{"id":"b85b6207-09b8-4b15-ae41-42aa05b956e4","arxiv_id":"2508.18165","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"For three particle flavors, the positivity cone C_W has exactly three families of extremal rays, one of which yields genuinely new inelastic constraints not implied by elastic bounds.","lead":"The paper classifies all extremal rays of the positivity cone for three-flavor effective field theories, completing the n=3 case. It also proves that for O(3), Z_2^3, and SO(2) symmetric amplitudes the elastic bounds generate the full positivity cone.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The classification rests on the unshown determinant elimination in Prop. 4.25; the displayed 6x6 matrix and kernel basis need an independent symbolic check before the third family in Theorem 4.1 is fully established.","rationale":"I read the paper primarily as a mathematical classification theorem, and the physical interpretation is carefully scoped by footnote 3 and the Appendix; I do not see the dispersion-relation identification as a threat to the mathematical claim. The most load-bearing risk is the algebraic elimination in Prop. 4.25, which the reader also identified as fragile. I agree with that part of the reader's weakest_assumption, but not with giving equal weight to the physical caveat for the central theorem. The surrounding proof is detailed and internally coherent, with independent support from the n=2 case in Theorem 3.17, but there is no machine-checked formalization and the determinant step is summarized. My recommended verdict is therefore unchanged rather than conditional: the omitted computation is mechanical and very likely correct, but a direct symbolic check would settle the one place where the completeness of Theorem 4.1 could fail.","tokens_in":54187,"tokens_out":25761,"duration_ms":245794,"concrete_test":"Run a computer-algebra verification of Prop. 4.25: (1) recompute the 6x6 matrix R(S) from the parameter definitions in the proof; (2) compute all principal minors symbolically and solve the system 'all 4x4 principal minors vanish, 1x1-3x3 principal minors are nonnegative, q>0'; (3) confirm that the solution set is exactly the stated parameter values together with g^2>1-d^2+dh, including the d=0 and g=0 subcases; (4) verify that each displayed kernel vector satisfies R(S)v=0 and that the three vectors span the kernel. If any check fails, identify the missing branch or typo and test whether the explicit S in Theorem 4.1 is still positive semidefinite and extremal.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central theorem's completeness depends on the step from the abstract extremality criterion (Prop. 4.15) to the explicit three-parameter family (Prop. 4.25). That step is the least secure point: the proof sets up a 6x6 Gram matrix for R(S) and then states, without displaying the elimination, that requiring the fourth principal minors to vanish, with lower principal minors nonnegative and q>0, yields a=1+d^2, b=1+dh, c=dg, e=1+g^2+h^2, f=g(d+h), b+q=d^2+g^2, and g^2>1-d^2+dh. It also asserts a three-vector kernel basis. No symbolic derivation is shown, and the ordering of the e1∨e3 and e2∨e3 columns in the preliminary matrix in the proof is not reconciled with the final displayed matrix. If the principal-minor conditions admit another branch or an additional inequality, then Theorem 4.1's third family either contains non-extremal tensors or misses extremal ones; if the kernel basis is mis-specified, the subsequent no-gamma^4 and no-(alpha∨beta)^2 checks in Prop. 4.15 lose their anchor. Since the inelastic bounds in Section 5 are exactly the dual inequalities from the third family, the applications inherit this dependence. This is not a claim that the theorem is false; it is a claim that the load-bearing algebraic computation is currently asserted rather than demonstrated.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the forward-limit positivity cone C_W for n flavors, viewed as a linear spectrahedron of four-tensors, and determines all its extremal elements for dim V = 3. The main result, Theorem 4.1, gives an if-and-only-if classification into three explicit families: rank-one totally symmetric tensors; tensors of the form (α1∨α2)^2 + (α1∧α2)^2; and a three-parameter family parameterized by (g,d,h) with g^2 > 1 - d^2 + dh. The proof splits according to the rank of R'(S), using a kernel-inclusion criterion for extremality in spectrahedra (Cor. 2.14) and a sequence of auxiliary propositions. The paper then derives the resulting elastic and inelastic positivity bounds, proves that for amplitudes with O(3), Z_2^3, or SO(2) symmetry the elastic bounds suffice, and applies the O(3) result to pion scattering in chiral perturbation theory.","tokens_in":54587,"tokens_out":7692,"duration_ms":76335,"significance":"If correct, the classification solves the extremal representation problem for the three-flavor forward-limit positivity cone, thereby completing the two-flavor result of Ref. [98] and providing the full set of positivity bounds for three-flavor EFTs, including the inelastic bounds that are absent in the two-flavor case. The general structural results—the kernel-based extremality criterion, the two-dimensional recovery in Theorem 3.17, and the necessary kernel-dimension bounds in Proposition 3.12—are clean and potentially reusable. The paper also delivers concrete, falsifiable inequalities for the symmetric cases and reproduces the known pion bounds. These are significant contributions if the algebraic core of the proof is fully secured.","major_comments":[{"comment":"The load-bearing step of the classification is asserted rather than demonstrated. After setting up the 6×6 matrix in terms of parameters a,b,c,d,e,f,g,h,q, the proof states without derivation that requiring the fourth principal minors to vanish, the first three principal minors to be nonnegative, and q>0 yields a=1+d^2, b=1+dh, c=dg, e=1+g^2+h^2, f=g(d+h), b+q=d^2+g^2 and g^2>1-d^2+dh. This elimination is the only step connecting the abstract extremality criterion (Propositions 4.15 and 4.24) to the explicit third family in Theorem 4.1, and the inelastic bounds in §5.1.2 