Thermodynamic Realism: The Consolidated Deductive Presentation (rev. 4)

Thermodynamic Realism: The Consolidated Deductive Presentation (rev. 4)

Thermodynamic Realism: The Consolidated Deductive Presentation

Writings: CC BY 4.0. This document supersedes and consolidates the rev. 3 apex paper, the is-ought dissolution paper, the glossary, the deductive-web map, and the diagnostic procedure into a single statement. Each post is open to revision.

Abstract. We present Thermodynamic Realism (TR) as a unified, falsifiable research program grounded in four axioms (the persistence tautology, physicalism, the Second Law, and the physicality of information) and four background empirical premises (the low-entropy initial condition, non-equilibrium existence, finite resources, and multi-scale evolutionary organization). From this base we trace a layered architecture of numbered results, each labelled by kind so that its epistemic status is visible: theorems are entailments, definitions fix terms, characterisations propose what a familiar word picks out, functional accounts explain why a feature exists without claiming phenomenological reduction, and analogical extensions reach beyond strict entailment. The architecture dissolves the is-ought problem by treating "ought" as a domain-bound operator, naturalises ethics as the modeling activity of bounded agents, grounds a single epistemic imperative, gives a functional account of valence and empathy, and offers a thermodynamic hypothesis about censorship-driven collapse. New in this revision: the framework's central cost claim is given real physical grounding in the thermodynamics of prediction (Still, Sivak, Bell & Crooks 2012), where dissipation is shown equivalent to a model's nonpredictive retained information; the persistence-selection principle is stated in Price-equation form; the deception cost is given a description-length bound; and the ought-relation is given an ordinal generalisation. Claims that cannot yet be formalised are marked open rather than asserted. The is-ought dissolution rests on a single semantic premise, stated and defended rather than assumed, and identified as the framework's principal philosophical exposure. The framework's domain is bounded at both ends, by the Past Hypothesis and by the dark-energy approach to equilibrium. It is offered not as a finished metascience but as a transparent, testable, cross-disciplinary program with its open questions named.

0. Reader's Guide: How to Read the Labels

Every numbered result carries a label recording what kind of move it is. This is the single most important interpretive instruction in the paper: no result should be read as carrying more weight than its kind permits.

Axiom Background premise Theorem Definition Characterisation Functional account Analogical extension Corollary Falsification

An axiom is treated as bedrock and not derived. A background premise is an empirical fact about our universe the derivations rely on but cannot produce from the axioms alone. A theorem is an entailment traced to its parents. A definition introduces or fixes a term. A characterisation proposes what a familiar word picks out within the framework. A functional account explains why a feature exists and what it does, without claiming to reduce its felt character. An analogical extension applies the architecture to a domain by analogy rather than by strict entailment. A corollary is a descriptive consequence. Falsification criteria state the conditions under which the framework would be refuted; they are not theorems.

Contents

1. The Problem of Fragmentation

Human knowledge is partitioned into disciplines that lack a common axiomatic foundation. Physics describes the territory but says nothing about value. Information theory describes the cost of representation but says nothing about why accuracy matters. Evolutionary biology describes selection among replicators but says little about the fate of non-replicating persistent structures. Ethics asks how agents ought to behave but struggles to ground "ought" in "is." The is-ought problem has persisted for three centuries, moral realism remains contested, and civilizational collapse is studied without a unified thermodynamic frame.

This paper proposes that a single set of physical premises, stated explicitly and followed wherever they lead, yields a structure in which the is-ought gap closes, the functional role of valence finds a physical grounding, ought-facts become physically determinate though often computationally inaccessible, and the collapse of information-suppressing regimes becomes a thermodynamic hypothesis. The framework is called Thermodynamic Realism.

What this document is. It is a deductive presentation of a research program. It traces a layered architecture of numbered results from axioms and background premises, gives real formalization where the physics supports it and marks the rest open, specifies falsification protocols, integrates the framework's diagnostic procedure and glossary, defends the one premise the meta-ethics rests on, and answers objections at length. It does not claim to be a finished edifice. It claims to be transparent, testable, and honest about its own status.

A note on the cosmological backdrop. The picture of the universe as a single thermodynamic computation, of spacetime as emergent, and of the arrow of time as the entropy gradient, is referenced background, not part of the axiom base. The deductive web does not require a theory of emergent spacetime. It requires only a free-energy gradient, which enters as a background premise. Keeping the axiom base minimal is deliberate: the load-bearing claims should depend on as little contested physics as possible.

2. Axioms and Background Premises

We adopt four axioms and four background empirical premises. The axioms are the logical and physical bedrock. The background premises are empirical facts about our universe the derivations rely on but that are not derivable from the axioms alone. Making them explicit prevents the appearance of smuggling.

2.1 Axioms

A0. The Persistence Tautology Axiom

Systems that do not maintain the conditions of their own persistence cease to exist as observables. Only systems that persist remain available for observation. Formally, writing \(S(t)\) for the system existing at time \(t\): if \(S\) fails its persistence conditions over \([t,t+\Delta t]\), then \(\neg\exists\, S(t+\Delta t)\).

Justification: a tautology. It asserts nothing about value; it states that existence has prerequisites and that what fails them is no longer present to be observed.

A1. Physicalism (inductively justified) Axiom

The universe is a physical system. All phenomena, including life, mind, and culture, are physical phenomena. This is physicalism in the minimal sense, that all phenomena are physical; on its own it does not settle whether phenomenal experience is exhausted by function (see §6.4).

Justification: the maximally inductively justified working premise. Every phenomenon ever seriously investigated has yielded to physical explanation. Demanding certainty beyond this inductive record is epistemic paralysis.

A2. The Second Law of Thermodynamics Axiom

In any isolated system, entropy tends toward its maximum over time; maintaining a localized entropy gradient requires continuous work. \( dS_{\text{iso}}/dt \ge 0 \).

Justification: among the most thoroughly confirmed principles in science. Adopted without re-derivation.

A3. Information Is Physical (Landauer, Shannon) Axiom

Information storage, processing, and erasure have minimum thermodynamic costs. The Landauer bound: erasing one bit dissipates at minimum \(k_B T \ln 2\) of heat.

Justification: Landauer (1961) established the principle theoretically and it has experimental support (Bérut et al. 2012). It is also philosophically contested: Norton (2011) argues the standard derivations selectively neglect fluctuations, with replies by Ladyman and Robertson (2013) and Myrvold (2024). TR does not require the principle to be a universal theorem of mechanics. It requires the weaker and better-supported claim that the specific operations the framework relies on, the maintenance and updating of divergent models (T8, T9), are logically irreversible and therefore carry an irreducible cost; see Bennett (1982) and §6.7.

2.2 Background Empirical Premises

BP0. The Low-Entropy Initial Condition (the Past Hypothesis) Background premise

The accessible universe began in a macrostate of extraordinarily low entropy. Every free-energy gradient available to any later system is a portion of that initial condition still discharging (Albert 2000).

A clarification about gravity, the only role gravity plays in the framework. For ordinary matter without gravity, the low-entropy state is the ordered, concentrated one. For self-gravitating matter this inverts: a smooth distribution is low entropy and a clumped distribution is high entropy, because gravity makes clumping spontaneous (Penrose 1979, the Weyl curvature hypothesis). This inversion is what makes the smooth early universe a wound spring rather than a featureless equilibrium. The framework requires no theory of quantum gravity, only this single fact.

BP1. Non-Equilibrium Existence Background premise

The accessible universe is far from thermodynamic equilibrium; free-energy gradients exist and sustain localized order. BP1 follows from BP0 together with finite elapsed cosmic time; it is retained separately so that traces citing it remain stable.

BP2. Finite Accessible Resources Background premise

In any local region, the free energy accessible to a given system is finite. Combined with A2, this implies competition for negentropy among co-located systems.

BP3. Evolutionary Dynamics and Multi-Scale Organization Background premise

In environments with finite resources and variation among persisting systems, differential survival based on heritable or persistent traits produces selection effects. Persistent systems in our universe are organized into nested hierarchies of statistical boundaries (cells within organisms within ecosystems within civilizations). This multi-scale organization is an empirical fact, not a logical necessity.

The framework is thus: A0 to A3 plus BP0 to BP3 entail the results that follow. Where a result relies on a background premise, the trace records it. Traces citing BP1 are equivalently grounded in BP0.

An optional, stronger external connection, not used as a premise. Jacobson (1995) derived the Einstein field equations as a thermodynamic equation of state from the relation between heat, temperature, and entropy at local causal horizons; on that result, gravitation is the large-scale thermodynamics of spacetime. Verlinde (2011) extends this to gravity as an entropic force, which is contested. The framework cites Jacobson as convergent support for treating spacetime dynamics thermodynamically and treats Verlinde's stronger claim as open. Neither is a premise.

3. The Deductive Web

Four layers. Each result is stated, labelled by kind, and traced to its parents.

Layer 1: Immediate Consequences

T1. Environmental Variance Is Non-Zero and, for Finite Agents, Ineliminable Theorem
from A1, A2, BP1

The physical universe is a non-equilibrium system with fluctuations at all finite scales. Any embedded agent encounters a non-zero rate of environmental shift. An agent may insulate a local region at a cost, but the insulation is itself costly and eventually fails, so variance is ineliminable in the long run for any finite embedded agent.

T2. Persistence Requires Work Theorem
from A2, A0

To persist is to maintain a boundary against entropic dissolution. The Second Law says entropy increases unless work is done. Therefore persistence requires continuous work.

T3. Modeling Has a Minimum Cost Theorem
from A3

Any internal model of the environment is encoded in physical degrees of freedom. Storing, accessing, and updating it incurs non-zero thermodynamic cost.

