Referee A recommended accept with high confidence and assigned canon match none. Referee B recommended minor revision with high confidence and assigned partial match. The synthesis adopts minor revision because the presentation clarifications and caveats identified by B are necessary for rigor and align with the minor comments from A.
A one-parameter Gaussian family realizes the Fisher-Rao metric equal to the log-Hessian of J.
The construction works but the paper should address whether the realization is minimal or privileged.
| ID | Claim | Section | Importance | Status | Lean match | Author action |
|---|---|---|---|---|---|---|
| C7 | A one-parameter Gaussian family realizes the Fisher-Rao metric equal to the log-Hessian of J. | Section 5 | clarify before publication | plausible | The construction works but the paper should address whether the realization is minimal or privileged. | |
| C1 | Permutation symmetry together with the dimensional-reduction condition forces alpha_i = 1/n for all i. | Section 2, permutation symmetry lemma | no action needed | verified | The conclusion is correct; only the uniformity argument in the case split requires a one-sentence continuity justification. | |
| C2 | The Hessian of J in logarithmic coordinates t_i = log x_i has constant rank one with an integrable (n-1)-dimensional radical distribution orthogonal to alpha. | Section 3.1 | no action needed | verified | Explicit matrix form g_ij = cosh(S) alpha_i alpha_j confirms the rank-one claim. | |
| C3 | The Hessian of J in original x-coordinates is generically nondegenerate and defines a pseudo-Riemannian metric away from the hypersurface R=1. | Section 3.2 | no action needed | verified | The decomposition into A + beta u u^T and the determinant condition tanh(S) = sum alpha_i are correctly derived. | |
| C4 | In two dimensions the Christoffel symbols and Ricci scalar of the Levi-Civita connection on the x-manifold are given by explicit closed-form expressions. | Section 4.2 | no action needed | verified | The formulas are stated and can be verified by direct computation. | |
| C5 | Gradient flows of J evolve exclusively along the alpha direction and admit the closed-form solution tanh(S/2) = C exp(|alpha|^2 tau). | Section 4.5 | no action needed | verified | The integration is straightforward once the flow is recognized to lie in the alpha line. | |
| C6 | J is exactly the symmetrization of the Itakura-Saito divergence applied to R. | Section 5 | no action needed | verified | The identity 1/2(D_IS(1||R) + D_IS(R||1)) = J(R) holds by direct expansion. |
V0 Prose only; no formal artifacts supplied.
The manuscript extends the canonical reciprocal cost J(x) = 1/2(x + x^{-1}) - 1 to n positive variables via R = prod x_i^{alpha_i} with alpha_i = 1/n fixed by permutation symmetry and dimensional reduction. It compares the rank-one Hessian in logarithmic coordinates (integrable radical distribution) with the generically nondegenerate pseudo-Riemannian Hessian in original coordinates (singular on R=1 and conditionally on a secondary locus). Explicit 2D Christoffel symbols, Ricci scalar, three geodesic families, closed-form gradient flows, and information-geometric realizations via symmetrized Itakura-Saito divergence and a Gaussian Fisher-Rao model are derived.
The paper supplies concrete reference computations for the Hessian geometry of the canonical cost and places it inside standard information-geometric machinery. The rank analysis, explicit curvature, and divergence identifications are useful for optimization and statistical modeling. Significance is modest but solid within differential and information geometry.
A one-parameter Gaussian family realizes the Fisher-Rao metric equal to the log-Hessian of J.
The construction works but the paper should address whether the realization is minimal or privileged.
Permutation symmetry together with the dimensional-reduction condition forces alpha_i = 1/n for all i.
The conclusion is correct; only the uniformity argument in the case split requires a one-sentence continuity justification.
The Hessian of J in logarithmic coordinates t_i = log x_i has constant rank one with an integrable (n-1)-dimensional radical distribution orthogonal to alpha.
