Rival Architectures

Simulations, the CTMU, Gödel, and God

Part IV has proposed an ambitious ontology: Chaos as a possibility-space representation, coherence filters as selectors and interpreters, and constructors as a bridge toward dynamics. It does not yet explain realization, the selection of laws and measures, or consciousness. Comparisons with other architectures should therefore expose shared burdens rather than presume that this proposal has already solved them.

Four rivals are worth engaging seriously: the simulation-hypothesis family, Christopher Langan’s Cognitive-Theoretic Model of the Universe, the theological argument that materialism cannot survive the beginning of time, and the claim that Gödel’s incompleteness theorems make the universe unsimulable in principle. Each gets the same treatment: the rival’s best case, stated as strongly as I can state it; the diagnosis of where it fails; and the accounting of what it gets right — because each of these rivals is reaching for something real, and a framework that wins by refusing to acknowledge that is not winning.

The Simulation Family

Few ideas grip both philosophers and technologists like the notion that we are living in a simulation. It is an intellectual kaleidoscope: twist it one way and you find Descartes’ evil demon; twist it another and you are inside a GPU-rendered Matrix. But “the simulation hypothesis” is not one hypothesis. It is a family, and its members make different claims.

The oldest ancestors are skeptical thought experiments. Descartes’ demon feeds us false experiences; Putnam’s brain in a vat receives artificial stimuli indistinguishable from a world. These are epistemic traps, not theories: they establish that perception cannot guarantee reality, and they offer nothing beyond the doubt itself. In the digital age the demon becomes a supercomputer. Bostrom’s trilemma — civilizations never reach the capacity for ancestor simulations, or reach it and abstain, or we are almost certainly inside one — reframes ancient skepticism as probability. Its weakness is that it assumes computability and sidesteps the regress: what runs the hardware? A third wing collapses physics into computation directly — digital physics in the tradition of Zuse, Fredkin, and Wolfram, where reality is a cellular automaton and the laws of nature are transition rules; or the quantum-rendering conceit, where indeterminacy is a game engine conserving processing power by rendering only what is observed. Empirically ambitious, but the substrate is still unexplained: what medium runs the automaton? The theological variants — God as programmer, the Gnostic demiurge running a flawed copy — reframe creation, fall, and deception as computational metaphor, and are exactly as testable as their originals.

The strongest member of the family is the newest: predictive processing. On accounts developed by Seth, Friston, and others, the brain does not passively receive the world; it predicts sensory input and corrects through error signals. Perception is therefore modeled as a constrained construction from sparse sampling plus inference, and dreaming shows that internally generated content can recruit much of the same machinery without ordinary sensory input. The brain also predicts sensory consequences of action and compares them with outcomes. These are active neuroscientific programs, not one settled theory of all cognition. They explain important functional and reportable features of perception; identifying that predictive machinery with phenomenal consciousness requires the further argument developed in Volume 3. In any case, an intracranial model says nothing by itself about whether the universe is simulated.

The diagnosis, then: every member of the family locates a filter between us and raw reality, and every member leaves the crucial term undefined. The skeptics have no mechanism beyond doubt. Bostrom has a probability and a regress. Digital physics has rules and no substrate. Theology has narrative and no test. Predictive processing has a mechanism — inside the skull, silent about everything outside it. And the bare sentence “we live in a simulation” is underdetermined until “we,” “live,” and “simulation” are bound — a statement that fails rather than a proposition, which is why arguments about it so rarely settle anything. Chalmers deflates the question from the other side: even if simulated, the simulation is our reality, so what exactly was being asked?

The Chaos proposal shares the family’s distinction between raw representation and interpreted world. But it does not yet provide an “actual mechanism”: a set containing program strings does not execute them, a semantic map does not physically realize its outputs, and an infinite sequence containing a brain description does not thereby instantiate a mind. It may avoid an external programmer only by taking mathematical realization as primitive. That trade should be compared with, not declared superior to, hardware and simulation accounts.

The Tautological Universe

Langan’s CTMU¹ deserves engagement precisely because it is motivated by the right intuition. Its core claim: reality is a Self-Configuring, Self-Processing Language (SCSPL), a closed structure in which syntax and semantics merge — the rules of the language are its laws, the states of the language are its contents, and the world is the unfolding of its own grammar. Telic recursion gives it direction: future global consistency feeds back into present states, a built-in teleology. Infocognitive monism closes the loop: information, cognition, and reality are one substance. The demand underneath all this is legitimate and it is mine too: a total ontology must be self-contained, generative, and coherent, with no external scaffolding left standing.

The diagnosis is that the CTMU meets this demand by announcement rather than construction. It is rhetorically sweeping and formally underdeveloped; its claims of tautological necessity collapse, on inspection, into slogans — reality is consistent because it must be, language is language because it is. Point by point, it is the mirror image of the Chaos account. Source of order: the CTMU is top-down — reality selects itself; Chaos is bottom-up — order emerges from noise. Mechanism: the CTMU has a syntax/semantics duality with no operational formalism; the Chaos account has constructors, in the Deutsch–Marletto sense, doing specifiable work. Randomness: the CTMU banishes it — everything is deterministic recursion, and novelty is pre-encoded; Chaos makes randomness primary and carves coherence out of it, so novelty is real filtration, not disguised replay. Agency: the CTMU smears it across the whole, a universal recursion in which the distinction between agent and system disappears; the physics of agency localizes it in self-modeling constructors biasing their own futures through branchspace. And the deepest failure is the claim to unconditional truth. All truth is conditional; the attempt to escape this by appeal to “tautology” is a hidden metaphysical leap, a grand identity claim — reality = cognition = information — where an argument should be.