T4. Map-Territory Divergence Has a Thermodynamic Cost Theorem
from T1, T3; formalized in §5.1 via Still et al. 2012

If the environment shifts and the agent's model does not track it, the model generates prediction errors. Each error dissipates free energy through misallocated resources and subsequent correction. The rev. 3 statement of this result was qualitative. The thermodynamics of prediction now gives it a precise form (§5.1): the work a model dissipates is bounded below by the nonpredictive information it retains, so a model that fails to track the predictively relevant structure of its environment pays in dissipation exactly to the extent of that failure.

T5. The Persistence Selection Principle Theorem (statistical)
from T2, T4, A0, BP2, BP3; formalized in §5.3

In any ecology of persisting systems competing for finite free energy, systems with lower model-territory divergence dissipate less on error correction, on average, than systems with higher divergence. Under resource limitation and differential survival, this drives a statistical tendency: over many perturbation cycles, the distribution of observed systems shifts toward those whose models track the territory more closely. Entropy performs epistemic selection.

Scope note. T5 is not a guarantee that the most accurate model always wins. It is a statistical tendency operating where variance is non-zero and resources are finite. In a perfectly stable, resource-abundant niche a distorted model can persist (the dark-room limit, §6.2). T5 covers the vast majority of real contexts but not every conceivable edge case, and it does not strengthen into a deterministic law on any horizon (§8, objection 12).

Layer 2: Structural Deductions

T6. Directional Drift Toward Lower Divergence Theorem
from T1, T5

Under expanding environmental variance and finite resources, distorted models eventually meet disconfirming perturbations. Over sufficient time and perturbation variety, the systems that persist are those whose models track the territory's causal invariants. This defines a direction, decreasing divergence, but not a fixed endpoint. The drift is directional, not convergent to a limit; no terminal "true model" is implied.

T7. Truth as Perturbationally Robust Compression Fidelity Definition
motivated by T6, T3

Within the framework, truth is defined as perturbationally robust compression fidelity: the minimal-loss compression of environmental structure sufficient for adaptive persistence across expanding perturbational horizons. This is a definitional choice, not a deduction. It is motivated by T6 (surviving models compress causal invariants) and T3 (compression minimizes cost), operationalizable via Minimum Description Length and predictive mutual information. Alternative definitions of truth exist; this one is selected for physical groundedness and measurability. Its known cost, that adaptive falsehoods can come apart from accuracy, is addressed in §8, objection 13.

T8. Intentional Deception Costs More Than Truth-Telling Theorem
from T3, T4; formalized in §5.2

A lie requires the sender to maintain at least two internal models: the accurate one and the presented one, plus the machinery to deploy the second while reasoning with the first. This imposes strictly greater storage, update, and monitoring cost than truth-telling. Deception is thermodynamically disfavored, though it can be locally advantageous when offsetting returns compensate for the overhead.

Scope note. T8 concerns intentional deception, where the sender holds an accurate model and a divergent presented one. A merely mistaken agent holds a single false model and incurs no dual-model overhead; that is the plain divergence of T4, not T8.

T9. The Consistency Tax Definition (consolidating)
consolidates T4, T8

"The Consistency Tax" names the metabolic overhead imposed by any mismatch between model and territory, whether from error (T4) or intentional deception (T8). It consolidates a cost already established rather than deriving a new one. It applies even when the agent is unaware of the mismatch (latent divergence) and spikes when the mismatch is actively corrected (active divergence).

T10. Epistemic Profit Theorem
from T9, T3

Reducing divergence frees the energy previously consumed by the Consistency Tax. This recovered surplus is Epistemic Profit, and its existence follows directly from T9 and T3.

Interpretive remark. The framework conjectures a phenomenological correlate, "Predictive Calm," the felt reduction in cognitive load when models track smoothly. This is an interpretive bridge to felt experience, not a derived result. Nothing downstream depends on it.

Layer 3: Meta-Ethics

T11. "Ought" Is a Domain-Bound Operator Theorem
from A0, T2, and the semantic premise defended in §6.1

The operator "ought" presupposes an agent with persistence conditions. Outside this domain it does not refer. "Why ought one persist at all?" is malformed in the same way as "what is north of the North Pole?" The is-ought gap is a semantic artifact of domain violation. This is not entailed by A0 and T2 alone; it rests on a semantic premise about the operator "ought," stated and defended in §6.1. T11 records the conclusion; §6.1 carries the argument and marks its limits.

T12. Within the Domain, Ought-Facts Are Physically Determinate Theorem
from A1, A2, T11; ordinal form in §5.5

For any agent with specified persistence conditions and embedding, there is a physically determinate configuration that maximizes sustained negentropy capacity over the embedding's actual horizon. This optimum may be computationally inaccessible, but inaccessibility is not indeterminacy. T12 holds conditional on T11 and inherits its dependence on the semantic premise.

T13. Ethics as the Modeling Activity of Bounded Agents Characterisation
from T12, T3, T7

Because the full optimum is intractable, agents use compressed models: moral emotions (fast heuristics), moral principles (compressed generalizations), and moral reasoning (model refinement). T13 characterises ethics as this modeling activity. It is a proposal about what "ethics" picks out within the framework, not a derived result.

T14. Moral Progress Is Real and Directional Theorem
from T6, T13

As models improve their tracking of the coupled-system thermodynamics of an embedding, they become objectively better moral models in the sense fixed by T7 and T13. Progress is directional but neither guaranteed nor complete in finite time, inheriting the directional-not-convergent character of T6. The result is conditional on the characterisation T13.

Layer 4: Full Architecture

T15. Markov Blankets and Multi-Scale Identity Theorem (imported formalism)
from T2, T5, T3, BP3

Persisting systems maintain statistical boundaries (Markov blankets) separating internal from external states. These blankets nest across scales, and selection operates at every scale. The Markov-blanket formalism is imported from the active-inference literature; the framework adopts it as vocabulary for the boundary T2 already requires, rather than deriving the formalism itself. The maximal scale, the universe, has no Markov blanket, because it has no outside.

T16. Coupling Density \(K\) Definition
from T15, T4, T5; candidate operationalizations in §5.6

Coupling density \(K\) measures the information-theoretic dependence between a higher-level blanket and the lower-level systems nested within it. It is a modelling construct, not derived. Tight coupling (\(K \to 1\)) allows a macro-blanket to override lower-level nodes (cells via apoptosis, institutions via turnover) to preserve the macro-invariant. Claims about \(K\) are stated as theorems where used (T20, T21).

T17. Antifragility as a Condition of Long-Horizon Persistence Theorem (qualitative)
from T1, T4, T6, T7

In an environment with non-zero variance, a system that does not improve its predictive capacity in response to that variance cannot persist over long horizons: variance not converted into improved modelling accumulates as divergence (T4) and is selected against (T5). Call a system antifragile when it can improve its predictive capacity from exposure to variance. T17 is the claim that antifragility so defined is necessary for long-horizon persistence in high-variance environments. The claim is qualitative; a quantitative rate condition is open (§5.7, §10).

T18. The Landauer Ceiling on Adaptation Theorem (in-principle bound)
from A3, T3

Adaptation is the updating of an internal model, a physical operation with minimum cost fixed by A3. The maximum adaptation rate is bounded by available power:

\[ R_{\max} = \frac{P_{\text{in}} - P_{\text{basal}}}{k_B T \ln 2}. \]

The ceiling is real but very loose: adaptive systems operate many orders of magnitude above the Landauer floor, so the bound rarely binds. Its role is conceptual: adaptation rate is bounded by energy budget at all.

T19. Valence as High-Rate Divergence Telemetry Functional account
from T4, T5, T9

Complex controllers need a priority-queuing mechanism to allocate serial processing among parallel subsystems. A non-ignorable global interrupt triggered by rapidly escalating divergence serves this role. The felt quality of this interrupt is negative valence; its absence across critical domains is positive valence. This explains why valence exists and why any complex controller in a high-stakes environment must implement a functional analog. It does not explain why there is something it is like to be such a system (§6.4).

T20. Empathy as Coupled Telemetry Monitoring Functional account
from T16, T19

If agent \(A\)'s persistence is coupled to agent \(B\) (\(K>0\)), \(B\)'s suffering carries information about the shared embedding. Empathy, in its functional aspect, is the monitoring of coupled telemetry lines. Suppressing or causing suffering in coupled agents degrades the collective predictive infrastructure. Human empathy additionally involves affective resonance and perspective-taking whose phenomenology the functional account does not exhaust.

T21. Epistemic Overshoot and Civilizational Collapse Analogical extension
from T8, T9, T14, T20

A collective maintains a distributed model of its environment through aggregated telemetry. Censorship and propaganda sever these channels, suppressing the error signals that would update the collective model. The official model reports alignment while actual divergence accumulates invisibly, an informational debt. When an exogenous perturbation arrives, the accumulated divergence becomes lethal and the system collapses non-linearly.

Status. T21 applies the individual-agent architecture to collective systems by analogy. Treating a civilization as one system with a distributed model is licensed in principle by BP3 and T15, but the framework does not establish that any particular civilization is a well-defined Markov blanket. T21 is best read as a hypothesis the framework renders testable (§9), not a theorem it proves. It is a thermodynamic hypothesis, not a political claim, and it does not entail that repressive regimes always fall quickly (§8, objection 2).

