Biosemiotics, 16, 1–13. https://doi.org/10.1007/s12304-023-09527-w
- - - -Sherman, J. (2017). Neither Ghost nor Machine: The Emergence and Nature of Selves. Columbia University Press.
- - - -Sherman, J. (2024). Biosemiotics' greatest potential contribution to biology. Chinese Semiotic Studies, 20(2), 231–253. https://doi.org/10.1515/css-2024-2013
- - - -Pattee, H. H. (2001). The physics of symbols: Bridging the epistemic cut. Biosystems, 60(1-3), 5–21. https://doi.org/10.1016/S0303-2647(01)00104-6
- - - -England, J. L. (2015). Dissipative adaptation in driven self-assembly. Nature Nanotechnology, 10(11), 919–923. https://doi.org/10.1038/nnano.2015.250
- - - -Hoel, E. P., Albantakis, L., & Tononi, G. (2013). Quantifying causal emergence shows that macro can beat micro. Proceedings of the National Academy of Sciences, 110(49), 19790–19795. https://doi.org/10.1073/pnas.1314922110
- - - -Ellis, G. F. R. (2012). Top-down causation and emergence: Some comments on mechanisms. Interface Focus, 2(1), 126–140. https://doi.org/10.1098/rsfs.2011.0062
- - - -Parr, T., Pezzulo, G., & Friston, K. J. (2022). Active Inference: The Free Energy Principle in Mind, Brain, and Behavior. MIT Press.
- - - -Kirchhoff, M., Parr, T., Palacios, E., Friston, K., & Kiverstein, J. (2018). The Markov blankets of life: Autonomy, active inference and the free energy principle. Journal of The Royal Society Interface, 15(138), 20170792. https://doi.org/10.1098/rsif.2017.0792
- - - -Hipólito, I., Ramstead, M., Convertino, L., Bhat, A., Friston, K., & Parr, T. (2021). Markov blankets in the brain. Neuroscience & Biobehavioral Reviews, 125, 88–97. https://doi.org/10.1016/j.neubiorev.2021.02.003
- - - -Ramstead, M. J. D., Kirchhoff, M. D., & Friston, K. J. (2020). A tale of two densities: Active inference is enactive inference. Adaptive Behavior, 28(4), 225–239. https://doi.org/10.1177/1059712319862774
- - - -Pezzulo, G., Parr, T., & Friston, K. (2024). Active inference as a theory of sentient behavior. Biological Psychology, 186, 108741. https://doi.org/10.1016/j.biopsycho.2023.108741
- - - -Ladyman, J., Ross, D., Spurrett, D., & Collier, J. (2007). Every Thing Must Go: Metaphysics Naturalized. Oxford University Press.
- - - -French, S. (2014). The Structure of the World: Metaphysics and Representation. Oxford University Press.
- - - -Rovelli, C. (1996). Relational quantum mechanics. International Journal of Theoretical Physics, 35(8), 1637–1678. https://doi.org/10.1007/BF02302261
- - - -Price, H. (1996). Time's Arrow and Archimedes' Point: New Directions for the Physics of Time. Oxford University Press.
- - - -Rovelli, C. (2018). The Order of Time. Riverhead Books.
