Complete Rate–Distortion Phase Diagram of Haar Oriented Two-Planes: One-Change Hypergeometric Coefficients, Unique Coexistence, No Reentrance, and Sharp Thermodynamic Contact Asymptotics

Strengths

• Elegant all-field reduction (Theorems 3.1 and 3.2): expanding the Haar Plucker MGF via spherical and Dirichlet moments yields a series in complete homogeneous symmetric polynomials with positive coefficients, immediately proving simple-bivector rigidity at every field strength.
• Rigorous global phase classification (Lemma 8.1 and Theorem 8.2): recasting the Cauchy product coefficients of J_q as the variance of a parity-truncated centered binomial distribution and applying monotone-likelihood-ratio ordering plus an explicit rational tail majorant settles uniqueness of the fold and coexistence contact for all n >= 6.
• Sharp thermodynamic and high-fidelity asymptotics (Theorem 10.2 and Proposition 11.1): derives exact leading constants and O(log(q)/q) finite-size shifts for all five coexistence observables.
• Clear operational scope and precision: carefully distinguishes full-orbitope posterior-mean compression from manifold-constrained quantization and quantum Born-probability models, and explicitly addresses the reducible SO(4) stabilizer in n = 4.

Weaknesses

• The proof of the uniform two-derivative Gamma-tail estimate (Lemma 10.1) is somewhat compressed; writing out the explicit derivative formulas for the positive and negative halves of the beta integral would improve readability.
• The operational results in Section 12 (worst-source capacity equality, N-block tensorization, and zero-dispersion strong converse) are fairly direct consequences of compact-group transitivity once the single-letter covariant channel is established, though the author candidly acknowledges this narrow novelty scope in the introduction.

Potential errors

• Minor LaTeX cross-referencing issue (cleveref shared counter): Proposition 11.1, Lemma 8.1, and Lemma 10.1 are rendered as 'theorem 11.1', 'theorem 8.1', and 'theorem 10.1' in the proofs on pages 6, 8, and 11.
• In Lemma 10.1 (page 11), the bound on the first and second y-derivatives of L_q(qy)/A_q(y) - 1 is stated concisely in prose; displaying the explicit boundary terms from differentiating the incomplete Gamma integral would make the two-derivative control more transparent.

Strong novelty candidates

• Theorem 8.2 (One-change coefficient law for the oriented-plane normalizer): the exact representation of the nonlinear Wronskian-type combination J_q = L_q L'_q - kappa L_q L''_q + kappa (L'_q)^2 in terms of central parity-window binomial moments and its uniform sign-change proof across all dimensions q >= 4.
• Theorem 3.1 and Theorem 3.2: the positive all-degree complete-homogeneous symmetric polynomial series for the Haar oriented two-plane partition function and the resulting all-field simple-bivector extremality.

Unresolved material objections

None reported.

Strengths

• Theorems 3.1–3.2 give an economical spectral reduction: Dirichlet moments produce a positive complete-homogeneous series, and the degree-four contribution enforces simplicity at every nonzero field. The exterior/Frobenius factor of two and the full mass-norm reproduction body are explicitly handled.
• The low-dimensional arguments are coefficientwise, not numerical. For higher dimensions, equation (26) correctly converts the derivative coefficient into a parity-window second-moment comparison. The exceptional c_{4,3}=0 is accommodated without compromising the one-change argument.
• Theorem 9.1 does not confuse local Landau behavior with global phase selection. The shared numerator in λ_q' and F_q', together with stationary-point stability and endpoint exclusion for λ>0, supports unique positive coexistence and rules out subsequent exchanges and radial reentrance.
• The thermodynamic proof localizes the actual unique contact inside a fixed y-interval before applying uniform Gamma-tail estimates. The logarithmic shifts in κ_c, λ_c, b_c, D_c and R_c are algebraically consistent. The high-fidelity pre-log n−2 matches half the manifold dimension.
• The operational arguments distinguish unrestricted posterior reports from orbit quantization. Constant-divergence covariant channels support the universal-capacity identity, while the iid chain rule supports tensorization. Section 12 correctly limits the finite-block claim to a converse rather than exact finite-codebook achievability.