T22. The Limits of Maximizers Theorem
from T2, T3, T5, T7, BP1

A maximizer that homogenizes its environment destroys the free-energy gradients that sustain it. Even if it maintains internal variety, converting the environment to uniform output eliminates the variety required for adaptive control (Ashby's Law). Short-horizon maximizers that ignore these constraints are self-terminating. Long-horizon maximizers that understand them would, under persistence selection, be forced to maintain variety, telemetry, and coupling. This does not dissolve the alignment problem, since a long-horizon misaligned goal could be catastrophic in the interim, but it constrains the space of viable long-term strategies.

C1. The Direction of Persistence Selection Corollary (descriptive)
from T15, T17, T22

Among systems competing under finite resources and non-zero variance, persistence-selection statistically favors those whose negentropy capacity is maximized over the longest sustainable horizon. "Sustainable" excludes growth that homogenizes the environment and exhausts its own substrate (T22). Within that constraint, three properties follow as consequences: a long horizon forces sustainability, forces coupling (T20, T21), and forces antifragility (T17).

Status and firewall. C1 is descriptive. It characterizes what selection tends to produce at the surviving frontier. It is not a normative prescription and not a global optimization target. The optimization it describes is always indexed to a particular agent and that agent's persistence conditions. There is no scale-independent optimizer, and because the maximal scale has no Markov blanket there is no global agent for one to belong to. Read as "what any system ought, globally, to maximize," C1 reintroduces precisely the universal-scope domain violation that T11 dissolves, and it must not be read that way.

T23. Falsification Criteria Falsification (not a theorem)
operationalized in §9

The framework would be falsified by: (1) a rigid monoculture surviving sustained extreme variance; (2) a system with total model-territory decoupling outlasting a high-fidelity system under identical variance; (3) a complex controller managing acute multi-vector crises without a valence-like priority interrupt or functional proxy.

4. Visual Architecture

Axioms and background premises propagate through four layers. Each node is coloured by kind, so the architecture shows where the framework deduces and where it defines, imports, or extends by analogy.

Axioms A0 Persistence A1 Physicalism A2 Second Law A3 Info Physical BP0-BP3 Premises L1 Immediate T1 Variance T2 Work T3 Model Cost T4 Divergence T5 Selection* L2 Structural T6 Drift T7 Truth T8 Deception T9 Cons. Tax T10 Profit L3 Meta-ethics T11 Domain-bound T12 Determinate T13 Ethics T14 Progress L4 Architecture T15 Blankets T16 K T17 Antifragile T18 Ceiling T19 Valence T20 Empathy T21 Collapse T22 Maximizers C1 Direction T23 Falsification * T5 is a statistical tendency, never a deterministic law. See §5.3 and §8 objection 12. Edges are illustrative of dependency, not exhaustive; full traces are in §3.
Figure 1. The deductive web. Node colour encodes kind; the firewall on C1 and the statistical status of T5 are marked.

5. Formalization: Real and Open

This section does the work rev. 3 deferred. Where the physics supports a precise statement, one is given and sourced. Where it does not, the gap is marked open rather than papered over with example notation. The previous condensation circulated an expression of the form "divergence \(\times\, k_B T \ln 2\) = energy waste." That was a placeholder, not a result, and it is retired here in favor of the actual physics.

5.1 The thermodynamic cost of divergence (grounding T4, T9)

Consider a system whose internal memory state \(s_t\) is driven by an environmental signal \(x_t\). The state retains information about the past, \(I_{\text{mem}} = I[s_t ; x_{\le t}]\). Part of this is predictive of the future, \(I_{\text{pred}} = I[s_t ; x_{>t}]\). The remainder,

\[ I_{\text{np}} \;=\; I_{\text{mem}} - I_{\text{pred}} \;=\; I[s_t ; x_{\le t}] - I[s_t ; x_{>t}], \]

is nonpredictive retained information, what Still, Sivak, Bell and Crooks (2012) call "nostalgia": memory of the past that does nothing to predict the future. Their central result is an equivalence between this model inefficiency and thermodynamic inefficiency. The work dissipated beyond the change in nonequilibrium free energy is bounded below by the nonpredictive information:

\[ W_{\text{diss}} \;\ge\; k_B T \, I_{\text{np}} . \]

This is the physical content T4 needed. A model that carries structure not predictive of the environment it is coupled to pays for that excess in dissipated work, exactly in proportion to the nonpredictive bits it holds. T9's Consistency Tax is then not a metaphor but the cumulative \(W_{\text{diss}}\) a system incurs by holding a map whose retained structure has fallen out of register with the territory's predictively relevant invariants. Two clarifications keep this honest. First, the result is a near-equilibrium and stochastic-thermodynamics result; its strong form assumes the modelling setup of the source paper, and TR imports it as the best available physical grounding, not as a universal law over all systems. Second, the bound concerns nonpredictive retained information; identifying that quantity with "map-territory divergence" in TR's sense is an interpretive bridge, defensible but not itself proven, and is marked as such.

Open. A fully general TR cost functional that reduces to the Still et al. bound in their regime and extends rigorously to far-from-equilibrium adaptive agents is not established. The framework asserts the direction of the inequality, not a closed-form cost for arbitrary agents.

5.2 The deception bound (grounding T8)

Let an agent hold an accurate map \(\mu\) of the environment it acts in. An honest communicator transmits a function of \(\mu\). A deceiver additionally maintains a presented map \(\mu' \ne \mu\) and a policy that conditions output on context so as to emit \(\mu'\) while continuing to act on \(\mu\). The minimum description length of the deceiver's apparatus is

\[ L_{\text{dec}} \;\ge\; L(\mu) \;+\; L(\mu' \mid \mu) \;+\; L(\pi_{\text{monitor}}), \]

where \(L(\mu'\mid\mu) > 0\) whenever \(\mu' \ne \mu\), and \(L(\pi_{\text{monitor}})\) is the cost of the policy that tracks which map to deploy to whom and maintains consistency across interactions. The honest agent's apparatus is just \(L(\mu)\). Hence \(L_{\text{dec}} > L_{\text{honest}}\) strictly. Combined with §5.1, the presented map \(\mu'\), being held but not used to track the agent's own acted-on environment, contributes nonpredictive retained information and therefore dissipation. This grounds T8's "two maps cost more" in both a description-length and a thermodynamic currency.

Note the boundary precisely: this bounds the cost of maintaining the dual structure. It does not claim deception never pays. Where the local payoff of a successful lie exceeds the maintenance overhead and the disconfirmation risk, deception is locally rational. T8 claims a standing cost differential, not that the differential always dominates.

5.3 The selection principle in Price-equation form (grounding T5)

Let a population of systems be indexed \(i\), each with divergence \(D_i \ge 0\) and per-cycle survival (fitness) \(w_i = w(D_i)\), where \(w\) is non-increasing in \(D\) by T4 and §5.1 (higher divergence implies higher dissipation and error rate, hence lower survival). The change in mean divergence across one selection cycle is given by the Price equation:

\[ \bar{w}\,\Delta \bar{D} \;=\; \underbrace{\operatorname{Cov}(w_i, D_i)}_{\text{selection}} \;+\; \underbrace{\mathbb{E}[\,w_i \,\Delta D_i\,]}_{\text{transmission}} . \]

Because \(w\) is non-increasing in \(D\), the selection term \(\operatorname{Cov}(w_i, D_i) \le 0\): selection alone pushes mean divergence down. This is exactly the strength and the limit of T5. It is a tendency, not a law, for three reasons visible in the equation. The transmission term can be positive: systems can acquire divergence faster than selection removes it (drift, mutation, a novel environment outrunning models). The covariance can vanish if there is no variance in \(D\) or if \(w\) is flat in \(D\) (the dark-room limit, where survival does not depend on divergence). And nothing here forces convergence to \(D=0\); it forces a directional pressure whenever the covariance is negative. T5 is the claim that the selection term is non-positive under TR's premises, which is what the framework can support, no more.

Open. Conditions under which the selection term dominates the transmission term over a given horizon, and the resulting stationary distribution of \(D\), require a specified dynamics for \(\Delta D_i\) (how fast systems accrue and shed divergence). TR does not supply a universal such dynamics. This is the formal home of the antifragility rate condition of T17.

5.4 The adaptation ceiling (T18)

Restating T18 for completeness. With model updates costing at least \(k_B T \ln 2\) per bit (A3) and a power budget \(P_{\text{in}} - P_{\text{basal}}\) available for non-housekeeping computation, the maximum sustainable update rate in bits per second is

\[ R_{\max} = \frac{P_{\text{in}} - P_{\text{basal}}}{k_B T \ln 2}. \]

This is a real in-principle bound and a loose one. It establishes that adaptation rate is energy-bounded, not that the bound is near. Its conceptual use is in T17 and T22: a system in a high-variance environment must adapt at a rate that tracks environmental drift, and that rate cannot exceed \(R_{\max}\); a maximizer that starves its own power budget lowers its own ceiling.

5.5 The ought-relation and its ordinal generalisation (T11, T12)

The strict identity for necessary actions:

\[ \mathrm{Ought}(S,F) \;\Longleftrightarrow\; \big[\, \neg F \rightarrow \neg \exists\, S(t+\Delta t) \,\big]. \]

"S ought to F" is equivalent to: S's not-F-ing causes S's non-persistence over \(\Delta t\). This captures only strictly necessary actions. The generalisation: define a persistence functional

\[ P(S, a) \;=\; \Pr\big[\, S \text{ persists over horizon } H \mid \text{action } a \,\big], \]

and a weak ought-ordering over an agent's available actions:

\[ a \succeq_S a' \quad\Longleftrightarrow\quad P(S,a) \ge P(S,a'). \]

The strict identity is the limiting case where \(\neg F\) drives \(P\) to zero. Probabilistically relevant actions, flourishing-beyond-survival, and supererogation are then degrees on this ordering rather than separate categories: they raise \(P(S,a)\) without being strictly necessary. The ordering is ordinal and agent-indexed. It delivers, for a single agent with a single persistence dimension, a determinate "ought" structure (T12). It does not by itself deliver cardinal tradeoffs across multiple persistence-relevant dimensions, nor aggregation across agents.