- - - -Dehaene, S., & Changeux, J. P. (2011). Experimental and theoretical approaches to conscious processing. Neuron, 70(2), 200–227. https://doi.org/10.1016/j.neuron.2011.03.018
- - - -Dehaene-Lambertz, G., Dehaene, S., & Hertz-Pannier, L. (2002). Functional neuroimaging of speech perception in infants. Science, 298(5600), 2013–2015. https://doi.org/10.1126/science.1077066
- - - -Tononi, G., & Koch, C. (2015). Consciousness: Here, there and everywhere? Philosophical Transactions of the Royal Society B, 370(1668), 20140167. https://doi.org/10.1098/rstb.2014.0167
- - - -Massimini, M., Ferrarelli, F., Huber, R., Esser, S. K., Singh, H., & Tononi, G. (2005). Breakdown of cortical effective connectivity during sleep. Science, 309(5744), 2228–2232. https://doi.org/10.1126/science.1117256
- - - -Casali, A. G., Gosseries, O., Rosanova, M., et al. (2013). A theoretically based index of consciousness independent of sensory processing and behavior. Science Translational Medicine, 5(198), 198ra105. https://doi.org/10.1126/scitranslmed.3006294
- - - -Mashour, G. A., Roelfsema, P., Changeux, J. P., & Dehaene, S. (2020). Conscious processing and the global neuronal workspace hypothesis. Neuron, 105(5), 776–798. https://doi.org/10.1016/j.neuron.2020.01.026
- - - -Owen, A. M., Coleman, M. R., Boly, M., Davis, M. H., Laureys, S., & Pickard, J. D. (2006). Detecting awareness in the vegetative state. Science, 313(5792), 1402. https://doi.org/10.1126/science.1130197
- - - -Giacino, J. T., Kalmar, K., & Whyte, J. (2004). The JFK Coma Recovery Scale-Revised: Measurement characteristics and diagnostic utility. Archives of Physical Medicine and Rehabilitation, 85(12), 2020–2029. https://doi.org/10.1016/j.apmr.2004.02.033
- - - -Andrews, K., Murphy, L., Munday, R., & Littlewood, C. (1996). Misdiagnosis of the vegetative state: Retrospective study in a rehabilitation unit. BMJ, 313(7048), 13–16. https://doi.org/10.1136/bmj.313.7048.13
- - - -Schnakers, C., Vanhaudenhuyse, A., Giacino, J., et al. (2009). Diagnostic accuracy of the vegetative and minimally conscious state: Clinical consensus versus standardized neurobehavioral assessment. BMC Neurology, 9, 35. https://doi.org/10.1186/1471-2377-9-35
- - - -Sebel, P. S., Bowdle, T. A., Ghoneim, M. M., et al. (2004). The incidence of awareness during anesthesia: A multicenter United States study. Anesthesia & Analgesia, 99(3), 833–839. https://doi.org/10.1213/01.ANE.0000130261.90896.6C
- - - -Alkire, M. T., Hudetz, A. G., & Tononi, G. (2008). Consciousness and anesthesia. Science, 322(5903), 876–880. https://doi.org/10.1126/science.1149213
- - - -Friston, K. (2010). The free-energy principle: A unified brain theory? Nature Reviews Neuroscience, 11(2), 127–138. https://doi.org/10.1038/nrn2787
- - - -Baars, B. J. (1988). A Cognitive Theory of Consciousness. Cambridge University Press.
- - - -Dehaene, S., Naccache, L., Cohen, L., Le Bihan, D., Mangin, J. F., Poline, J. B., & Rivière, D. (2001). Cerebral mechanisms of word masking and unconscious repetition priming. Nature Neuroscience, 4(7), 752–758. https://doi.org/10.1038/89551
- - - -Rounis, E., Maniscalco, B., Rothwell, J. C., Passingham, R. E., & Lau, H. (2010). Theta-burst transcranial magnetic stimulation to the prefrontal cortex impairs metacognitive visual awareness. Cognitive Neuroscience, 1(3), 165–175. https://doi.org/10.1080/17588921003632529
- - - -Lau, H., & Rosenthal, D. (2011). Empirical support for higher-order theories of conscious awareness. Trends in Cognitive Sciences, 15(8), 365–373. https://doi.org/10.1016/j.tics.2011.05.009
- - - -Block, N. (2011). Perceptual consciousness overflows cognitive access. Trends in Cognitive Sciences, 15(12), 567–575. https://doi.org/10.1016/j.tics.2011.11.001
- - - -Rosenthal, D. M. (2005). Consciousness and Mind. Oxford University Press.
- - - -Varela, F. J., Thompson, E., & Rosch, E. (1991). The Embodied Mind: Cognitive Science and Human Experience. MIT Press.
- - - -Libet, B., Gleason, C. A., Wright, E. W., & Pearl, D. K. (1983). Time of conscious intention to act in relation to onset of cerebral activity (readiness-potential): The unconscious initiation of a freely voluntary act. Brain, 106(3), 623–642. https://doi.org/10.1093/brain/106.3.623
- - - -Libet, B. (1985). Unconscious cerebral initiative and the role of conscious will in voluntary action. Behavioral and Brain Sciences, 8(4), 529–539. https://doi.org/10.1017/S0140525X00044903
- - - -Chalmers, D. J. (1995). Facing up to the problem of consciousness. Journal of Consciousness Studies, 2(3), 200–219.