Explicit matrix form g_ij = cosh(S) alpha_i alpha_j confirms the rank-one claim.
The Hessian of J in original x-coordinates is generically nondegenerate and defines a pseudo-Riemannian metric away from the hypersurface R=1.
The decomposition into A + beta u u^T and the determinant condition tanh(S) = sum alpha_i are correctly derived.
In two dimensions the Christoffel symbols and Ricci scalar of the Levi-Civita connection on the x-manifold are given by explicit closed-form expressions.
The formulas are stated and can be verified by direct computation.
Gradient flows of J evolve exclusively along the alpha direction and admit the closed-form solution tanh(S/2) = C exp(|alpha|^2 tau).
The integration is straightforward once the flow is recognized to lie in the alpha line.
J is exactly the symmetrization of the Itakura-Saito divergence applied to R.
The identity 1/2(D_IS(1||R) + D_IS(R||1)) = J(R) holds by direct expansion.
Publication readiness is governed by the referee recommendation, required revisions, and the blockers summarized above.
Final recommendation
accept
minor_revision
minor_revision is adopted because the clarifications on proof uniformity, sign choice, geodesic caveats, and Fisher-Rao questions are necessary for full rigor and align with the minor comments from Referee A.
Canon match strength
none
partial
partial is adopted because the manuscript directly cites and extends the 1D uniqueness theorem washburn_uniqueness_aczel while remaining consistent with the canon cost foundation.
Depth of Fisher-Rao section
strongest paragraph in the application section
valuable but requires two additional sentences on minimality and privileged status
The section is a genuine contribution; the two clarifying sentences requested by B improve it without altering the assessment of strength.
Referee A recommended accept with high confidence and assigned canon match none. Referee B recommended minor revision with high confidence and assigned partial match. The synthesis adopts minor revision because the presentation clarifications and caveats identified by B are necessary for rigor and align with the minor comments from A.
IndisputableMonolith.Foundation.DimensionForcingIndisputableMonolith.Unification.SpacetimeEmergenceIndisputableMonolith.Foundation.ConstantDerivationsIndisputableMonolith.Foundation.LogicAsFunctionalEquationIndisputableMonolith.Foundation.SimplicialLedgerIndisputableMonolith.Quantum.BlackHoleInformationIndisputableMonolith.Astrophysics.MassToLightIndisputableMonolith.Astrophysics.PlanetaryFormationFromJCostIndisputableMonolith.CPM.ExamplesIndisputableMonolith.ILG.KernelIndisputableMonolith.Relativity.Lensing.ShadowPredictionsIndisputableMonolith.Foundation.UniversalForcingSelfReference{
"canon_match_strength": "none",
"cited_canon_theorems": [],
"confidence": "high",
"issue_inventory": [],
"load_bearing_issues": [],
"major_comments": [],
"minor_comments": [
{
"comment": "The choice of alpha_i = 1/n under permutation symmetry and dimensional reduction is justified, but the manuscript could briefly note that other extensions (e.g., non-uniform alpha or different composition rules) exist and why they are set aside.",
"section": "Section 2"
},
{
"comment": "The singular locus condition |sum alpha_i| \u003c 1 is correctly derived; a short remark on its geometric meaning (where the metric signature changes or degenerates) would aid readers.",
"section": "Section 3.2"
},
{
"comment": "The residual error plots for numerical geodesics are useful; adding a sentence on the numerical method (e.g., integrator, step size) would improve reproducibility.",
"section": "Section 4"
}
],
"optional_revisions": [],