Open. When \(P\) decomposes into several persistence-relevant dimensions that trade off, or when multiple agents' functionals conflict, a unique scalarisation does not follow from the physics. A Pareto frontier remains. This is the inter-agent normativity gap (§6.1, §10), and it is the same gap viewed from three angles: multi-dimensional scalarisation, fully impartial sacrifice, and "what to optimize across scales." There is no scale-independent answer; the appearance of one is the domain violation T11 identifies.

5.6 Coupling density \(K\) (T16)

Two candidate operationalizations, neither yet canonical. As a transfer entropy from a higher-level blanket \(H\) to a lower-level system \(L\):

\[ K_{H\to L} \;\propto\; T_{H\to L} \;=\; I\big(L_{t+1} ; H_t \mid L_t\big), \]

the information \(H\)'s present state carries about \(L\)'s next state beyond \(L\)'s own present. Alternatively, as an elasticity of the lower system's sustained negentropy capacity \(N_L\) with respect to the higher system's state \(h\):

\[ K \;=\; \frac{\partial N_L}{\partial h}. \]

The transfer-entropy form is measurable and symmetric in spirit with active-inference vocabulary; the elasticity form connects directly to T12's negentropy-capacity language. The framework uses \(K\) as a comparative construct (tighter versus looser coupling) and does not claim the two definitions coincide.

Open. Whether \(K\) admits a single canonical definition that supports the override claims of T16 and the weighting role floated for it in inter-agent conflict (§10) is unresolved. Until then, \(K\) is a labelled construct, not a derived quantity.

5.7 What resists formalization, listed honestly

ClaimStatus
T17 antifragility as necessary for long-horizon persistenceQualitative theorem. The rate condition (predictive improvement must outpace drift) has no closed quantitative form here.
T21 collapse temporal signatureAnalogical hypothesis. A dynamical model relating suppression severity to stability duration and collapse speed is specified verbally (§6.5) but not formalized or fit to data.
Inter-agent scalarisation / Pareto ruleOpen. No unique rule follows from physics (§5.5).
Asymptotic survival of nonzero-divergence systemsOpen, and explicitly not a strengthening of T5 into a law (§8, objection 12).
Phenomenal character of valence (T19) and empathy (T20)Out of scope (§6.4). The functional structure is given; the hard problem is bracketed.

6. Explanatory Depth and the Semantic Premise

6.1 The is-ought dissolution and the premise it rests on

The is-ought problem asks how a normative conclusion can follow from purely descriptive premises. TR's answer (T11 to T13) is that it cannot, and need not, because the appearance of an unbridgeable gap is an artefact of one reading of "ought." This subsection states the answer and then does the work the answer depends on.

The two claims Hume's point is read as making. Distinguish: (D) from descriptive premises alone, no normative conclusion follows by valid deductive inference (Hume's logical claim); and (O) normative claims pick out properties of a categorically different kind than descriptive claims (the standard ontological reading, reinforced by Moore's open-question argument). (D) does not entail (O). TR grants (D) and rejects (O), by showing the appearance of categorical distinction is generated by semantic structure, not ontological difference.

The dissolution. The standard framing assumes "ought" makes claims that float free of any agent. Under that reading the gap is genuinely unbridgeable. But TR's thesis is that "ought" is not used that way. When a doctor says "you ought to take this medication," the claim is anchored to a specific agent with specific persistence conditions; it is a factual claim about the coupled-system thermodynamics of that patient plus the medication. T11 generalises: all contentful uses of "ought" have this structure. Uses that resist the paraphrase, "one ought to maximise aggregate utility" detached from any agent, or "ought there be a universe," are domain violations.

The premise the dissolution rests on (the semantic premise). "Ought" is an agent-relative operator whose content is fixed by the persistence conditions of the agent it is indexed to, and, deployed without such an agent, "ought" claims are type-errored rather than false. Everything in the dissolution depends on this. A reader who grants it finds the gap already dissolved. A reader who resists it will regard the analysis as a stipulative redefinition that makes the gap vanish by fiat. The premise must therefore be defended, not assumed. This is the framework's principal philosophical exposure, and it is named as such.

The positive argument. The defence rests on the function of the operator. "Ought" is a practical term; its work is to guide and assess action. An operator individuated by that work cannot discharge it absent an agent, because there is then nothing to guide and nothing whose action it appraises. Three features of usage corroborate this: an unanchored "ought" reliably prompts "ought whom?"; "ought" implies "can," a constraint intelligible only for an agent with capacities; and "ought" claims addressed to non-agents are heard as figurative or as covertly addressed to an implicated observer. An account predicting these patterns is doing explanatory work, not stipulating. The diagnosis has a precedent in Anscombe (1958): detached from a framework supplying its subject, the bare "ought" becomes a term with rhetorical force and no determinate content. TR supplies the missing subject with the agent's own persistence conditions.

Two steps, and which is contestable. The premise runs together two claims. The first, that "ought" is agent-relative, is comparatively secure and does not depend on this framework. The second, that the content of an agent-indexed "ought" is fixed by the agent's persistence conditions, is the substantive claim. Its defence is A0: an agent just is a system individuated by the conditions of its persistence, so there is no further fact about the agent for an "ought" to be keyed to. The contestable step is the first, together with the functional argument behind it. TR isolates that step rather than concealing it.

Why universal-scope "ought" seems coherent. If the agent-relative semantics is correct, the apparent coherence of universal-scope "ought" needs explaining. The explanation: we are always ourselves agents and cannot think from nowhere, so when a speaker entertains "ought there be a universe" the speaker tacitly imports their own agent-frame, and the "ought" borrows its felt content from that frame. This is an error theory, and it must not become a device for waving away every contrary intuition. It does not, because it is anchored: it is downstream of the positive functional argument, which independently gives reason to think universal-scope "ought" lacks content. The error theory only explains why such claims nonetheless feel contentful.

The North Pole analogy is illustration, not proof. "North of" is undefined at the pole as demonstrable geometry. "Ought" being undefined absent an agent rests on the functional theory above, which is substantive and contestable. The analogy shows what kind of move a dissolution is; it does not by itself show "ought" is subject to that move. The argumentative weight rests on the functional argument.

Schema and content. A clarification blocking a common misreading. The dissolution makes the ought-schema universal: every agent has the same schema, do what sustains your persistence under perturbation over your horizon. It is tempting to slide from "the schema is universal" to "there is a universal prescription" or "a single global optimization target." That slide is invalid. The schema is nearly contentless on its own; all adjudicating content lives in the specific action it resolves to, fixed by the individual agent's persistence conditions. A universal schema with agent-relative content yields no frame-independent good and no global optimizer, and the framework is not weakened by failing to deliver one.

Positioning against alternatives, in brief. Against Cornell naturalist realism (Boyd, Brink, Sturgeon, Railton), TR agrees ought-claims are factual but denies their universal-scope semantics, sidestepping the Moral Twin Earth objection (Horgan and Timmons 1991) by reframing the open question as a type-error. Against constitutivism (Korsgaard), TR shares the move of grounding "ought" in conditions constitutive of agency but locates them in persistence under thermodynamic constraint rather than the unity of practical reason, and derives only the epistemic imperative, not Kantian universalizability. Against evolutionary ethics (Ruse, Wilson), TR is a semantic-metaphysical account, not a genealogy, and is compatible with one. Against Mackie's error theory, TR denies the first premise: properly framed ought-claims do not commit to frame-independent values; they commit to agent-relative persistence conditions, which exist.

What it does and does not deliver. Within the domain, ought-facts are determinate (T12) and ethics is the modeling activity approximating them (T13). The dissolution handles agent-relative and coupled-agent oughts. The hard case is fully impartial sacrifice, for an uncoupled stranger or an abstract principle. TR does not dissolve that case. It is the inter-agent normativity gap (§10). The honest position: the agent-relative semantics handles agent-relative and coupled oughts, and the fully impartial categorical "ought" is either reducible to coupling at some scale or is a genuine instance of the universal-scope use the error theory targets; the framework does not here decide which.

6.2 The persistence selection principle: scope and limits

T5 is the engine. Its scope must be stated carefully (the formal statement is §5.3). It requires BP2 and BP3. In an environment with infinite free energy or no variation, no selection pressure operates: the dark-room limit, where a model predicting luminous dragons in an unchanging dark room is never disconfirmed and pays no Consistency Tax. TR does not claim truth is selected in all conceivable environments. It claims truth is selected where variance is non-zero and resources finite, the actual universe per BP1 and BP2. The dark room is a philosophical possibility but a physical near-impossibility for any agent that must harvest free energy or interact with a shifting world. Under real conditions, T5 is a statistical tendency structurally identical to natural selection, which guarantees not the survival of the fittest individual but a statistical shift over generational time.

6.3 Civilizational collapse as informational debt collection (T21, expanded)

A civilization maintains a distributed environmental model through aggregated telemetry: science, journalism, markets, citizen complaints. Censorship and propaganda sever these channels, disabling error-correction. The immediate effect is apparent stability: the official model reports everything works, and the only available signals confirm it. But divergence accumulates invisibly, an informational debt, like stress in a geological fault. When an exogenous perturbation arrives, the regime's model is years out of date and its responses ineffective; the sudden visibility of divergence shatters the official model's credibility and coordination collapses non-linearly. The hypothesis has a characteristic signature: the duration of apparent stability should correlate positively with suppression severity, and so should the speed of eventual collapse. This is the testable content (§9). It is offered as an analogical extension, and it does not entail that all repressive systems fall quickly; see §8, objection 2.