- - - -Chalmers, D. J. (1996). The Conscious Mind: In Search of a Fundamental Theory. Oxford University Press.
- - - -Chalmers, D. J. (2016). The combination problem for panpsychism. In G. Brüntrup & L. Jaskolla (Eds.), Panpsychism: Contemporary Perspectives (pp. 179–214). Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199359943.003.0002
- - - -Coleman, S. (2014). The real combination problem: Panpsychism, micro-subjects, and emergence. Erkenntnis, 79(1), 19–44. https://doi.org/10.1007/s10670-013-9431-x
- - - -Goff, P. (2017). Consciousness and Fundamental Reality. Oxford University Press.
- - - -Strawson, G. (2006). Realistic monism: Why physicalism entails panpsychism. Journal of Consciousness Studies, 13(10-11), 3–31.
- - - -Dennett, D. C. (1988). Quining qualia. In A. J. Marcel & E. Bisiach (Eds.), Consciousness in Contemporary Science (pp. 42–77). Oxford University Press.
- - - -Dennett, D. C. (1991). Consciousness Explained. Little, Brown and Company.
- - - -Frankish, K. (2016). Illusionism as a theory of consciousness. Journal of Consciousness Studies, 23(11-12), 11–39.
- - - -Jackson, F. (1982). Epiphenomenal qualia. Philosophical Quarterly, 32(127), 127–136. https://doi.org/10.2307/2960077
- - - -Levine, J. (1983). Materialism and qualia: The explanatory gap. Pacific Philosophical Quarterly, 64(4), 354–361. https://doi.org/10.1111/j.1468-0114.1983.tb00207.x
- - - -Nagel, T. (1974). What is it like to be a bat? Philosophical Review, 83(4), 435–450. https://doi.org/10.2307/2183914
- - - -Popper, K. R. (1959). The Logic of Scientific Discovery. Hutchinson.
- - - -Lakatos, I. (1978). The Methodology of Scientific Research Programmes. Cambridge University Press.
- - - -Flew, A. (1950). Theology and falsification. University, 1, 1–8.
- - - -Whitehead, A. N. (1925). Science and the Modern World. Macmillan.
- - - -Yunkaporta, T. (2020). Sand Talk: How Indigenous Thinking Can Save the World. HarperOne. (Originally published 2019 by Text Publishing, Australia.)
- - - -Kelly, L. (2017). The Memory Code: The Secrets of Stonehenge, Easter Island and Other Ancient Monuments. Pegasus Books. (Originally published 2016 by Allen & Unwin, Australia.)
- - - -Kimmerer, R. W. (2013). Braiding Sweetgrass: Indigenous Wisdom, Scientific Knowledge and the Teachings of Plants. Milkweed Editions.
- - - -Reser, D., Simmons, M., Johns, E., Ghaly, A., Quayle, M., Dordevic, A. L., Tare, M., McArdle, A., Willems, J., & Yunkaporta, T. (2021). Australian Aboriginal techniques for memorization: Translation into a medical and allied health education setting. PLOS ONE, 16(5), e0251710. https://doi.org/10.1371/journal.pone.0251710
- - - -Deloria, V., Jr. (2003). God Is Red: A Native View of Religion (30th Anniversary ed.). Fulcrum Publishing.
- - - -Deloria, V., Jr., & Wildcat, D. R. (2001). Power and Place: Indian Education in America. Fulcrum Publishing.
- - - -Cordova, V. F. (2007). How It Is: The Native American Philosophy of V. F. Cordova (K. D. Moore, K. Peters, T. Jojola, & A. Lacy, Eds.). University of Arizona Press.
- - - -Norton-Smith, T. M. (2010). The Dance of Person and Place: One Interpretation of American Indian Philosophy. SUNY Press.
- - - -Waters, A. (Ed.). (2004). American Indian Thought: Philosophical Essays. Blackwell Publishing.
- - - -Wilson, S. (2008). Research Is Ceremony: Indigenous Research Methods. Fernwood Publishing.