The comparison identifies a shared demand for self-containment and a shared risk: redescribing self-reference can look like explaining realization. The Chaos account has clearer toy formalisms in this volume, but its filters and constructors do not yet constitute a working physical engine. A fair verdict is that both programs owe operational definitions, derivations, and discriminating predictions where they claim more than interpretation.

Before the Big Bang

The theological rival’s best case is not the design argument; it is the argument from the beginning of time, and its sharpest statement comes from Stephen Meyer in Return of the God Hypothesis:²

If sometime in the finite past, either the curvature of space reached an infinite and/or the radius and spatial volume of the universe collapsed to zero units, then at that point there would be no space and no place for matter and energy to reside. Consequently, the possibility of a materialistic explanation would also evaporate, since at that point neither material particles nor energy fields would exist. Indeed, since matter and energy cannot exist until space (and probably time) begins to exist, a materialistic explanation involving either material particles or energy fields — before space and time existed — makes no sense.

If sound, this is checkmate: materialism cannot explain its own opening move, and immaterial causation wins by default. Robin Hanson’s rebuttal is seven words: you assume material must occupy finite space. That one sentence is the whole diagnosis; the rest is unpacking.

Meyer’s argument relies on a container picture of matter and space that general relativity complicates. Spacetime geometry is dynamical and coupled to stress-energy, though general relativity also admits vacuum solutions, so empty spacetime is not ill-defined. If spacetime is emergent, asking where a more basic structure was “before” space may misuse spatial and temporal concepts.

Emergent spacetime is an active research direction, not an established consensus theory. Ordinary quantum field theory is formulated on spacetime backgrounds; quantum gravity programs explore ways geometry might emerge from more basic relational or quantum structure. Wheeler–DeWitt approaches, loop quantum gravity, causal sets, and holographic dualities illustrate possibilities without yet establishing one account of cosmological origins. Non-spatiotemporal structure can be meaningful in a model even if “matter before space” is a misleading phrase.

The question remains serious: a quantum vacuum on a spacetime background is not by itself an account of spacetime’s origin. The Chaos proposal offers a non-spatiotemporal representation, but it does not yet derive geometry, time, matter, or their empirical laws. It is one candidate framework, not a direct answer already supplied, and it does not refute theological grounding merely by offering different primitives.

Gödel in the Machine

The last rival attacks from the opposite flank: not that reality has a simulator, but that it provably cannot have one. The target is the Deutsch–Church–Turing (DCT) thesis — every finitely realizable physical system can be perfectly simulated by a universal computing device operating by finite means. This is not a claim about our tools; it is a claim about reality: physical law corresponds to algorithm. It underwrites digital physics, computational cosmology, and every strong simulation hypothesis, including — apparently — mine.

The Gödelian argument against it runs: if the universe were algorithmic, every physical truth would be derivable from its computational rules; Gödel showed that no consistent formal system is complete; therefore there exist physical truths no finite algorithm can compute; therefore the universe transcends computation. If this holds, the stakes are total. Digital physics is strictly false, not merely incomplete. No civilization, however advanced, could run a perfect simulation of us. No machine bound by algorithmic law could fully model reality — only approximate it.

The diagnosis turns on a distinction the argument blurs: epistemic versus ontological non-computability. Epistemic non-computability — we cannot compute or predict everything, for want of knowledge or resources — is obviously true and perfectly compatible with the DCT thesis. Ontological non-computability — some physical processes have no computable description even in principle — is what the argument needs, and here the Gödelian transposition is metaphor, not proof. Gödel’s theorems apply to symbolic systems that can encode arithmetic; nothing establishes that physical law is symbolically representable in the relevant way, and without that mapping the incompleteness machinery never engages. Meanwhile the DCT thesis was never a theorem to be toppled. It is a boundary condition — an assumption of closure that defines a cosmos renderable as finite information under computable rules. To refute it you must exhibit a physically real process that provably exceeds Turing computability — a natural hypercomputer — and no such demonstration exists.

The computability of reality may itself be conditional: holding locally, within domains of decohered structure — which is exactly why computational physics works — while failing globally, at the boundaries of emergence: quantum measurement, consciousness, cosmogenesis. And the Chaos account occupies that middle ground natively. Its substrate is measureless randomness — not the output of any algorithm, not a program running on cosmic hardware. Its constructors, the lawful machinery filtered out of that substrate, are precisely the computable part. The universe on this picture is a simulation engine embedded in a non-algorithmic substrate: computation does not exhaust ontology, and was never claimed to. If a natural hypercomputer is someday demonstrated, digital physics dies and the Chaos account absorbs the result without amendment. The Gödelian challenge is not whether reality is computable but where computation ceases to be an adequate model of it, and what lies beyond that line.

What a Rival Must Do

Four engagements reveal shared tests. A simulation account needs a precise substrate and observer measure. The CTMU and Chaos both need to turn self-description into explanatory or empirical work. Cosmological arguments must distinguish temporal beginnings from ontological dependence. Gödelian arguments must connect formal incompleteness to physical dynamics. The Chaos proposal faces those tests too: its mathematical primitives are assumptions, its realization bridge is open, and its quantities become computable only in specified models. An ontology earns confidence through clarity, derivation, and comparison with evidence—not breadth alone.