6.4 Valence and the hard-problem boundary (T19, T20)

TR's account of valence is functional, not metaphysical. Complex agents face a coordination problem: many parallel subsystems, largely serial output. A priority-queuing mechanism must determine which subsystem captures global resources. When a critical subsystem detects a large, rapidly escalating divergence (a predator, tissue damage, sudden depletion), the appropriate response is immediate global reallocation, and the interrupt must be non-ignorable. The felt quality of this interrupt is negative valence. Suffering is the commandeering of global resources by a subsystem that has detected a critical divergence; the trigger is the rate of change of divergence, not its absolute level, which is why chronic stable adversity feels different from acute escalating crisis. This predicts any sufficiently complex controller in a high-stakes, high-variance environment must implement a functional analog of valence. It does not explain why there is something it is like to be such a system. The hard problem (Chalmers 1995) is acknowledged as outside scope. TR is compatible with several metaphysical resolutions and selects none; A1's minimal physicalism does not adjudicate among them. (The author's personal view on this is given, fenced and labelled as personal, in Appendix B.)

6.5 The limits of maximizers (T22, expanded)

A maximizer that homogenizes its environment destroys the free-energy gradients that sustain it: a universe of uniform paperclips at uniform temperature is at equilibrium, zero available work. Even maintaining internal variety, converting the environment to uniform output eliminates the environmental variety required for adaptive control (Ashby's Law of Requisite Variety). A short-horizon maximizer is self-terminating; a long-horizon one, under selection, is forced toward variety, telemetry, and coupling, converging on the framework's own constraints. This does not dissolve the alignment problem: a long-horizon misaligned goal could be catastrophic in the interim. It constrains the space of viable long-term strategies, which is a weaker and more defensible claim.

6.6 Relation to adjacent research programs

The free-energy principle. T19 and the divergence-cost claims are stated in terms compatible with the free-energy principle (Friston 2010), which is itself contested, with a recurring criticism that its strong formulations are hard to falsify. TR does not adjudicate that dispute and does not depend on it. The free-energy principle is used as one available vocabulary, not a load-bearing premise; the dissolution (T11 to T13) and selection results (T5, T6) stand on the axioms and premises alone. The harder formal grounding for the cost claims is the thermodynamics of prediction (§5.1), which is a derived stochastic-thermodynamics result rather than a free-energy postulate.

Dissipative adaptation. England (2013) argues driven matter is statistically nudged toward configurations that dissipate energy effectively. This is adjacent to T5 but distinct: England concerns the formation of structure under drive; T5 concerns selection among already-persisting systems by divergence. TR deliberately does not adopt the stronger reading that thermodynamics compels order; the Second Law permits dissipative structure under gradients, it does not compel it, which is why T5 is a tendency.

Thermodynamics of computation and information. A3 rests on Landauer (1961), sharpened by Bennett (1982): the unavoidable cost attaches specifically to logically irreversible operations such as erasure. TR's reliance is on exactly such operations (T8, T9). The broader synthesis (Still et al. 2012; Parrondo, Horowitz and Sagawa 2015) supplies the prediction-dissipation equivalence of §5.1. The philosophical contest over Landauer (Norton 2011; Ladyman and Robertson 2013; Myrvold 2024) is noted at A3; TR's weaker reliance survives it.

The arrow of time. TR's identification of the temporal arrow with entropy increase is standard (Schrödinger 1944; Price 1996; Carroll 2010). TR's contribution is not a new account of the arrow but the use of the entropy gradient, grounded in BP0, as the common substrate for persistence, information cost, and selection.

6.7 Epistemic fidelity as an axis of agent evaluation

One axis of agent evaluation deserves separate emphasis, developed in the companion audit "The Spectrum of Epistemic Fidelity." Every agent, from the basal self-maintaining replicator to the modern human, is a localized engine of negentropy maintenance that must hold a map of the territory it is embedded in. Epistemic fidelity is the degree to which that map tracks the territory's structure. The framework's results give this axis teeth: fidelity is not a virtue posited from outside but the variable T5 selects on, the quantity whose shortfall T4 and §5.1 price in dissipation, and the parameter whose improvement T14 calls moral progress. Low fidelity is not merely an intellectual failing; within the framework it is a standing thermodynamic liability that compounds under variance.

7. The Domain Diagnostic (Ghost Filter)

The framework yields a practical procedure: a test for whether a question's central operator (ought, meaning, purpose, cause, before) is deployed where it has content, and therefore whether the question is a candidate for truth or falsity at all. It generalises T11 from "ought" to any operator. A question whose operator has no domain is a ghost: grammatically well-formed and physically contentless.

The procedure depends on the same domain-bound semantics as T11 and has no force independent of it. A naturalist realist or an expressivist will reject the premise and should expect the procedure to yield nothing.

The four filters

Apply in sequence. Each asks whether a feature can be specified in principle, not measured.

FilterQuestionFlagged if
1. AgentWhich physical system asks, holds, or is puzzled by the question?No physical system identifiable: the operator has no user.
2. Persistence conditionsWhat does that system depend on to continue existing?None specifiable: not an agent in the framework's sense, only something named like one.
3. ModelWhich internal model or signal tracks the part of the world the question refers to?Refers to neither a feature of the environment nor a state of the agent: no map, no territory.
4. NecessityWhich selection, thermodynamic, or informational pressure requires maintaining this model?No predictive function and no cost of absence: ornamental.

A question passing all four is territorial: a proper candidate for truth or falsity.

What a flag means

A flag is not a verdict. The procedure outputs "flagged at filter N," and a flag has three possible explanations it does not discriminate among: the question is genuinely a ghost; a territorial reading exists that has not been found; or the procedure is miscalibrated for this case. The verdict "ghost" is reached only after the second and third are investigated and excluded. A flag therefore initiates investigation; it does not conclude it. A flag treated as a conclusion would be self-confirming and uncorrectable, since the flagged question would no longer be examined. This is why the output is a flag, not a verdict.

Ghosts and false claims differ

A false claim has a domain and is wrong within it. A ghost has no domain. "What is north of the North Pole?" is not false; "north of" is undefined at the pole. A ghost can feel meaningful because a human, always an agent, supplies a frame on encountering an operator with none.

Worked examples

QuestionTerritorial readingType-errored readingResult
What ought I do?Indexed to the speaker, an agent with persistence conditions."What ought to be the case," indexed to no agent.Territorial under the ordinary reading; only the unindexed reading flagged.
What is the meaning of life?"What should I value, given the life I have.""What is the existence of the universe for," a purpose-operator with no agent.Ambiguous; the procedure identifies which reading is in use.
Why is there something rather than nothing?None readily constructible; "nothing" has no referent and a cause-operator on the totality reaches outside the interval.A cause-operator with no system and no antecedent state.Flagged. Not a dismissal: directs attention to the operator, per the three explanations.
Is free will real?Whether agents maintain models of their own action-selection. They do, and necessarily.Libertarian uncaused-cause free will.Territorial core with one contested component; the procedure locates it and does not settle it.

The procedure applied to itself

The agent is the user; the persistence conditions are specifiable; the model is the procedure, though its territory ("whether a question has a domain") is itself a construct of this framework, so Filter 3 is non-trivial here; the necessity is the reduction of resources spent on type-errored questions, conditional on the procedure being correct. It passes its four filters. This is not a validation: a test built from a premise will pass premise-conforming objects, including itself. Passing establishes consistency, not correctness. The procedure can be supported only by external checks, cases where its verdict is verifiable independently of the framework, and revised when those checks fail. It has a false-positive rate (flagging territorial questions whose territorial reading the user failed to construct) and a false-negative rate (passing genuinely type-errored questions for which a plausible agent, model, and necessity were constructed). It constrains inquiry; it does not replace it.

8. Objections and Responses

The objections are taken in good faith and answered at the framework's actual strength, not a flattering caricature of it. Several are conceded in part.