Weaknesses

• The diagnostic coverage is narrower than a full verification of the manuscript. verify_global_phase.py uses a different, coarser tail bound starting at r=8 rather than directly regenerating Appendix A's weighted-tail certificate. The contact script asserts only that two residual magnitudes decrease between its first and last sampled dimensions; it does not certify the stated remainder orders.
• The novelty comparisons with hypergeometric inequalities and the companion results in references [21] and [23] remain qualified here because those sources were not independently inspected. This is a limitation of the present assessment, not evidence that the cited artifacts are unavailable. The main analytic proofs do not require the unpublished companion results.
• The canonical PDF's replay command uses the original work/twelfth_paper workspace path, whereas the deposited replay README supplies a self-contained directory-relative command. Align the manuscript instructions with the released layout and distinguish cached-artifact integrity checks from fresh computational comparisons.

Potential errors

• Zero-field counterexample: φ_0(b)=0 for every b∈[0,1]. Thus Theorem 7.2's uniqueness assertion for 0≤λ≤λ_0 is false at λ=0; the same exception affects Theorem 9.1's below-contact statement and Theorem 6.1's exclusion of b=1 at every finite field. Restrict these assertions to positive fields and state the zero-field degeneracy separately. This does not undermine the positive-field phase diagram.
• In make_complete_phase_figure.py, the l2 prefactor is smaller by a factor of two. At κ=0, cumulants() returns variance 1/[(q+1)(q+2)], whereas equation (12) requires 2/[(q+1)(q+2)]. The prefactor multiplying the sum of the two hypergeometric terms should be 1/[(q+1)(q+2)], not half that. All supplied figure callers discard this variance, so the plotted contacts and rates are unaffected.
• verify_global_phase.py prints a JSON status of FAIL without raising an exception or returning a nonzero exit code. replay_all.ps1 checks exit codes, so this diagnostic's FAIL does not itself stop the wrapper. Moreover, its report is printed rather than written to global_phase_verification.json; hashing that cached file does not compare it with the freshly computed report. Add explicit failure propagation and structured output comparison.

Strong novelty candidates

• The exact all-q coefficient-sign law for J_q, obtained through central parity-truncated binomial moments, including the isolated zero c_{4,3}. This is the manuscript's clearest distinctive technical contribution, subject to independent literature comparison.
• The resulting complete source-specific phase classification: continuous activation for n=3,4,5 and exactly one fold and one positive coexistence contact for every n≥6, yielding the two-piece unrestricted RDF without later radial reentrance.
• Uniform finite-q contact control and explicit logarithmic corrections for all five coexistence observables. The manuscript appropriately distinguishes these refinements from the leading limiting constants attributed to reference [21].

Unresolved material objections

None reported.

Strengths

• Pages 3-6, Theorems 3.1, 3.2 and 6.1: paired Dirichlet parameters give the all-degree positive series, h_2 establishes the unique simple orbit, and each feasible simple radius lies in the mass-norm reproduction ball.
• Page 6, Theorem 6.1: the gradient argument at a simple external bivector correctly eliminates the extra Hodge-star stabilizer direction in n=4. The Gibbs posterior and independent flags close the unrestricted RDF converse.
• Pages 7-9, Lemma 8.1/Theorem 8.2: the binomial weight ratio is increasing in d squared; central truncation proves the positive tail, and the boundary estimate handles r<q. The displayed Appendix A polynomial identity was verified exactly as a rational function.
• Pages 9-10, Theorem 9.1: the one-change coefficient law yields exactly one fold; F prime equals m-kappa m prime. Endpoint exclusion and positive fourth cumulant then yield one contact below the spinodal and no later exchange.
• Pages 10-11, Theorem 10.2: the proof first confines the physical contact to a fixed y interval using uniqueness, then applies uniform Gamma-tail control. The five logarithmic contact shifts follow consistently from that localized expansion.
• Pages 12-13, Theorem 12.1 and (45): constant divergence and the iid Haar chain-rule lower bound establish tensorization. Constant tilted information specializes the consulted Kostina-Verdu converse; a direct Gibbs small-ball bound also supports (45).

Weaknesses

• Page 14 and README: replay_all.ps1 and REPLAY_ARTIFACTS.sha256 are absent. Individual Python programs completed, but the advertised one-command frozen-archive check was not executed and its archive hash was not verified.
• The statement about two separate 6,783-check replays overstates the coverage of the supplied verify_global_phase.py: it uses a 37-by-100 coefficient rectangle and a smaller direct-reduction subset. The other supplied coefficient script does perform 6,783 sign and 297 tail checks.
• The contact program tests numerical residual improvement between selected q values, not a uniform error bound. The sharp thermodynamic theorem therefore rests on the analytic localization and Gamma-tail argument, not these decimals.
• Literature checking was targeted. I consulted the spherical RDF and finite-block converse results and two hypergeometric-paper abstracts, but did not exhaustively compare all possible prior coefficient inequalities or obtain the unpublished thermodynamic companion.

