The direct URL moved. Let me find the current page on hedweb.com.
I found the key pages. Let me read the full essay sections.
I have the core text. Let me quickly verify a few details on Standish's and Schmidhuber's ensembles to make the comparison precise.
I have what I need. The question is deeply comparative, so let me load the rich-ui skill for a structured comparison alongside the analysis.
The comparison is qualitative, so a structured Markdown comparison plus analysis will be clearer than a chart. Let me verify canonical sources for Tegmark and Schmidhuber, and check whether Standish engages the zero ontology directly.
All five views are "plenitude" theories: they answer "why anything?" with "everything". The differences sit in what the ensemble is made of, why it exists at all, and whether the totality is required to be informationless. I'll work from the current hedweb essay sections, Standish's arXiv papers, Schmidhuber's "Algorithmic Theories of Everything", and the standard MUH and modal realism literature.
The five positions at a glance
What exists
One quantum-coherent superposition realizing all self-consistent possibilities
All mathematical structures (Level IV)
All formally describable program outputs (dovetailer)
The set of all possible descriptions
All concrete possible worlds
Why the ensemble exists
Necessary: zero information is the default condition; any restriction costs bits
Posited: mathematical existence is physical existence
Posited via algorithmic probability; "P is formally describable" is an assumed bias
Assumed pre-existence of all descriptions; agnostic whether that is all there is
Posited for modal-semantic utility (recombination)
Total info content
Exactly zero
Small but nonzero (algorithmic complexity of laws)
Small but nonzero (dovetailer program length)
Exactly zero as a specification set ("no bits are specified")
No claim; no totality that could cancel
Quantum commitment
Constitutive: superposition principle formalizes inexistence; no collapse
None required; Everett optional
None; universes are discrete computed histories
QM derived as most general linear theory satisfying anthropics, given computable observers
None; worlds are isolated, non-interfering
Consciousness
Intrinsic field-theoretic what-it's-like-ness; no digital sentience
Tegmark. The MUH identifies physical reality with mathematical structure and gets Level IV for free; Pearce shares the physics-mathematics identification, but the essay's decisive move is exactness: "even if the algorithmic information content of the multiverse as a whole is small, perhaps even close to zero [as proposed by Max Tegmark], it isn't exactly zero, as predicted by the zero ontology" (§7). Small-but-nonzero leaves a residue of brute contingency: why these axioms? The informationless criterion forbids Tegmark's CUH retreat to computable structures, which is precisely a bit-cost restriction motivated by Gödel worries (MUH overview). And where Tegmark leaves consciousness as self-aware substructure computations, the zero ontology's third equation, phenomenology = 0, answers the "fire in the equations" gap by making experience the intrinsic nature of the physical.
Schmidhuber. His ensemble presupposes a describable probability distribution over histories, a universal machine, and the dovetailer program: small but irreducibly nonzero information, plus a generator (Algorithmic Theories of Everything). The criterion removes both the generator and the bias: no machine choice, no formal-describability restriction. The deeper clash is the classical/quantum one. A classical "library of Babel" has zero information as a set, but withdrawing any single definite book creates information; only coherent superposition keeps the total at zero, which is why the essay's Library of Babel must be tightened into a quantum one (§1, and your recent IAI piece on amplitudes summing to zero: IAI).
Standish. The nearest neighbour: he independently formulated the "zero information principle" ("the information content of this complete set is precisely zero, as no bits are specified. It is ontologically equivalent to Nothing"), derived Occam's razor from inhabiting a least-information anthropically viable world, and remained agnostic about whether the description ensemble exhausts reality (Why Occam's Razor; Theory of Nothing). The zero ontology is the stronger, less agnostic reading: the criterion is not a bookkeeping fact about a specification set but a constraint on what is physically real, which is what generates the Everett commitment and the interference structure. Note also the direction-of-explanation difference: Standish derives QM from anthropics plus computable consciousness; the zero ontology treats the superposition principle as the formalization of inexistence itself, prior to any observer.
Lewis. Modal realism posits the plurality for its theoretical utility in analysing modality, guarantees plenitude via recombination, and treats actuality indexically; it offers no cancellation structure and no totality that could equal zero (On the Plurality of Worlds, 1986). The zero ontology shares indexical actuality (the doppelgänger sees the dead cat) and plenitude ("all self-consistent alternative possibilities"), but converts brute plurality into necessity, and monism replaces isolation: the branches are relative states of one superposition that genuinely interferes before decoherence, whereas Lewis's worlds cannot. The trade: Lewis buys an analysis of counterfactuals and properties; the zero ontology buys an answer to the existence question, not a modal semantics.
What the informationless criterion adds
Necessity instead of posit. Zero information is the unique condition demanding no explanation; every restriction is a bit-cost that demands one. So the unrestricted totality, the multiverse, "inherits such default status": existence becomes obligatory rather than gratuitous (§7). None of Lewis, Tegmark, or Schmidhuber has this argument form; only Standish approaches it.
Exactness. The gap between "simple" and "zero" is the entire explanatory gap: it forecloses residual arbitrary bits (choice of axioms, machine, or restriction) and forces maximal variation across vacua, geometries, constants, and bylaws.
Physical embodiment. The criterion lives inside our best theory: superposition, decoherence, information conservation. The classical ensembles of Schmidhuber and Standish satisfy it only as specification sets; the quantum Library satisfies it dynamically.
Refutability. Objective collapse (GRW, Penrose) would create information ex nihilo and falsify the conjecture; traditional collapse readings are explicitly inconsistent with it (§7). Lewis and Tegmark's views are compatible with collapse.
Phenomenology inside the zero. Physics = mathematics = phenomenology = 0 is the equation none of the other four can state. The criterion is thereby applied to the hard problem: the world's intrinsic nature is what-it's-like-ness whose total cancels.
What it removes
Arbitrariness: no creator, no random sampling from the ensemble, no privileged machine or language, no Gödel-motivated cut-down of the ensemble.
Absolute actuality and real collapse: actuality becomes indexical; information is never created, so "where did information come from?" is dissolved rather than answered (§1).
Computationalism about mind: the most striking removal, since the criterion sounds information-theoretic. Schmidhuber's and Tegmark's self-aware substructures and Standish's computable observation map all drop out; experience becomes the non-computational intrinsic nature of the zero's physical embodiment.
Where the criterion is still doing unfinished work
Candidly, three pressure points remain. First, the measure problem: Born weights are extra structure, and a non-uniform norm is arguably nonzero information unless the amplitude formalism is uniquely fixed as the formalization of inexistence, which is the boldest step in the whole construction. Second, totality definition: "all possible (by)laws" presupposes a space of possibility, and specifying that space is a bit-cost unless QM is alogical necessity. Third, Witherall's objection that the zero "helps itself to both totalities" and that the phenomenological cancellation is under-defended (the colours-summing-to-white argument doesn't yet give an identity) (Witherall). The Standish contrast sharpens this: his zero information is a property of a description set, the zero ontology's is a property of reality, and the bridge between them is exactly what the quantum argument is meant to supply.
Nov 8, 8:40am
This is a shared conversation and may include unverified content that doesn't reflect Mistral AI's views. Login to resume chatting.
This is a shared conversation and may include unverified content that doesn't reflect Mistral AI's views. Login to resume chatting.