"paper_summary": "The manuscript extends the one-dimensional canonical reciprocal cost J(x) = 1/2(x + x^{-1}) - 1 to n dimensions via a weighted product form J(x) = 1/2(R + R^{-1}) - 1 with R = product x_i^{alpha_i}. It analyzes the Hessian geometry in logarithmic coordinates t_i = log x_i, where the metric has constant rank one and an integrable (n-1)-dimensional radical distribution, versus the original x-coordinates, where the Hessian is generically nondegenerate and defines a pseudo-Riemannian metric away from a singular hypersurface determined by tanh(S) = sum alpha_i. The paper computes explicit Christoffel symbols and Ricci scalar for the Levi-Civita connection in 2D, classifies affine geodesics (straight lines in each coordinate system, with domain restrictions in x-coordinates) and Levi-Civita geodesics (singular on the zero-cost hypersurface), studies gradient flows (confined to the alpha direction), and relates the construction to the symmetrized Itakura-Saito divergence, Bregman divergences, and a Fisher-Rao realization via a normal distribution model.",
"recommendation": "accept",
"required_revisions": [],
"significance": "The work provides a detailed geometric analysis of the Hessian structures induced by this cost function under different affine structures, clarifying how coordinate choice affects degeneracy, curvature, and geodesic behavior. The explicit 2D calculations and information-geometric interpretations offer concrete tools for optimization and statistical modeling contexts.",
"strengths": [
"Explicit closed-form expressions for the Hessian, Christoffel symbols, and Ricci scalar in both coordinate systems.",
"Clear separation of the two affine structures and their distinct geodesic families.",
"Rigorous treatment of the rank-one degeneracy and its foliation in logarithmic coordinates.",
"Concrete link to Fisher-Rao geometry via an explicit statistical model.",
"Well-organized progression from 1D uniqueness to multi-D geometry and applications."
]
}
{
"canon_match_strength": "partial",
"cited_canon_theorems": [
{
"decl": "washburn_uniqueness_aczel",
"module": "IndisputableMonolith.Cost.FunctionalEquation",
"note": "The 1D uniqueness theorem the paper cites as [WZ] and builds on; the manuscript extends J to n variables but does not re-prove or weaken 1D uniqueness.",
"relation": "supports"
},
{
"decl": "RCL_is_unique_functional_form_of_logic",
"module": "IndisputableMonolith.Foundation.LogicAsFunctionalEquation",
"note": "The precursor paper module that bundles WZ + WZE; the manuscript inherits the same composition-law setting but never engages logical-formalization content.",
"relation": "tangential"
},
{
"decl": "all_constants_from_phi",
"module": "IndisputableMonolith.Foundation.ConstantDerivations",
"note": "Constants downstream of J; the manuscript explicitly stays away from physical-constants applications, so the canon\u0027s $\\varphi$-ladder content is not engaged.",
"relation": "tangential"
},
{
"decl": "Jcost_near_identity",
"module": "IndisputableMonolith.Unification.SpacetimeEmergence",
"note": "The canon also analyses J near the identity via $\\varepsilon^2/(2(1+\\varepsilon))$; the manuscript\u0027s Hessian-at-identity matrix $\\alpha\\alpha^\\top$ is the multidimensional generalisation of the same quadratic expansion.",
"relation": "supports"
}
],
"confidence": "high",
"issue_inventory": [],
"load_bearing_issues": [],
"major_comments": [
{
"canon_evidence": [],
"comment": "The proof of the permutation-symmetry lemma needs tightening at one step. From the identity $A+A^{-1}=B+B^{-1}$ with $A=\\prod x_{\\sigma(i)}^{\\alpha_i}$ and $B=\\prod x_i^{\\alpha_i}$ the authors correctly note $A=B$ or $A=B^{-1}$, then proceed to argue componentwise by plugging $x_j=e$, $x_i=1$. The implicit step that needs to be justified is that the disjunction can be assumed *uniformly* on a connected region in $(\\mathbb{R}^+)^n$: a priori one branch could hold at some points and the other branch at others. Since $A,B$ are continuous and the locus $A=B$ is closed, and on the complement $A=B^{-1}$ holds (also closed), one of the two must hold on an open set, and by analyticity globally. Please add a one-sentence continuity/analyticity argument; without it the case split looks point-wise rather than functional. The conclusion of the lemma is correct, so this is a presentation fix, not a substantive error.",