- - - -Wildcat, M., & Voth, D. (2023). Indigenous relationality: Definitions and methods. AlterNative: An International Journal of Indigenous Peoples, 19(2), 432–440. https://doi.org/10.1177/11771801231168380
- - - -Pratt, S. L. (2002). Native pragmatism: Rethinking the roots of American philosophy. Indiana University Press.
- - - -Durant, F., Morokuma, J., Fields, C., Williams, K., Adams, D. S., & Levin, M. (2017). Long-term, stochastic editing of regenerative anatomy via targeting endogenous bioelectric gradients. Biophysical Journal, 112(10), 2231–2243. https://doi.org/10.1016/j.bpj.2017.04.011
- - - -Durant, F., Bischof, J., Fields, C., Morokuma, J., LaPalme, J., Hoi, A., & Levin, M. (2019). The role of early bioelectric signals in the regeneration of planarian anterior/posterior polarity. Biophysical Journal, 116(5), 948–961. https://doi.org/10.1016/j.bpj.2019.01.029
+ +Durant, F., Morokuma, J., Fields, C., Williams, K., Adams, D. S., & Levin, M. (2017). Long-term, stochastic editing of regenerative anatomy via targeting endogenous bioelectric gradients. Biophysical Journal, 112(10), 2231–2243. https://doi.org/10.1016/j.bpj.2017.04.011
+ + + +Durant, F., Bischof, J., Fields, C., Morokuma, J., LaPalme, J., Hoi, A., & Levin, M. (2019). The role of early bioelectric signals in the regeneration of planarian anterior/posterior polarity. Biophysical Journal, 116(5), 948–961. https://doi.org/10.1016/j.bpj.2019.01.029
+ + + +Chis-Ciure, R., & Levin, M. (2025). Cognition all the way down 2.0: Neuroscience beyond neurons in the diverse intelligence era. Synthese, 206, article 257. https://doi.org/10.1007/s11229-025-05319-6.
@@ -5398,19 +3067,26 @@ -Hoffman, D. D., Singh, M., & Prakash, C. (2015). The interface theory of perception. Psychonomic Bulletin & Review, 22(6), 1480–1506. https://doi.org/10.3758/s13423-015-0890-8
- - - -Hoffman, D. D., & Prakash, C. (2014). Objects of consciousness. Frontiers in Psychology, 5, 577. https://doi.org/10.3389/fpsyg.2014.00577
- - - -Bagwell, J. N. (2023). Debunking interface theory: Why Hoffman's skepticism (really) is self-defeating. Synthese, 201(25). https://doi.org/10.1007/s11229-022-04021-1
- - - -Martínez, M. (2019). Usefulness drives representations to truth: A family of counterexamples to Hoffman's interface theory of perception. Grazer Philosophische Studien, 96(3), 319–341. https://doi.org/10.1163/18756735-09603004
+Hoffman, D. D., Singh, M., & Prakash, C. (2015). The interface theory of perception. Psychonomic Bulletin & Review, 22(6), 1480–1506. https://doi.org/10.3758/s13423-015-0890-8
+ + + +Hoffman, D. D., & Prakash, C. (2014). Objects of consciousness. Frontiers in Psychology, 5, 577. https://doi.org/10.3389/fpsyg.2014.00577
+ + + +Bagwell, J. N. (2023). Debunking interface theory: Why Hoffman's skepticism (really) is self-defeating. Synthese, 201(25). https://doi.org/10.1007/s11229-022-04021-1
+ + + +Martínez, M. (2019). Usefulness drives representations to truth: A family of counterexamples to Hoffman's interface theory of perception. Grazer Philosophische Studien, 96(3), 319–341. https://doi.org/10.1163/18756735-09603004
+ + + + + +Prakash, C., Stephens, K. D., Hoffman, D. D., Singh, M., & Fields, C. (2021). Fitness Beats Truth in the Evolution of Perception. Acta Biotheoretica, 69(3), 319–341. DOI: 10.1007/s10441-020-09400-0.