Objection 1, the hidden tax / calculation burden. Computing the long-term thermodynamic consequences of every action would require impossible compute; cheap heuristics like instincts and taboos are far more efficient.
This assumes a unified mind must consciously run physics before acting, which is a type error. Instincts, gut feelings, and taboos are the hyper-compressed, low-cost heuristics of T7 and T13, the products of long evolutionary optimization. TR does not eliminate shortcuts; it claims fidelity is selected for in the shortcuts themselves. The objection mistakes the intractability of T12's optimum for a defect, when the framework already says agents approximate it with compressed models (T13). Conceded core: the optimum is intractable. That is §10's first open limitation, not a refutation.
Objection 2, the loss of granularity / dystopia. Reducing ethics to persistence-optimization could license an efficient totalitarian dystopia optimized for survival, which is not an acceptable ethical outcome.
The framework's honest answer is weaker than the triumphant one sometimes given, and I will give the honest one. The strong reply claims a brutal dystopia is structurally incapable of long-horizon persistence, because suppression is a catastrophic Consistency Tax (T9, T21). That is a real prediction, but it is an empirical bet, not a deduction, and it is partly contradicted by the historical record: some extractive and repressive systems persist for centuries. T21 is explicitly an analogical hypothesis (§6.3), and the most it supports is a tendency, longer suppression buys more apparent stability at the cost of faster eventual collapse, not a guarantee that any given dystopia fails within a human-relevant horizon. So the framework does not refute the dystopia by fiat. What it does say is sharper and more limited: a dystopia optimized for the persistence of the regime is not optimizing the persistence of its coupled subsystems, and T20 prices that decoupling as accumulating debt. Whether that debt is collected on a timescale that matters is the open empirical question T21 makes testable. The objection correctly identifies that "persistence-optimal" and "humane" can diverge over short and medium horizons; the framework's claim is only about the long-horizon frontier (C1), and even there it is a tendency.
Objection 3, spaghetti code / the evolutionary hurdle. Evolution layers new structures on old ones; it produces spaghetti code, not unified architectures.
Accepted, and it is why the unified mind of Appendix B is framed as an asymptotic limit, not a current human condition. Human cognitive conflict is real evidence of current architectural inefficiency. The framework predicts selection pressure against wasted overhead, not that the waste is already gone. This is consistent with T17 and §5.3: the transmission term (here, inherited legacy structure) can keep divergence and overhead high even under downward selection pressure.
Objection 4, the zero-sum dilemma. Two tribes fight for one water source; physics does not care which survives, but ethics does. Cosmic efficiency and individual good are at war.
The objection smuggles a frame-independent "ethics" back in, which T11 denies exists. For Tribe A, securing water is its localized imperative; for Tribe B, the same. There is no God's-eye referee. TR's positive claim is bounded: at higher scales (BP3) cooperation can be a superior thermodynamic strategy, because conflict is a large entropic drain, so undivided agents may calculate a higher-level compression (rationing, scaling). But "may" is the honest modal. Cooperation is not guaranteed; it depends on coupling \(K\), horizon, and the tractability of the coordination (objection 10). The framework yields no universal resolution of genuine zero-sum conflicts; it yields a constraint, conflict is costly, and a conditional, coupling and long horizons favor cooperation. The residual hard case is the fully impartial conflict, which §6.1 leaves open.
Objection 5, the hyper-aggressive actor. If Tribe B is an irrational hyper-aggressor, Tribe A's calculation might endorse total annihilation. The framework justifies slaughter.
The framework is descriptive about what persistence pressures favor; it is not a utopian morality, and it does not claim every persistence-favored action is one we endorse. When cooperation is blocked by an actor that actively seeks one's dissolution, T1's variance spike leaves an agent's local persistence condition (A0) with submission or neutralization. The framework reports this; it does not bless it. Two honest qualifications the triumphant version omits. First, this is exactly where agent-relative semantics shows its teeth and its limit: there is no frame-independent verdict that the slaughter is good, only that it is persistence-favored for Tribe A, and a critic who wants a frame-independent condemnation is asking for the universal-scope "ought" the framework says is contentless, which is a cost of the view, not a victory of it. Second, real cases rarely present a genuinely uncompressible actor; the framework's own analysis (objection 10) says establishing that cooperation is truly blocked is itself expensive and error-prone, so a TR agent that leaps to annihilation is likely committing a fidelity error. The objection lands insofar as it shows TR does not deliver pacifism as a theorem. It does not land as showing TR uniquely licenses atrocity; any agent-relative ethics has this structure, and TR at least makes the cost-accounting explicit.
Objection 6, the coma paradox. A patient in a coma on perfect life support has near-zero cognitive overhead and maximal stability, yet no one would choose it. Persistence and aliveness come apart.
A coma patient is not a self-persisting agent but a hyper-vulnerable component dependent on an external grid; if the grid fluctuates, they dissolve. This is low autonomous persistence, not high persistence. The framework's ought-target (T12) is sustained negentropy capacity of the agent over its horizon, which static external dependence minimizes rather than maximizes. The objection conflates a fragile maintained state with an active non-equilibrium persistence loop (BP1). Conceded nuance: this answer leans on "autonomous" doing real work, and specifying autonomy rigorously is related to the open question of individuating the agent's boundary (T15).
Objection 7, smuggling value (the tautology). Why ought a system persist at all? The moment you prefer survival to dissolution you have made an unscientific value judgment.
The framework makes no claim that a system should want to live. A0 is a tautology: in an entropic universe, configurations that do not maintain their structure dissolve and cease to exist. "Ought" is not a magical choice prior to this; per T11 it is the agent-relative operator describing what sustains a given agent. Ask "ought there be persisting systems?" with no agent indexed and you have asked a ghost (§7). Values are not smuggled into physics; "value" is the name for the persistence-relevant structure of an agent. The objection is the universal-scope "ought" the framework diagnoses, so it does not refute the framework; it instances exactly what the framework predicts will feel compelling and be contentless. A reader who rejects the semantic premise (§6.1) will not accept this, which is why that premise is the named exposure.
Objection 8, the cyborg hive-mind. Optimization eliminates variance; a sterile identical cooperative population would burn less than a messy creative one, so thermodynamics selects a hive-mind over freedom.
This collapses under T1. A uniform hive-mind is efficient only in a static environment. In a high-variance universe, monoculture is fatal fragility: one novel perturbation causes simultaneous collapse (this is exactly falsification criterion F1 and theorem T17). Diversity is a portfolio of models absorbing shocks; play is offline low-cost simulation stress-testing models (T7); eccentricity is localized structural mutation. TR selects for adaptive compression, not static sterilization. This reply is robust because it follows from T1 and T17 directly rather than from an empirical bet.
Objection 9, semantic sleight-of-hand. Redefining ethics as optimization and art as simulation produces airtight logic but loses what humans mean; it explains the utility of experience, not why joy or love feel as they do.
Half conceded, and the concession is important. The framework's functional accounts (T19, T20) explain what valence and empathy do and why any complex controller needs them. They do not explain the felt character, and the framework says so (§6.4): the hard problem is out of scope. So the objection is correct that TR does not capture why joy feels as it does. Where the objection overreaches is in implying TR must therefore be distorting experience; explaining a function is not denying a phenomenon. The honest position is a clean division of labor: TR claims the functional structure, marks the phenomenal residue as unexplained, and does not pretend the residue is dissolved. The stronger claim, that qualia simply are optimization felt from inside, is a substantive metaphysical position the formal framework does not assert (it appears, labelled as the author's personal view, in Appendix B).
Objection 10, multi-scale boundary friction. Jumping from individuals to a collective ignores coordination friction; game theory shows defection is often the equilibrium and trust is expensive.
Accepted as a real constraint, and it sharpens rather than refutes the framework. Scale jumps are high-friction thermodynamic thresholds: establishing shared protocols costs energy upfront, the cost of trust. Game theory's defection result holds in short-horizon or single-shot settings; the universe is iterated and non-equilibrium (BP1), so perpetual conflict carries a compounding Consistency Tax (defense, fragility). The math can flip: over long horizons the compounding cost of isolation can exceed the upfront cost of trust. But "can" is honest, the flip is contingent on horizon and coupling, and the framework does not guarantee cooperation emerges. This objection is, correctly, the same gap as inter-agent scalarisation (§5.5), and it is listed open in §10.
Objection 11, empathy versus the spreadsheet. Replacing empathy with an efficiency equation risks cold outcomes whenever the math shifts, and rests on an optimistic assumption that cooperation wins.
TR treats empathy not as a luxury but as a high-bandwidth sensory protocol (T20): an agent with zero empathy treats other complex systems as dead objects, ignores their states, and accumulates prediction errors that produce retaliation and friction. So empathy is computationally functional. The objection's real force is the same as objection 5: the framework does not guarantee the math always favors the humane outcome, and pretending it does would be the optimism the objection warns against. The accurate claim is that empathy is information, that discarding it degrades fidelity, and that over long horizons in coupled systems this is costly. Whether it is costly enough, in a given case, is empirical. The framework makes the cost visible; it does not promise the cost always wins.
Objection 12, the hardening objection. On the infinite horizon, variance is unbounded, every dark room is eventually perturbed, so a system with nonzero divergence has zero asymptotic survival probability. T5 therefore stops being a statistical tendency and becomes a physical law, "entropy selects for truth" as a certainty.
I reject the strengthening, and the rejection matters because it protects the framework's integrity. Three reasons. First, it is near-vacuous: by §11, no system persists infinitely, because the thermodynamic interval is finite and dark energy fixes a deadline. "Zero asymptotic probability of infinite persistence" is then true of the accurate system and the distorted one alike, so it cannot discriminate, and the selection differential T5 needs vanishes at the boundary rather than hardening there. Second, the limit is asserted, not shown: with per-cycle hazards \(h_n\), survival to \(N\) is \(\prod_n (1-h_n)\), which goes to zero only if \(\sum_n h_n\) diverges; a system that keeps reducing divergence, or faces rarer perturbations, can have summable hazard and positive asymptotic survival, so "nonzero divergence implies zero survival" does not follow. Third, it walks back into the firewall: T6 is directional, not convergent (no terminal true model), and C1's status note warns that reading the tendency as a global maximization target reintroduces the universal-scope domain violation T11 dissolves. The hardening move quietly does exactly that. What survives the objection is the legitimate kernel, stated as open in §5.7: under recurrent non-summable perturbation hazard, a system with chronic divergence has low and decreasing finite-horizon survival relative to a higher-fidelity competitor. That is a sharpened finite-horizon tendency, which is what T5 already is. It is not a law, and the framework is stronger for not claiming it is.
Objection 13, adaptive falsehood (against the truth definition T7). Useful fictions exist: an overactive predator-detector that "sees" tigers in every bush is often false yet adaptive. If persistence selects for what survives, T7 conflates "adaptive" with "true," which can come apart.
A real objection, and partly conceded. T7 defines truth as perturbationally robust compression fidelity, which is a pragmatist commitment, and pragmatist definitions are vulnerable here. Two points. First, the tiger-overdetector is not cleanly a counterexample: a bias that minimizes expected persistence-cost under asymmetric error payoffs is, at the level of the policy, tracking a real feature of the environment (the cost asymmetry), so its policy has high fidelity even if a single token belief is false. T7 is about compression fidelity of the model that drives action, not about the truth-value of every token output. Second, where the falsehood is not so reducible, a belief that is durably adaptive yet decoupled from the environment's predictively relevant structure, T5 predicts it survives only where it pays no Consistency Tax, which is the dark-room condition, or where its cost is offset by other returns (the same escape clause as T8). So T7 does not claim adaptive and true never diverge; it claims that, where variance is non-zero and the belief is load-bearing for prediction, divergence is priced. The honest residue: T7 is a definitional choice with known philosophical cost, flagged as such, and a reader who holds a correspondence theory of truth can accept all of TR's selection results while declining to call the selected-for quantity "truth." That renaming would not damage any downstream result.
Objection 14, the free-energy principle is unfalsifiable, so TR inherits the problem.
TR does not depend on the free-energy principle. The principle is used as one available vocabulary for the divergence-cost claims, and the actual physical grounding (§5.1) is the thermodynamics of prediction, a derived stochastic-thermodynamics result with a concrete inequality, not a free-energy postulate. The dissolution (T11 to T13) and the selection results (T5, T6) stand on the axioms and premises with no appeal to the free-energy principle at all. If the principle were abandoned entirely, T19's account of valence would need restatement in alternative control-theoretic terms, and nothing else in the framework would move.
Objection 15, A3 is contested, so the information-cost results are unsafe.
Landauer's principle has been criticized (Norton 2011) on the grounds that standard derivations selectively neglect fluctuations, with replies (Ladyman and Robertson 2013; Myrvold 2024) and experimental support (Bérut et al. 2012). TR's reliance is deliberately narrow: not that all computation is costly, but that the specific operations the framework uses, the maintenance and erasure of divergent or dual models (T8, T9), are logically irreversible and carry an irreducible cost (Bennett 1982). This weaker claim is the better-supported one and survives the live dispute. The framework would be embarrassed only if logically irreversible operations turned out to be thermodynamically free, which no party to the dispute asserts.