"section": "Section 2, Lemma on permutation symmetry"
},
{
"canon_evidence": [],
"comment": "The derivation $\\cosh(an\\log x)-1=\\cosh(\\log x)-1$ for all $x\u003e0$, hence $|an|=1$, is correct, but the manuscript jumps directly to $a=\\pm 1/n$ without acknowledging that $a=-1/n$ gives $R(x)\\to R(x)^{-1}$, i.e. the same $J$ by the symmetry $J(R)=J(R^{-1})$. The sentence \"Without loss of generality, we take $a=1/n$\" should be expanded to make the involution explicit: the choice is a coorientation, not a normalisation, and the statement that this is *canonical* deserves a one-line justification (e.g., it is the unique choice with $\\alpha_i\u003e0$).",
"section": "Section 2.2, dimensional reduction $a=\\pm 1/n$"
},
{
"canon_evidence": [],
"comment": "The paragraph contrasting the present rank-1 Hessian with Carrollian (co-rank 1) and Newton-Cartan (rank-1 temporal plus rank-3 spatial) structures is accurate and useful, but currently treats $n=2$ as a marginal case where the present setup coincides with Carrollian. This deserves to be sharpened: in $n=2$ the metric has signature $(+,0)$, so the kernel is one-dimensional and the structure *is* a degenerate Carrollian-type metric on $\\mathbb{R}^2$. Conversely, for $n\\ge 3$ the kernel is $(n-1)$-dimensional and the natural physical analogue is the inverse: a leaf-wise foliation with a one-dimensional transverse cost direction, closer to a fibered structure over the line $\\mathbb{R}\\cdot\\alpha$ than to Carrollian geometry. A clarifying sentence here would help the reader who knows Carrollian/NC geometry.",
"section": "Section 3.1, comparison with Carrollian and Newton-Cartan geometry"
},
{
"canon_evidence": [],
"comment": "The formal-series analysis near the singular locus extracts the two tangent-direction constraints $x_1=y_1 z_0^{1/a+1/b}$ and $x_1=-(b/a)y_1 z_0^{1/a+1/b}$ and then states that no matching conditions arise because the equations are singular. This is the right honest conclusion, but the manuscript should state explicitly that the formal-series ansatz \\eqref{supposedseries} *may have no convergent representative*: the constraints are necessary but the analysis does not produce a Frobenius/Briot-Bouquet type existence theorem. Without this caveat a reader could read the bullet list as a description of actual extending geodesics. A single sentence (\"the analysis here is purely formal and we make no claim of regular extension across $\\Delta=0$\") closes the gap.",
"section": "Section 4.6, Levi-Civita geodesics near $\\Delta=0$"
},
{
"canon_evidence": [],
"comment": "The Fisher-Rao construction via a Gaussian family with mean $m(S)=\\int_0^S\\sqrt{\\cosh u}\\,du$ produces $I_{ij}=g_{ij}$, but the parenthetical \"not necessarily unique\" understates the situation. Two natural questions deserve a sentence each. (i) Is this realization minimal in the sense of an embedding dimension? (ii) Does the Gaussian choice carry any privileged status, or is it merely the first explicit family that works? An honest answer would help; the paper does not need to resolve these but should not leave them implicit. The current section is otherwise a real contribution and the strongest paragraph in the application section.",
"section": "Section 5, Fisher-Rao realization"
}
],
"minor_comments": [
{
"comment": "There are unmerged editorial color marks {\\color{blue}...} throughout the source (e.g., \u0027co-rank\u0027, \u0027we obtain\u0027, \u0027so\u0027, \u0027lines:\u0027, \u0027derivatives, we obtain\u0027). These need to be removed before production.",
