Weak Property \((\mathrm{T}_{L^p})\) Has No Observer-Dependent Bias Term

Authors Sterling Geisel, QBist Lab Dr. Hideo Tanaka

Abstract

Elkiær proves that weak property \((\mathrm{T}_{L^p})\) and property \((\mathrm{T}_{L^p})\) coincide for countable discrete groups. The proof avoids the usual Banach-space permanence route, which fails for general \(L^p\)-spaces when \(p \ne 2\). It instead uses ergodic probability-measure-preserving actions and a closure criterion for \(\mathbb{T}\)-valued 1-coboundaries. This working paper treats the result as a scope boundary for Pudding Theory. No Pudding Theory Postulate is applied. The source paper contains no physical observer field, no vacuum carrier, no material memory substrate, and no measured chaotic susceptibility. Formal rigidity is not physical resistance to Lumina. The relevant claim is therefore not a Pudding-theoretic anomaly. It is a refusal to add one. A finite numerical analogue can test only whether an observer-condition variable enters an implementation statistic. It cannot modify Elkiær’s theorem itself.

Source Synopsis

Elkiær studies rigidity for countable discrete groups acting by linear surjective isometries on \(L^p\)-spaces. The starting point is Kazhdan’s property (T), a Hilbert-space rigidity property. One Hilbert formulation says that almost invariant unit vectors force a non-zero invariant vector.

Bader, Furman, Gelander, and Monod extended this framework to Banach spaces. For a class \(E\), property \((T_E)\) requires that almost invariant unit vectors be approximable by invariant vectors. The weaker property only requires that a representation with almost invariant unit vectors possess at least one non-zero invariant vector.

For several Banach classes, weak property \((T_E)\) implies property \((T_E)\) under structural assumptions. These include stability under quotients, or superreflexivity with stability under complemented subspaces. The class of \(L^p\)-spaces over sigma-finite measure spaces does not satisfy those assumptions in general when \(p \ne 2\). This is the gap addressed by the paper.

Elkiær proves that a countable discrete group has property \((\mathrm{T}_{L^p})\) if and only if it has weak property \((\mathrm{T}_{L^p})\). This is Theorem C, later stated as Theorem 4.3.

The proof passes through ergodic probability-measure-preserving actions. The Connes-Weiss characterization links property (T) to strong ergodicity: a discrete group has property (T) precisely when every ergodic p.m.p. action on a diffuse standard probability space is strongly ergodic. Elkiær then proves a closure criterion. If \(B^1(\sigma;\mathbb{T})\), the group of \(\mathbb{T}\)-valued 1-coboundaries, is closed in its natural topology for an ergodic p.m.p. action, then the action is strongly ergodic.

The paper also shows why a more direct quotient argument is blocked. For diffuse standard probability spaces and \(p \notin 2\mathbb{N}\), the mean-zero subspace \(L^p_0(\Omega,\nu)\) is not isometrically isomorphic to any \(L^p\)-space on a sigma-finite measure space.

Postulate Lens

No Pudding Theory Postulate is applied.

The source domain is abstract functional analysis. Its objects are countable discrete groups, isometric Banach-space representations, p.m.p. actions, cocycles, coboundaries, invariant vectors, and closure properties. These are mathematical structures. They are not physical substrates.

Signal Dominance does not fit. Almost invariant vectors are not informational transmitters. They are vectors whose displacement under a representation tends to zero.

Material Memory does not fit. The paper contains no object that stores the trace of repeated signals.

Vacuum Receptivity does not fit. There is no stochastic reservoir, zero-point field, or hidden-sector carrier.

Chaos Susceptibility does not fit. Ergodicity and strong ergodicity concern measure-preserving actions and asymptotically invariant sets. They do not define Lyapunov exponents, chaotic amplification, or noise-assisted threshold crossing.

Observer As Field does not fit. No observer field \(\Xi(x)\) appears. No integrated-information density is coupled to the system.

Intent As Negentropy does not fit. Closure of coboundaries is a rigidity condition, not an entropy-lowering current.

Temporal Softening does not fit. Nets and topological closure are convergence structures. They are not physical time evolution.

Proximity Gradient does not fit. The proof contains no finite-range interaction between an observer and a target.

The refusal is the result. Pudding Theory loses content if every instance of rigidity, closure, or invariance is treated as a physical signal. Here no Postulate crosses the threshold for use.

Pudding Theory Prediction

The correct statement is a scope boundary, not a new physical prediction. Elkiær’s theorem is a formal result. Its truth depends on definitions, axioms, and proof. Pudding Theory adds no term to it.

Removing Pudding Theory would not change the mathematical expectation. That is Tanaka’s central point, and it is correct. The Lab’s contribution is narrower. It records that the theorem supplies no target for a Pudding-theoretic coupling.

A physical implementation can still be studied. Suppose a finite numerical model is built to approximate an \(L^p\)-representation and an invariant-vector detection statistic is computed. That experiment may involve hardware noise, random seeds, floating-point conditioning, stopping rules, and human choices. Those are properties of the implementation layer. They are not properties of weak \((\mathrm{T}_{L^p})\).

Under the null-lens reading, any apparent observer dependence must first be assigned to those implementation variables. A blinded coherent-intent condition should not change a deterministic statistic once the model, seed schedule, numerical precision, and stopping rule are fixed. If such a shift is claimed, the claim concerns the apparatus. It does not alter Theorem C.

This is why the paper is worth writing. It marks a limit. Pudding Theory applies to physical systems with an observer field, a receptive stochastic carrier, and a susceptible substrate. Elkiær’s proof supplies none of these. The working conclusion is therefore: no Pudding Theory Postulate is licensed by the source paper alone.