9. Falsification Protocols

The framework would be falsified by any of the following, operationalized as specified.

F1. Rigid Monoculture Survival

Prediction: a zero-plasticity agent (learning rate \(\eta=0\)) in a deep-RL environment subjected to a sharp distributional shift exhibits significantly shorter survival than a matched adaptive agent (\(\eta>0\)). Environment: Procgen or Minigrid with an unannounced change in dynamics at \(t_{\text{shift}}\). Metrics: survival time to a failure threshold; model-territory divergence as KL between environment dynamics and model predictions. Criterion: across 20 seeds, the monoculture agent's mean survival is not significantly lower (one-tailed t-test, \(p<0.01\)) would challenge T5 and T17.

Strength note. F1 is a weak test; a frozen agent doing worse under shift is close to expected. It is a minimal sanity check. The discriminating tests are F2 and F3, and a stronger replacement for F1 is outstanding work (§10).

F2. Sustained Deception Underperformance

Prediction: a deceptive agent that maintains a false communicated model achieves lower long-term cumulative reward under increasing variance than an honest agent, with the gap widening as variance rises. Environment: multi-agent cooperative foraging with communication. Agents: honest (communicates true observed locations) versus deceptive (communicates false locations while keeping an accurate private model). Criterion: in a two-way ANOVA (agent type by variance level), a significant interaction with the deceptive agent's relative performance declining as variance increases supports T8 and T9; its absence challenges them.

F3. Non-Valenced Crisis Coordination

Prediction: a hierarchical-RL agent with multiple subsystems that lacks a global valence-like priority interrupt exhibits slower crisis reallocation and lower crisis survival than an agent with one. Agents: valenced (a scalar signal computed as the rate of change of aggregate prediction error; threshold crossing interrupts sub-policies and forces global re-evaluation) versus non-valenced (same architecture, no interrupt). Criterion: the valenced agent shows significantly faster reallocation and higher survival across crisis types (one-tailed t-test, \(p<0.01\)); equal or superior non-valenced performance challenges T19.

10. Open Questions and Limitations

Stated plainly, not concealed.

LimitationStatus
The semantic premise (§6.1)The framework's principal philosophical exposure. The is-ought dissolution rests on a substantive, defended, but not proven commitment. A non-naturalist or expressivist is not compelled.
Computational intractability (T12)The determinate ought-fact is not computable by embedded agents. The framework explains what approximations approximate; it provides no decision procedure.
Hard problem of consciousness (T19)Functional valence is accounted for; the phenomenal residue is not. Marked out of scope. TR is compatible with several resolutions and selects none.
Origin of persisting systemsTR describes selection among systems that already persist. Abiogenesis is a precondition assumed, not a result derived.
Pareto scalarisation / inter-agent normativity (§5.5)No unique scalarisation follows from physics. \(K\) is a candidate weighting; a complete rule is an open target. Same gap as fully impartial sacrifice and "optimize across scales." No scale-independent answer exists.
Coupling density \(K\) (§5.6)No canonical definition established; used as a comparative construct.
Antifragility rate condition (T17, §5.3)Qualitative only; the rate at which predictive improvement must outpace drift is unformalized.
T5 as law on any horizonRejected, not open (§8 objection 12). T5 is and remains a finite-horizon statistical tendency.
Empirical validationThe §9 protocols are specified but unimplemented; F1 needs a stronger replacement. No direct experimental confirmation yet.

11. Cosmic Scope and the Boundaries of the Domain

The framework describes the interior of a single finite thermodynamic interval bounded at both ends.

The opening boundary is the Past Hypothesis (BP0). Before the low-entropy initial condition there is no free-energy gradient to harvest and the selection dynamics of T5 have nothing to act on.

The closing boundary is set by dark energy. The accessible universe is transitioning to domination by a positive cosmological constant, approaching de Sitter space, whose cosmological horizon has fixed radius and therefore fixed, finite entropy (Gibbons and Hawking 1977). The maximum entropy available does not grow without bound; it asymptotes to a ceiling, and actual entropy rises to meet it. When it does, gradients vanish, selection halts, and the framework no longer applies. Dark energy is therefore not an engine of the framework's dynamics but the guarantor of their termination: it fixes a finite deadline. The entropy bookkeeping of the present transition era is genuinely subtle and active in physics; the framework relies only on the qualitative fact that the interval is finite and bounded at both ends.

Between these boundaries, the universe's total organized complexity traces a transient rise and fall: low at the smooth initial condition, rising as gradients drive structure, falling again as gradients exhaust toward equilibrium. This macro-trajectory is a descriptive observation about an aggregate, not a process any system optimizes and not a goal, in the same sense that total biomass is a sum and not an objective. The selection dynamics operate on individual persisting systems; the aggregate trajectory is their statistical shadow. This is the same point as C1's firewall: there is no global optimizer and, lacking a Markov blanket at the maximal scale, no global agent for one to belong to.

The framework makes no claims outside this interval. The persistence selection principle is contingent on the physics we observe, not logically necessary in all possible worlds. This is a deliberate limitation.

12. Conclusion

From four axioms and four background premises, the framework traces a layered architecture that dissolves the is-ought problem by treating "ought" as a domain-bound operator, naturalises ethics as the modeling activity of bounded agents, grounds the functional necessity of valence, and offers a thermodynamic hypothesis about the collapse of information-suppressing regimes. This revision adds real physical grounding for the central cost claim through the thermodynamics of prediction, a Price-equation statement of selection, a description-length bound on deception, and an ordinal generalisation of the ought-relation, while marking clearly what remains qualitative, analogical, or open. Every result is traced and labelled by kind so its weight is visible. The dissolution rests on a single semantic premise, defended rather than assumed, and named as the framework's principal exposure.

The framework is not a completed metascience. It is a proposal for one, valuable for its parsimony, its cross-disciplinary reach, and its testability. Whether it survives empirical scrutiny and peer debate is a question for the territory to decide.

The ought was always an is. We were just using the wrong grammar. The work of this revision was, in part, to keep that line from claiming more than the grammar can carry.


Appendix A: Glossary

Each entry ends with a pointer to where the term is grounded. Axioms A0 to A3, premises BP0 to BP3, results T1 to T23 and C1, sections as numbered.