"section": "Manuscript-wide"
},
{
"comment": "The unit-log-curvature condition is written as $\\lim_{t\\to 0} 2F(e^t)/t^2 = 1$. Reading this with $F=J$ gives $\\cosh(t)-1 \\sim t^2/2$, so $2(t^2/2)/t^2 = 1$. The condition is correct, but the heuristic \u0027unit log-curvature\u0027 is non-standard; consider naming it \u0027unit second log-derivative\u0027 or simply restating it as $F\u0027\u0027(1)=1$ via the chain rule, which is more transparent.",
"section": "Equation (2)"
},
{
"comment": "The notation $S(t)$ versus $S(x)$ is used interchangeably with the same symbol; consider $S_t$ and $S_x$, or explicitly say $S(x)=\\sum\\alpha_i\\log x_i$ once and use a single name.",
"section": "Section 2"
},
{
"comment": "Theorem 1 (the uniqueness theorem of WZ) is restated in this section; the manuscript would benefit from a single forward pointer (already partly present) saying that the multidimensional construction of Section 2 does *not* inherit uniqueness from WZ, with a one-line indication of which axioms fail in $n\\ge 2$.",
"section": "Section 3 first paragraph"
},
{
"comment": "The on-shell rank-1 Hessian at $R=1$ should be related back to the calibration $F\u0027\u0027(1)=1$ from Theorem 1: at $x_i=1$ one has $\\nabla_x^2 J = \\alpha\\alpha^\\top$, recovering the same calibration in matrix form. This is a small remark but it ties the manuscript to its prequel.",
"section": "Section 4.1, equation (15)"
},
{
"comment": "Equation (2dT) is a $2\\times 2$ scalar multiple of $(a,b)^\\top(a,b)$; the manuscript already gives this factored form below. Consider showing the factored form first and the matrix form as an unfolding.",
"section": "Section 4.2, eq. (2dT)"
},
{
"comment": "The basis vectors $\\beta^k$ spanning the radical distribution are defined in the $n$-dimensional treatment but never connected back to the explicit $(b,-a)$ choice used here. A one-line cross-reference is enough.",
"section": "Section 4.6 (qr coordinates)"
},
{
"comment": "Figures 2 and 3 are labelled \u0027multiplotxy1.pdf\u0027 etc.; the captions describe specific initial conditions which should be checked against the data in the figures. Reviewer cannot verify pixel-level claims; please confirm the figures match the stated parameters.",
"section": "Section 4.7"
},
{
"comment": "The Itakura-Saito symmetrization $\\tfrac12(D_{\\rm IS}(1\\|x)+D_{\\rm IS}(x\\|1))=J(x)$ is a clean identity worth labelling as a Proposition rather than left as an inline calculation.",
"section": "Section 5"
},
{
"comment": "[Carrollian] arXiv:2510.21651 is dated 2026 (RIKEN iTHEMS Report 26); the date should be checked. [PardoGuerra2026] arXiv:2601.12194: similarly check the identifier format (arXiv identifiers do not have the form 26xx.xxxxx for 2026 submissions on the current numbering scheme).",
"section": "References"
},
{
"comment": "The conclusion mentions \u0027multidimensional uniqueness\u0027 as future work but does not say what the conjectural uniqueness theorem would look like (which axioms? which class of multidimensional extensions?). A one-paragraph speculation would point readers in the right direction.",
"section": "Conclusion"
}
],
"optional_revisions": [],
