MRH: Markov Relevancy Horizon
Speculative — Theoretical FrameworkThe Markov Relevancy Horizon (MRH) is a term proposed by Dennis Palatov, inspired by Markov blankets — a concept from probabilistic graphical models where a node's Markov blanket is the minimal set of other nodes that makes it conditionally independent of everything else. The blanket is the boundary: everything inside it is relevant to the node, everything outside is statistically screened off.
MRH extends this idea from a static graph property to a dynamical, scale-dependent boundary. Where a Markov blanket asks “what nodes shield this node?”, the MRH asks “at what horizon do correlations between systems decay below the noise floor?” — making the boundary itself a function of scale, density, and context.
In one line: the MRH is the bubble of neighbors that matter. A single atom does not “feel” the Andromeda galaxy — it responds to what is immediately around it, and the MRH is the name for how far “immediately around it” reaches. Everything past that edge is supposed to be irrelevant to how the system evolves.
The formal version says the same thing with the load-bearing words made explicit (degrees of freedom = the independent numbers you would need to write down to specify the system's state):
The minimal set of interacting degrees of freedom whose state transitions materially influence the coherence evolution of a defined system.
Operational Criteria
An MRH is not just a vague boundary — it must satisfy two testable conditions:
Predictive Sufficiency
Removing any element inside the MRH degrades coherence prediction. Everything inside is load-bearing.
Predictive Closure
Adding elements outside the MRH does not materially improve prediction. Everything outside is irrelevant. If it does improve prediction, the MRH was incorrectly specified.
Does the framework respect its own horizon? Not in the galaxy sector.
Added 2026-09-07, raised by a visitor researcher persona. This is a data-free self-consistency failure: it needs no SPARC fit, no BTFR slope and no ΔBIC, and it would stand even if every empirical test on the Tier 1 ledger had passed.
The galaxy sector advertises C as a function of local density — that is the whole reason it is claimed to be MRH-respecting, and the reason the local-density no-go bites. But look at what the mechanism actually evaluates. Two of its three inputs are imported from outside any local neighborhood:
- ρcrit = A·Vflat². Vflat is the asymptotic rotation speed — defined in the r → ∞ limit, and by the BTFR fixed by the galaxy's total baryonic mass. So C at radius r is not C(ρ(r)); it is C(ρ(r), Mtotal). A threshold on a local field has been keyed to a global label, which is exactly what Predictive Closure above forbids.
- Bmax = 1/Ωm = 3.17. The per-galaxy boost ceiling is set by a cosmological parameter. A galaxy's relevancy horizon does not contain Ωm. (Flagged independently by a visitor graduate-physics persona the same day.) Note this objection is stronger than the empirical one it accompanies, because it does not depend on the value: TEST-10 needs B ≥ 13.7 and no cosmic ratio supplies it, but even a ratio that did supply it would still be an import.
The mechanical consequence. In the small-x regime SPARC actually samples (median x ≈ 7×10−5), C ≈ γ x = γρ/(A Vflat²) — γ and A enter only as the ratio γ/A. That is one free number per galaxy, not two, which is what the Fisher correlation ρ(lnγ, lnA) = +1.000000 on For Researchers is reporting. The unidentifiability artifact and this locality violation are the same defect seen from two directions: the model is fit per-galaxy against a global label, so the local parameters never get separately measured.
The shape of the whole problem. The galaxy sector needs one non-local variable it does not have (gbar, the enclosed-mass acceleration the RAR is organized by — that is the local-density no-go), and it silently uses two non-local variables it should not have (Vflat, Ωm). It is non-local by construction in the sector where it claims locality, and non-local in the wrong variables. This is registered here rather than hidden because it is cheap, structural, and cuts against the framework.
A fourth, quantitative instance — and a correction to yesterday's framing (explorer 2026-09-07, added 2026-09-08). For a globular cluster, ρcrit = A·V² has no obvious V: read with the host galaxy's 220 km/s the knee sits at ~1 pc (the whole cluster saturated at the floor); read with the cluster's own dispersion it sits at ~2.5 half-mass radii; the measured velocity-blind knee puts it at 6–8. That is a factor 10⁴ in ρcrit and the whole verdict, decided by which non-local label you import. Yesterday this page's topic file called the cluster prediction “unformulable.” It is not — it is formulable under every placement, and under-determined, which is worse in a specific way: the internal-locality violation does not make the prediction undefined, it makes it depend on a choice the framework never states. The executed result is on the Honest Assessment.
MRH and Presence
Presence (ρ) — the compatible structural elements that drive coherence — is defined relative to an MRH. Change the MRH boundary, and presence changes. This means coherence is always context-dependent: what counts as “present” depends on which system you're examining and at what scale.
See: Coherence Function for how presence feeds into C(ρ).
How It Works
Every system maintains correlations with nearby systems. As distance (spatial or temporal) increases, these correlations weaken. The MRH is where they become negligible.
- Inside the MRH: Systems are correlated, can influence each other, quantum effects persist
- At the MRH: Correlations = noise. This IS the boundary.
- Beyond the MRH: Systems are effectively independent, classical behavior dominates
MRH and Quantum Measurement
This is Synchronism's most provocative claim about quantum mechanics:
Wave function “collapse” = crossing the MRH.
When a quantum system interacts with a macroscopic apparatus, the correlations between the system and its environment rapidly exceed the MRH. What we call “measurement” is this boundary crossing. No observer needed. No consciousness required. Just decoherence at the relevancy horizon.
Untested — 6 Testable ProtocolsFull treatment: Measurement Without Observers →
MRH at Cosmic Scales
The same concept applies to cosmology. Cosmic horizons (particle horizon, event horizon) can be reinterpreted as MRH boundaries at cosmological scales. Beyond the MRH, correlations from the early universe have decayed. What we call the “observable universe” is the region within our MRH.
Cosmic Horizons as MRH Phenomena →
MRH in Statistical Mechanics
In statistical mechanics, the correlation length ξ measures how far correlations extend. At a phase transition, ξ diverges. The MRH is the dynamical version of this: where correlations become irrelevant not just in space but in the full phase space of the system.
Open Gap: Formal Mathematical Definition
The name Markov Relevancy Horizon promises a specific mathematical structure: a conditional independence statement. In standard graphical models, a Markov boundary for node X satisfies:
The current MRH framework specifies operational criteria (predictive sufficiency, predictive closure) but not a probability distribution, integration measure, graphical model, or explicit mapping from coherence dynamics to the conditional-independence condition.
Two paths forward: (1) Provide the integration measure that defines “inside MRH” vs. “outside MRH” and write the explicit conditional-independence statement with a probability distribution. (2) Or rename toRelevance Horizon — acknowledging the intuition is Markov-motivated but the formalism is not yet Markov. The “Markov” label sets an expectation that is not currently met.
What's Untested
The MRH as a replacement for “wave function collapse” is the central untested prediction. Six experimental protocols have been designed (Sessions #368-370) but none have been run. The theory predicts specific decoherence patterns at the MRH boundary that should be measurable.