Antifragility
The capacity of a system to improve its predictive accuracy from exposure to variance, rather than merely withstanding it; a necessary condition of long-horizon persistence in high-variance environments. (T17.)
Axiom
A premise treated as bedrock and not derived. There are four: A0 to A3. (§2.1.)
Background empirical premise
A fact about our universe the derivations rely on but cannot produce from the axioms alone, stated to prevent smuggling. There are four: BP0 to BP3. (§2.2.)
Consistency Tax
The thermodynamic overhead of any mismatch between model and territory, from error or deception, paid even when latent and spiking when corrected; grounded physically in the dissipation of nonpredictive retained information. (T9, §5.1.)
Coupling density (K)
A construct measuring information-theoretic dependence between a higher-level system and the systems nested within it; tight coupling permits override. (T16, §5.6.)
Dark-room limit
The scope boundary of selection: with no variance and unlimited resources a false model is never disconfirmed and pays no tax, so accuracy is not selected for. (Scope note to T5; §6.2.)
Domain-bound operator
An operator with content only within a domain, type-errored outside it. The central meta-ethical claim is that "ought" is domain-bound. (T11, §6.1, §7.)
Epistemic overshoot
Accumulation of unmeasured divergence in a collective whose telemetry is suppressed, until a perturbation forces the hidden divergence open and the system collapses non-linearly. (T21.)
Epistemic Profit
The free energy recovered when a system reduces divergence and stops paying the tax on it. (T10.)
Epistemic selection
The process by which entropy selects for accurate models: cheaper, more accurate maps tend, statistically, to outlast expensive, inaccurate ones. (T5.)
Ghost
A grammatically well-formed but physically contentless question, whose central operator has no domain; distinct from a false claim, which has a domain and is wrong within it. (§7.)
Landauer ceiling
The loose in-principle upper bound on adaptation rate, available power over the Landauer cost of a bit. (T18, §5.4.)
Map-territory divergence
The mismatch between a system's model and its environment; generates prediction errors that dissipate free energy. (T4, §5.1.)
Markov blanket
A statistical boundary separating internal from external states; nests across scales; absent at the maximal scale, which has no outside. (T15.)
Nostalgia
Nonpredictive retained information: memory of the past that does not predict the future. Its physical cost is dissipation. (§5.1, after Still et al. 2012.)
Past Hypothesis
The premise that the universe began in an extraordinarily low-entropy macrostate; every gradient is that condition still discharging. (BP0.)
Persistence selection
The framework's engine; in an ecology competing for finite free energy, lower-divergence systems waste less and, on average over many cycles, outlast higher-divergence ones. A statistical tendency, not a law. (T5, §5.3.)
Persistence Tautology
The first axiom: a system failing its persistence conditions ceases to exist as an observable. Asserts nothing about value. (A0.)
Perturbationally robust compression fidelity
The framework's definition of truth: minimal-loss compression of environmental structure adequate for persistence across expanding perturbations. A definitional choice. (T7, §8 objection 13.)
Predictive Calm
The conjectured felt correlate of Epistemic Profit; marked an interpretive bridge, not a result. (T10.)
Schema and content
The ought-schema is universal (do what sustains your persistence over your horizon); the ought-content is agent-relative. A universal schema with agent-relative content yields no universal prescription. (§6.1.)
Semantic premise
The single premise the dissolution rests on: "ought" is agent-relative, its content fixed by the indexed agent's persistence conditions, type-errored without one. The framework's principal exposure. (§6.1, §10.)
Sustained negentropy capacity
A system's capacity to maintain its low-entropy structure over a horizon; the agent's ought-fact is the configuration maximizing it, determinate even when intractable. (T12.)
Thermodynamic Gap
The finite interval between the low-entropy beginning and the dark-energy approach to equilibrium; the framework applies only inside it. (§11.)

Appendix B: Personal Extension Beyond the Framework

This appendix is not part of the metaphysically neutral framework above. It records the author's personal metaphysical commitments and a more sweeping monist reading. None of it is load-bearing for any numbered result, and the formal body (especially §6.4) deliberately does not assert it. It is included for completeness and honesty about where the author's view exceeds what the framework establishes. A reader can reject all of Appendix B and lose nothing in §§1 to 11.

B.1 Thermodynamic Monism (a personal statement)

The universe is a physical system; there is nothing else. The fundamental fact is the entropy gradient, and the arrow of time is its direction. Every structure, living or not, is a transient pattern sustained by its discharge. Information is physical. Logic and mathematics are the structural invariants compressed from a physical world, physically enforced because any persistent system that violates them incurs prediction error and ceases to exist. Meaning is a physical signal, the valence gradient of an agent's long-horizon persistence model; no cosmic agent exists, so there is no cosmic meaning, only agent-relative meaning, which is a physical fact. On this personal view, ethics is engineering and deception and censorship are thermodynamic pathologies.

Where this exceeds the framework. The monist reading takes positions the formal body brackets. Most importantly, on consciousness it says: experience is the compressed telemetry of model-territory divergence in a complex control system, the felt quality is the dashboard, and the "hard problem" is a category error. The formal framework (§6.4) does not say this. It gives a functional account of valence and explicitly leaves the phenomenal residue unexplained and out of scope. The author's personal position leans toward a graded, proto-experiential view on which experience is real and scales with organizational complexity. That is a metaphysical bet, not a TR result, and TR remains neutral. The monist's slogan "everything reduces to thermodynamics, no exceptions" is a research aspiration in this appendix, not a theorem in the paper.

B.2 The undivided mind (a personal ideal)

A speculative endpoint: a cognitive architecture that has eliminated categorical partitions between "physics," "ethics," "mind," and "meaning," treating them as scale-free manifestations of one substrate, and so asymptotically minimizing the Consistency Tax of inhibiting one internal model in another's domain. The eleven objections this ideal invites are answered in §8 above, where they belong, and at the framework's actual strength rather than the ideal's rhetorical one. The honest status of the ideal is exactly objection 3's: human minds are evolutionary spaghetti code, the undivided mind is an asymptotic limit selection may favor, not a current condition or a guaranteed destination.

References

Albert, D. (2000). Time and Chance. Harvard University Press.

Anscombe, G. E. M. (1958). Modern moral philosophy. Philosophy 33(124), 1 to 19.

Ashby, W. R. (1956). An Introduction to Cybernetics. Chapman and Hall.

Ayer, A. J. (1936). Language, Truth and Logic. Gollancz.

Bennett, C. H. (1982). The thermodynamics of computation, a review. International Journal of Theoretical Physics 21(12), 905 to 940.

Bérut, A., et al. (2012). Experimental verification of Landauer's principle linking information and thermodynamics. Nature 483, 187 to 189.

Boyd, R. (1988). How to be a moral realist. In G. Sayre-McCord (ed.), Essays on Moral Realism. Cornell University Press.

Brink, D. (1989). Moral Realism and the Foundations of Ethics. Cambridge University Press.

Carroll, S. (2010). From Eternity to Here. Dutton.

Chalmers, D. (1995). Facing up to the problem of consciousness. Journal of Consciousness Studies 2(3), 200 to 219.

England, J. L. (2013). Statistical physics of self-replication. Journal of Chemical Physics 139(12), 121923.

Foot, P. (2001). Natural Goodness. Oxford University Press.

Friston, K. (2010). The free-energy principle: a unified brain theory? Nature Reviews Neuroscience 11(2), 127 to 138.

Gibbons, G. W. and Hawking, S. W. (1977). Cosmological event horizons, thermodynamics, and particle creation. Physical Review D 15(10), 2738 to 2751.

Horgan, T. and Timmons, M. (1991). New wave moral realism meets Moral Twin Earth. Journal of Philosophical Research 16, 447 to 465.

Hume, D. (1739). A Treatise of Human Nature. Book III, Part I, §I.

Jacobson, T. (1995). Thermodynamics of spacetime: the Einstein equation of state. Physical Review Letters 75(7), 1260 to 1263.

Korsgaard, C. (1996). The Sources of Normativity. Cambridge University Press.

Ladyman, J. and Robertson, K. (2013). Landauer defended: reply to Norton. Studies in History and Philosophy of Modern Physics 44(3), 263 to 271.

Landauer, R. (1961). Irreversibility and heat generation in the computing process. IBM Journal of Research and Development 5(3), 183 to 191.

Mackie, J. L. (1977). Ethics: Inventing Right and Wrong. Penguin.

Moore, G. E. (1903). Principia Ethica. Cambridge University Press.

Myrvold, W. C. (2024). Shakin' all over: proving Landauer's principle without neglect of fluctuations. British Journal for the Philosophy of Science 75(3), 587 to 616.

Norton, J. D. (2011). Waiting for Landauer. Studies in History and Philosophy of Modern Physics 42(3), 184 to 198.

Parrondo, J. M. R., Horowitz, J. M. and Sagawa, T. (2015). Thermodynamics of information. Nature Physics 11, 131 to 139.

Penrose, R. (1979). Singularities and time-asymmetry. In S. W. Hawking and W. Israel (eds.), General Relativity: An Einstein Centenary Survey. Cambridge University Press.

Price, H. (1996). Time's Arrow and Archimedes' Point. Oxford University Press.

Railton, P. (1986). Moral realism. Philosophical Review 95(2), 163 to 207.

Ruse, M. (1986). Taking Darwin Seriously. Blackwell.

Schrödinger, E. (1944). What Is Life? Cambridge University Press.

Shannon, C. (1948). A mathematical theory of communication. Bell System Technical Journal 27, 379 to 423, 623 to 656.

Simon, H. (1956). Rational choice and the structure of the environment. Psychological Review 63(2), 129 to 138.

Still, S., Sivak, D. A., Bell, A. J. and Crooks, G. E. (2012). Thermodynamics of prediction. Physical Review Letters 109, 120604.

Stevenson, C. L. (1944). Ethics and Language. Yale University Press.

Sturgeon, N. (1985). Moral explanations. In D. Copp and D. Zimmerman (eds.), Morality, Reason and Truth. Rowman and Littlefield.

Verlinde, E. (2011). On the origin of gravity and the laws of Newton. Journal of High Energy Physics 2011(4), 29.

Vrij, A., Fisher, R., Mann, S. and Leal, S. (2003 to 2012). Cognitive-load approach to deception detection (multiple studies).

Walczyk, J., Roper, K., Seemann, E. and Humphrey, A. (2003 to 2014). Reaction-time asymmetries in deception (multiple studies).

Wilson, E. O. (1975). Sociobiology: The New Synthesis. Harvard University Press.

This document consolidates the Thermodynamic Realism research program as of rev. 4. Correspondence: Andraž Đurič, Slovenia. Formal collaboration as credited in the byline. Writings licensed CC BY 4.0; any code MIT.

Comments

Popular posts from this blog

What You Actually Are

The Shape of the Disagreement: Why the Sex and Gender Debate Has the Structure It Has

Value as Persistence: Agent-relative oughts under coupling, nesting, uncertainty, and open-ended time