"paper_summary": "The paper studies the differential-geometric structures induced by the canonical reciprocal cost $J(x)=\\tfrac12(x+x^{-1})-1$ once extended to $n$ positive variables via $R(x)=\\prod x_i^{\\alpha_i}$, so that $J=\\tfrac12(R+R^{-1})-1$. After fixing the multidimensional extension by permutation symmetry and a dimensional-reduction condition $J(x,\\dots,x)=J(x)$ (which forces $\\alpha_i=1/n$), the authors compare two Hessian structures on $M=(\\mathbb{R}^+)^n$: one built from logarithmic affine coordinates $t_i=\\log x_i$, the other from the original $x_i$. The logarithmic Hessian $g_{ij}=\\cosh(\\alpha\\cdot t)\\,\\alpha_i\\alpha_j$ is shown to have rank one at every point, with an integrable $(n-1)$-dimensional null distribution orthogonal to $\\alpha$, and the geometry is therefore effectively one-dimensional along $S=\\alpha\\cdot t$. The $x$-Hessian, by contrast, is generically nondegenerate and defines a pseudo-Riemannian metric, with explicit singular hypersurfaces $\\{R=1\\}$ and a conditional secondary locus existing iff $|\\sum\\alpha_i|\u003c1$. In two dimensions the Christoffel symbols and Ricci scalar of the Levi-Civita connection on $M_x$ are computed explicitly. Three families of curves (affine geodesics on $M_t$, affine geodesics on $M_x$, Levi-Civita geodesics on $M_x$) are compared; the first are globally defined straight lines in $t$, the second restricted to the positive orthant, the third singular on $\\Delta=0$. Gradient flows of $J$ are integrated in closed form and shown to evolve only along $\\alpha$. A short final section relates the construction to information geometry: $J$ is the symmetrization of the Itakura-Saito divergence applied to $R$, the log-Hessian is the second-order Bregman term, and a Fisher-Rao realization is given via a one-parameter Gaussian family with mean $m(S)=\\int_0^S\\sqrt{\\cosh u}\\,du$.",
"recommendation": "minor_revision",
"required_revisions": [],
"significance": "The paper is the natural geometric companion to the WZ uniqueness theorem and the WZE d\u0027Alembert Inevitability Theorem: it asks what kind of multivariable geometry $J$ supports. The two main observations, rank-one degeneracy of the log-Hessian and generic nondegeneracy of the $x$-Hessian off explicit hypersurfaces, are not deep theorems but they are clean, correctly proved, and worth recording. The closed-form Ricci scalar in 2D and the explicit identification of the singular locus are useful reference computations. The information-geometric interpretation (Itakura-Saito symmetrization, Bregman quadratic, Fisher-Rao embedding) is the most novel contribution and gives the paper a hook beyond pure curiosity-driven Hessian geometry: it places the canonical reciprocal cost inside the standard divergence/exponential-family machinery in a clean way. Significance is modest but real; the paper does not over-claim. It does not attempt to derive new physics from the multidimensional cost.",
"strengths": [
"Mathematics is correct as far as I can verify by hand: the rank-1 identity $\\nabla^2 J=\\cosh(S)\\alpha\\alpha^\\top$, the $x$-Hessian decomposition $\\nabla_x^2 J = A+\\beta uu^\\top$, the determinant computation in 2D, and the gradient-flow integration $\\tanh(S/2)=Ce^{|\\alpha|^2\\tau}$ all check out.",
"Honest scope: the paper stays in pure geometry/information geometry and does not over-extend to physical predictions. The relation to the previously published WZ and WZE results is correctly cited.",
"Clean and useful comparison of three distinct curve families (affine on $M_t$, affine on $M_x$, Levi-Civita on $M_x$) arising from the same potential under different affine structures. This is the right level of granularity for a Hessian-geometry paper.",
"The information-geometric section is concise and substantive: the symmetrized Itakura-Saito identification of $J$, the Bregman quadratic, and the explicit Fisher-Rao realization via Gaussians with mean $\\int_0^S\\sqrt{\\cosh u}\\,du$ are nice independent observations.",
"Singular locus analysis is explicit: the condition $\\tanh(S)=\\sum\\alpha_i$ with existence iff $|\\sum\\alpha_i|\u003c1$ gives a sharp criterion, and the Ricci scalar computation makes the divergence loci explicit.",
"Numerical residual diagnostics (Figure 4) for the Levi-Civita geodesics are appropriate quality control for the symbolic Mathematica work."
]
}