+ @@ -5418,7 +3094,7 @@ -Cerullo, M. A. (2015). The problem with phi: A critique of integrated information theory. PLOS Computational Biology, 11(9), e1004286. https://doi.org/10.1371/journal.pcbi.1004286
+Cerullo, M. A. (2015). The problem with phi: A critique of integrated information theory. PLOS Computational Biology, 11(9), e1004286. https://doi.org/10.1371/journal.pcbi.1004286
@@ -5426,7 +3102,14 @@ -Doerig, A., Schurger, A., Hess, K., & Herzog, M. H. (2019). The unfolding argument: Why IIT and other causal structure theories cannot explain consciousness. Consciousness and Cognition, 72, 49–59. https://doi.org/10.1016/j.concog.2019.04.002
+Doerig, A., Schurger, A., Hess, K., & Herzog, M. H. (2019). The unfolding argument: Why IIT and other causal structure theories cannot explain consciousness. Consciousness and Cognition, 72, 49–59. https://doi.org/10.1016/j.concog.2019.04.002
+ + + + + +Tsuchiya, N., Andrillon, T., & Haun, A. (2020). A reply to “the unfolding argument”: Beyond functionalism/behaviorism and towards a science of causal structure theories of consciousness. Consciousness and Cognition, 79, 102877. DOI: 10.1016/j.concog.2020.102877.
+ @@ -5434,19 +3117,19 @@ -Bruineberg, J., Kiverstein, J., & Rietveld, E. (2018). The anticipating brain is not a scientist: The free-energy principle from an ecological-enactive perspective. Synthese, 195(6), 2417–2444. https://doi.org/10.1007/s11229-016-1239-1
- - - -Kirchhoff, M., Parr, T., Palacios, E., Friston, K., & Kiverstein, J. (2018). The Markov blankets of life: Autonomy, active inference and the free energy principle. Journal of The Royal Society Interface, 15(138), 20170792. https://doi.org/10.1098/rsif.2017.0792
- - - -Bruineberg, J., Dolega, K., Dewhurst, J., & Baltieri, M. (2022). The emperor's new Markov blankets. Behavioral and Brain Sciences, 45, e183. https://doi.org/10.1017/S0140525X21002351
- - - -Friston, K. (2022). Maps and territories, smoke, and mirrors: A reply to commentary on "The emperor's new Markov blankets." Behavioral and Brain Sciences, 45, e183. https://doi.org/10.1017/S0140525X22000632
+Bruineberg, J., Kiverstein, J., & Rietveld, E. (2018). The anticipating brain is not a scientist: The free-energy principle from an ecological-enactive perspective. Synthese, 195(6), 2417–2444. https://doi.org/10.1007/s11229-016-1239-1
+ + + +Kirchhoff, M., Parr, T., Palacios, E., Friston, K., & Kiverstein, J. (2018). The Markov blankets of life: Autonomy, active inference and the free energy principle. Journal of The Royal Society Interface, 15(138), 20170792. https://doi.org/10.1098/rsif.2017.0792
+ + + +Bruineberg, J., Dolega, K., Dewhurst, J., & Baltieri, M. (2022). The emperor's new Markov blankets. Behavioral and Brain Sciences, 45, e183. https://doi.org/10.1017/S0140525X21002351
+ + + +Friston, K. (2022). Maps and territories, smoke, and mirrors. Behavioral and Brain Sciences, 45, e195. https://doi.org/10.1017/S0140525X22000073
@@ -5454,7 +3137,7 @@ -Tegmark, M. (2000). Importance of quantum decoherence in brain processes. Physical Review E, 61(4), 4194. https://doi.org/10.1103/PhysRevE.61.4194
+Tegmark, M. (2000). Importance of quantum decoherence in brain processes. Physical Review E, 61(4), 4194. https://doi.org/10.1103/PhysRevE.61.4194
@@ -5481,142 +3164,14 @@Olivelle, P. (Trans.). (1996). Upaniṣads. Oxford University Press.