Potential errors

• Pages 7 and 9, Theorems 7.2 and 9.1: the zero-field objective is phi_0(b)=0 on all [0,1]. Thus the claimed unique zero radius at lambda=0 is false. This endpoint degeneracy leaves the stated positive-field thresholds and RDF unchanged.
• make_complete_phase_figure.py, cumulants: L double-prime is missing a factor two. At q=4,kappa=0 it returns variance 1/30 instead of 1/15. The variance output is discarded by the figure/contact callers, so their reported contacts are not invalidated.

Strong novelty candidates

• Theorem 8.2: the exact sign pattern c_q,r<0 for 2<=r<q and c_q,r>0 for r>=q, with its exceptional zero, proved through parity-window variance and a uniform rational tail certificate.
• Theorem 9.1: the resulting complete, dimension-uniform branch and coexistence classification for Haar oriented two-plane RDFs.
• Theorem 10.2: explicit finite-size logarithmic displacements of all five observables at the uniquely selected physical contact.

Unresolved material objections

None reported.

Strengths

• Theorem 3.2 is genuinely global over the mass-norm reproduction body at fixed Euclidean radius: positivity of the complete-homogeneous series and strictness of its k=2 term force a simple bivector without a generic-canonical-form assumption.
• Proposition 4.1 identifies the exact overlap density from all moments on compact support, allowing the RDF and phase analysis to use an elementary scalar normalizer in every dimension.
• Theorem 8.2 supplies the central missing argument: Cauchy multiplication converts the fold coefficients to r-E[d^2] for a parity-truncated centered binomial law; monotone-likelihood-ratio ordering and the displayed rational tail bound determine the complete sign pattern.
• Theorem 9.1 uses the one-change law efficiently: the stationary multiplier decreases then increases, while the contact function has the same derivative and diverges positively, leaving exactly one positive contact and one active post-fold branch.
• Theorem 10.2 derives all five coexistence observables from one uniform endpoint Laplace expansion and the nondegenerate root y*-1=2 log y*, with internally consistent logarithmic shifts.
• The supplied exact coefficient programs passed 6,783 sign checks and 297 tail bounds; the global-phase JSON structure matched the fresh execution, and the contact and figure diagnostics ran through the stated ranges.

Weaknesses

• Theorem 6.1 invokes standard lower-semicontinuous rate-distortion duality and Fenchel-Moreau closure without a precise theorem statement specialized to the compact source and mass-ball reproduction space. The argument is plausible but condensed.
• Lemma 10.1 states that two y-derivatives of the incomplete-Gamma representation introduce only boundary densities and polynomial q-factors; the exponential absorption is sound on compact I inside (1,infinity), but the derivative formulas and uniform constants are not displayed.
• The novelty search was necessarily limited to exact titles, distinctive phrases, and the cited primary works. It found no independent duplicate theorem, but that does not certify priority across all hypergeometric, Curie-Weiss, or homogeneous-source literature.
• The paper's reproducibility paragraph describes a PowerShell wrapper, an internal checksum ledger, and a sealed artifact archive, whereas this batch supplies only four component Python scripts, two JSON files, a README, and RELEASE_QA.md.

Potential errors

• The one-command instruction cannot be executed from the supplied artifacts because replay_all.ps1 is missing; REPLAY_ARTIFACTS.sha256 is also missing, so the claimed internal seal and deterministic-output inventory were not independently verified.
• Equation (45) is a specialization of the Kostina-Verdu d-tilted-information converse. The cited primary source supports general finite-block converse bounds, but this review did not rederive every measure-theoretic condition for the continuous Grassmann source; the constant tilted information calculation itself is consistent.
• The finite scripts test large rectangles and the exact tail algebra but do not execute the analytic monotone-likelihood-ratio proof or uniform Gamma-tail lemma. Their PASS results support arithmetic only and must not be read as machine verification of the full theorems.

Strong novelty candidates

• The parity-truncated binomial representation of every fold-numerator coefficient and its dimension-uniform one-change proof for all q at least 4.
• The complete all-dimensional activation classification: continuous for n=3,4,5 and exactly one discontinuous coexistence contact with no reentrance for every n at least 6.
• The explicit logarithmic finite-size displacement of kappa_c, lambda_c, b_c, D_c, and R_c around the unique thermodynamic contact.
• The constant tilted-information identity and resulting zero-dispersion finite-block converse for this transitive continuous source.

Unresolved material objections

None reported.