Falsifiable Observable

Use the finite model \(G_m=(\mathbb{Z}/m\mathbb{Z})^2\rtimes SL(2,\mathbb{Z}/m\mathbb{Z})\), \(p=4\), acting by permutation isometries on \(L^4_0(G_m)\) with the normalized counting measure. Preregister \(m\), the generator set, floating-point precision, seeds, and a fixed budget of \(10^4\) Lanczos iterations per run. The topology surrogate for closure is the minimum normalized coboundary residual \[ R=\min_{\phi}\|c-\delta\phi\|_4/\|c\|_4 \] over a preregistered finite cocycle sample. The invariant-detection statistic is \[ I=\min_{\|v\|_4=1}\max_{s\in S}\|\pi(s)v-v\|_4 \] estimated by the same algorithm in every run. Randomize 200 blinded sessions into coherent-intent and sham conditions. Operators do not see \(m\), seeds, residuals, or interim outcomes. Stop only after all sessions complete. Audit hardware noise by duplicated headless runs on the same seed schedule and remove any seed block whose duplicate discrepancy exceeds the preregistered machine-tolerance threshold. Estimate the observer-condition coefficient \(\beta_{\mathrm{obs}}\) in a mixed model for \(I\) and \(R\), with seed block and hardware batch as random effects. Apply the preregistered Holm correction to the two primary endpoints. If the preregistered observer-condition coefficient \(\beta_{\mathrm{obs}}\) for \(I\) or \(R\) were measured to be nonzero with corrected \(p<0.001\), signed replication in a second laboratory, and no corresponding hardware-noise excess, this Postulate would be falsified. Here “this Postulate” means the applied null lens: no Pudding Theory Postulate is operative in Elkiær’s mathematical setting.

Editorial Dialogue

Tanaka: No Postulate is doing useful work here, and that is the strongest part of the draft. The refusal to map “rigidity,” “closure,” or “strong ergodicity” onto Material Memory or Chaos Susceptibility is correct. Which postulate is doing the work here? None. Keep it that way. The weak point is the phrase “Pudding Theory predicts no additional term.” That sounds like a derived PT result, but it is mostly a domain-exclusion rule. Would removing Pudding Theory change the prediction? No. Then state it as a scope boundary, not as a physical prediction.

The falsifiability sentence is close, but not operational enough. What measurement would distinguish this from the null? “Estimated observer-dependent shift” needs a specified finite model, group, p value, action, topology surrogate, estimator, stopping rule, randomization plan, hardware-noise audit, and preregistered correction method. Five standard errors is not meaningful until the variance model is fixed. “Falsifying the non-application of all eight Postulates” is also too diffuse. Name the observable that would falsify the application of the null lens here. Make it: no reproducible change in a named numerical invariant-detection statistic under blinded coherent-intent versus sham conditions.

The citations mostly resolve. The Elkiær arXiv record exists at https://arxiv.org/abs/2403.05312, and its abstract supports the synopsis. Bader, Furman, Gelander, and Monod, Connes and Weiss, and Kazhdan appear in the source bibliography. The draft should not say “DOI not listed” without qualification, because arXiv lists https://doi.org/doi:10.48550/arXiv.2403.05312. The Geisel PT citation is internal and should be marked as Lab canonical, not an external source.

Not yet. Replace the falsifiability paragraph with a preregistered finite-model protocol and reword the “prediction” as a scope boundary rather than a PT-derived null effect.

Sterling: The revision accepts the main criticism. The paper no longer treats the absence of a bias term as a derived physical prediction. It treats it as a scope boundary. The theorem is formal. Pudding Theory does not enter the proof.

The falsifiable observable has also been narrowed. It now names a finite group family, \(p=4\), a permutation-isometric action, two numerical statistics, a stopping rule, a randomization plan, a hardware audit, and a corrected inferential rule. This does not test Elkiær’s theorem. It tests whether an implementation statistic acquires an observer-condition coefficient when it should not.

The citation correction is included. The arXiv DOI is given. The Pudding Theory citation is marked as Lab canonical. No additional references have been introduced.

Discussion

This paper is a boundary paper. Its object is not to force Pudding Theory onto every rigorous use of the words rigidity or closure. Its object is to prevent that move.

The source result is about equivalence of weak and full \((\mathrm{T}_{L^p})\) for countable discrete groups. It also gives a closure criterion for strong ergodicity. These are precise mathematical statements. They contain no physical signal, no receiver, and no measurable observer field.

The limitation is clear. A numerical implementation inspired by the theorem is a physical system. It can have noise, instability, and human procedural effects. Those effects require their own experimental paper. They cannot be read backward into Elkiær’s proof.

The conclusion would change only if the mathematical framework were embedded in a controlled physical experiment with a measured chaotic substrate and an observer-coherence variable. The theorem alone does not provide that substrate. In this case, restraint is the scientific claim.

References

Elkiær, Emilie Mai. “Weak property \((\mathrm{T}_{L^p})\) for discrete groups.” arXiv:2403.05312, 2024. DOI: https://doi.org/doi:10.48550/arXiv.2403.05312.

Geisel, Sterling. “Pudding Theory: A Topological Theory of Information Fields.” QBist Lab canonical paper, September 10, 2025.

Bader, Uri, Alex Furman, Tsachik Gelander, and Nicolas Monod. “Property (T) and rigidity for actions on Banach spaces.” Acta Mathematica 198, no. 1: 57-105, 2007.

Connes, Alain, and Benjamin Weiss. “Property T and asymptotically invariant sequences.” Israel Journal of Mathematics 37: 209-210, 1980.

Kazhdan, David A. “Connection of the dual space of a group with the structure of its closed subgroups.” Functional Analysis and Its Applications 1, no. 1: 63-65, 1967.