- -Kleiner, J., & Hoel, E. (2021). Falsification and consciousness. Neuroscience of Consciousness 2021(1): niab001. DOI: 10.1093/nc/niab001 Hornik, K., Stinchcombe, M., & White, H. (1989). Multilayer feedforward networks are universal approximators. Neural Networks 2: 359-366. DOI: 10.1016/0893-6080(89)90020-8 Turing, A. M. (1937). On computable numbers, with an application to the Entscheidungsproblem. Proceedings of the London Mathematical Society 2(42): 230-265. DOI: 10.1112/plms/s2-42.1.230 Koch, C. (2019). The Feeling of Life Itself: Why Consciousness Is Widespread but Can't Be Computed. MIT Press. Doerig, A., Schurger, A., Hess, K., & Herzog, M. H. (2019). The unfolding argument: Why IIT and other causal structure theories cannot explain consciousness. Consciousness and Cognition 72: 49-59. DOI: 10.1016/j.concog.2019.04.002 Albantakis, L., & Tononi, G. (2019). Causal composition: Structural differences among dynamically equivalent systems. Entropy 21(10): 989. DOI: 10.3390/e21100989 Hutter, M. (2004). Universal Artificial Intelligence: Sequential Decisions Based on Algorithmic Probability. Springer. DOI: 10.1007/b138233 Wolfram, S. (1984). Cellular automata as models of complexity. Nature 311: 419-424. DOI: 10.1038/311419a0 Tsuchiya, N., Andrillon, T., & Haun, A. (2019). A reply to "the unfolding argument": Beyond functionalism/behaviorism and towards a truer science of causal structural theories of consciousness. PsyArXiv. DOI: 10.31234/osf.io/8yqvm Casarotto, S., et al. (2024). Perturbational complexity index confirms diagnostic threshold across disorders of consciousness. Annals of Neurology 96(5): 976-988. DOI: 10.1002/ana.27053 Tsuchiya, N., et al. (2021). Enriched category theory as a framework for consciousness science. Neuroscience Research 175: 63-71. DOI: 10.1016/j.neures.2021.12.001
- - - -Every framework in consciousness studies rests on assumptions, many implicit. Making these explicit reveals which are empirical (testable), which are methodological (useful but not claims about the world), and which are metaphysical commitments that cannot be tested. A framework's hidden assumptions are often its most vulnerable points.
- - - -| Framework | Core Assumption | Type | Falsifiable? | Hidden Commitment |
|---|---|---|---|---|
| Constraint-based (this paper) | Consciousness = organizational closure with self-modeling | Structural identification | Yes (5 loss conditions) | Assumes closure is sufficient, not merely necessary |
| IIT | Consciousness = integrated information (Φ) | Identity claim | Partially | Assumes intrinsic existence from causal structure; panpsychist implications at edges |
| GWT | Consciousness = global broadcast/availability | Functional claim | Yes | Equates access with phenomenology; may miss unreportable experience |
| Panpsychism | Proto-experience pervades matter | Metaphysical primitive | No | Assumes combination is possible without mechanism |
| HOT | Consciousness requires meta-representation | Architectural claim | Partially | Assumes unconscious HOT can confer consciousness on target; regress problem |
| FEP | Systems minimize free energy | Mathematical identity | No | Too broad to fail; any behavior redescribable as free energy minimization |
| Orch-OR | Consciousness requires quantum coherence | Physical claim | Yes (timescale) | Assumes coherence survives decoherence; conflicts with thermal physics |
| Process Philosophy | Experience is fundamental | Metaphysical primitive | No | Cannot specify what would show experience isn't fundamental |
| Interface Theory (Hoffman) | Consciousness is fundamental; spacetime is an interface | Idealist ontology | No | Assumes evolutionary selection presupposes consciousness; parasitic on realist assumptions it claims to replace |
| Framework | How We Know Consciousness | Hidden Commitment |
|---|---|---|
| Constraint-based | Measurable signatures: integration, perturbational complexity, graded degradation | Assumes organizational features are accessible to third-person measurement |
| IIT | Compute Φ from causal structure | Assumes Φ calculation is tractable; in practice relies on proxies |
| GWT | Reportability and behavioral indicators | Assumes report tracks phenomenology; may miss "overflow" |
| HOT | Presence of higher-order representation | Assumes metacognitive capacity indicates consciousness; problematic for infants/animals |
| Panpsychism | Inference from own case to all matter | Assumes analogy is valid despite no shared features with electrons |
| FEP | Model parameters and prediction errors | Assumes Bayesian quantities map onto subjective quantities |
| Interface Theory (Hoffman) | Mathematical derivation from evolutionary games | Assumes mathematical structure directly generates phenomenology; confuses description with causation |
| Framework | Research Strategy | Hidden Commitment |
|---|---|---|
| Constraint-based | Identify falsification conditions; test organizational markers | Assumes falsification-first methodology is appropriate for consciousness |
| IIT | Axiomatic derivation from phenomenology | Assumes phenomenological axioms are self-evident and complete |
| GWT | Contrastive analysis (conscious vs. unconscious processing) | Assumes the contrast is exhaustive; no third category |
| FEP | Computational modeling under variational principles | Assumes brain literally implements Bayesian inference |
| Panpsychism | Conceptual analysis and inference to best explanation | No experimental methodology; relies on perceived elegance |
| Interface Theory (Hoffman) | Evolutionary game theory and mathematical modeling | No falsification conditions; any observation reinterpretable as "interface" |
Frameworks can be compared by how many assumptions they require and how testable those assumptions are:
- - - -| Framework | Ontological Assumptions | Testable | Metaphysical Load |
|---|---|---|---|
| Constraint-based | 3 (closure, self-modeling, thermodynamic cost) | All 3 | Low |
| GWT | 2 (global broadcast, capacity limits) | Both | Low |
| IIT | 5 (axioms) + identity claim | Partially | Moderate |
| HOT | 2 (HOT necessity, can be unconscious) | Partially | Moderate |
| FEP | 1 (but unfalsifiable) | No | Low but empty |
| Panpsychism | 2 (ubiquity, combination possible) | Neither | High |
| Orch-OR | 3 (coherence, objective reduction, microtubules) | All 3 but 2 falsified | High |
| Process Philosophy | 1 (experience fundamental) | No | High |
The constraint-based framework makes explicit, falsifiable assumptions with minimal metaphysical overhead. Competing frameworks either make unfalsifiable assumptions (panpsychism, process philosophy, FEP), make assumptions that have been falsified (Orch-OR), or make assumptions that are partially testable but carry hidden commitments (IIT, HOT).
- - - -For transparency, here are the explicit assumptions of the constraint-based framework:
- - - -What this framework does NOT assume:
- - - -The framework makes a structural identification (consciousness = organizational closure with self-modeling) rather than an identity claim (consciousness = X) or a primitivist claim (consciousness is fundamental). This identification is testable: find systems with closure and self-modeling that lack consciousness signatures, or find consciousness signatures without closure, and the framework fails.
- - - -The framework this essay describes is not complete. It is not final. It is falsifiable. If it survives, that will be evidence that it tracks something real. If it fails, that will be progress. Either way, we will have learned something, which is more than unfalsifiable frameworks can offer.
- - - -The claim that phenomenology must be "more than" any physical or functional description is not just unfalsifiable. It is anti-scientific in structure, indistinguishable from other immunized metaphysical claims once stripped of anthropocentric sentiment. This framework does not need to defeat that claim on its own terms. It shows that the claim never earned standing in the first place.
- - - -Survival under falsification pressure is evidence of tracking structure, not proof of final truth. The framework does not claim certainty. It claims testability. That is all any honest inquiry can offer.
-+ + + +Turing, A. M. (1937). On computable numbers, with an application to the Entscheidungsproblem. Proceedings of the London Mathematical Society, 2(42), 230–265. DOI: 10.1112/plms/s2-42.1.230.
+ + + # ADDENDUM: final internal cross-reference, prior specific-repo blob 54bbefbe to synchronized current reading edition --- article-specific-before-last-synchronization.html +++ canonical-article-current.html @@ -841,7 +841,7 @@ -A stronger interpretation proposes that biological structures causally access an independently existing Platonic morphospace. When that is offered as an explanation for a physical result, it needs a bridge. What property of the alleged ingression interface is measured? Which intervention or prediction distinguishes access to a pre-existing nonphysical form from a well-specified dynamical, developmental, or organizational account? The Chladni Plate Challenge to Michael Levin’s Biological Platonism develops a related dynamical alternative.
+A stronger interpretation proposes that biological structures causally access an independently existing Platonic morphospace. When that is offered as an explanation for a physical result, it needs a bridge. What property of the alleged ingression interface is measured? Which intervention or prediction distinguishes access to a pre-existing nonphysical form from a well-specified dynamical, developmental, or organizational account? The Chladni Plate Challenge to Michael Levin’s Biological Platonism develops a related dynamical alternative (see the Michael Levin evidence and scholarly criticism index for the complete research collection).