depend on it directly. The authors should display the elimination, provide a verifiable symbolic computation, or give a rigorous argument showing that no other branch and no additional inequality can arise.","section":"§4.4, Proposition 4.25"},{"comment":"The preliminary matrix displayed in the proof and the final matrix in the proposition are not reconciled. With the stated basis {e1^2, e2^2, e3^2, e1∨e2, e1∨e3, e2∨e3}, the preliminary matrix has entries (4,5)=0, (4,6)=d, (5,6)=g, (6,6)=b+q, whereas the displayed final matrix has (4,5)=d, (4,6)=g, (5,6)=0, (6,6)=1. Unless a column/row reordering or a rescaling of basis vectors is being performed, the two matrices are not the same. A reader cannot reproduce the claimed kernel basis without knowing which convention is used. The three cases of the kernel basis (d,g≠0; d=0; g=0) are also presented without derivation; since these vectors are used to check the no-nontrivial-solution conditions in Proposition 4.15, each case needs an explicit verification that the displayed vectors do indeed span the kernel.","section":"§4.4, Proposition 4.25 (matrix and kernel basis)"},{"comment":"The proof of extremality for the rank-3, rank-R'=1 case relies on the claim that each of the three displayed systems of quadratic equations has no non-trivial real solution. This claim is stated without proof. Because Corollary 2.14 requires exactly this vanishing condition, a short case-by-case justification (for instance, by sign analysis or by reducing to a contradiction with g^2>1-d^2+dh) should be included. Without it, the iff direction of Theorem 4.1 for the third family is incomplete.","section":"§4.2, Proposition 4.15"}],"minor_comments":[{"comment":"There is a typo: 'Secrion 3' should be 'Section 3'.","section":"§5.2.2, Lemma 5.5"},{"comment":"The proof's case analysis for vectors with four or more non-vanishing components is concise and somewhat informal; the statement that 'there will always remain a non-trivial constraint' for four non-vanishing components is asserted after only one example. Since Proposition 5.2 is used to characterize elastic bounds, a more systematic enumeration or a reference to the full verification would improve readability.","section":"§5.1.1, Proposition 5.2"},{"comment":"The caption says the regions were approximated by numerically sampling extremal rays, but no details on the sampling algorithm or accuracy are given. A brief description in the text or a supplementary file would make the figure reproducible.","section":"Figure 1"},{"comment":"The list of restrictions in footnote 3 is dense and would benefit from being expanded in the main text, since it delimits the physical scope of the otherwise purely mathematical classification.","section":"Footnote 3"},{"comment":"The reference to 'the appendix of [98]' for the classification of extremal rays of the Z_2^3-invariant cone should be more specific, since the appendix is long and the relevant result is used for an alternative proof of Theorem 5.6.","section":"§5.2.2, Remark 5.8"}],"recommendation":"major_revision","confidential_remarks":"The main theorem is likely correct, but the paper currently asks the reader to accept a nontrivial determinant elimination and kernel-basis computation in Proposition 4.25 without showing it. Because the inelastic bounds and the symmetry applications all depend on this computation, I recommend asking the authors to supply the full elimination (or a machine-checkable symbolic computation) and to reconcile the two matrix displays. If that is provided, the paper would be a strong contribution to the math-ph and positivity-bounds literature."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The bottom line: this is the first complete classification of extremal rays of the three-flavor positivity cone, and I believe it is correct. The genuinely new pieces are the third family in Theorem 4.1 — the inelastic family that generates all inelastic bounds — and the elastic-sufficiency proofs for O(3), Z_2^3, and SO(2). The n=2 case was already in [98]; this is a substantial extension, not a routine one. The χPT application also reproduces the known bounds ℓ2 ≥ 0 and ℓ2 ≥ −ℓ1, a decent sanity check.\n\nThe main soft spot is Proposition 4.25. The proof sets up a 6×6 Gram matrix for R(S), then states that requiring the fourth principal minors to vanish, with the lower ones nonnegative and q>0, yields a=1+d², b=1+dh, c=dg, e=1+g²+h², f=g(d+h), b+q=d²+g², and g²>1−d²+dh. That elimination is the load-bearing step for Theorem 4.1, and it is asserted, not demonstrated. To make matters worse, the ordering of the e1∨e3 and e2∨e3 basis vectors in the preliminary matrix does not match the final displayed matrix, so an independent check is harder than it should be. The same pattern appears in Proposition 4.15, where the claim that three systems of quadratics have no nontrivial real solutions is stated without the algebra. I don't think the theorem is false; the structural lemmas, the rank-of-R′ case split, and the kernel-inclusion extremality criterion are coherent. But the classification currently rests on unverified symbolic algebra, and that needs to be fixed in revision.\n\nThe physics side is honestly scoped. Footnote 3 restricts to tree-level, quartic-in-momentum, non-truncated cones without full crossing constraints, and the identification of C_W with the physical positivity cone is imported from Refs. [94,98]. The applications are conditional on that import, and the authors say so. No circularity beyond the usual.\n\nWho is this for: people working on positivity bounds in multi-field EFTs, and anyone who wants a worked example of spectrahedral extremal structure. It is long, but Section 5 delivers the payoffs.\n\nMy recommendation: send to peer review. The referee should ask for a displayed derivation of the Proposition 4.25 elimination—or an ancillary computer-algebra file—and a corrected matrix with consistent basis ordering. If that checks out, accept.","headline":"A credible first classification of the three-flavor positivity cone whose key symbolic elimination in Prop. 4.25 is asserted rather than shown—worth refereeing, with a request for a checkable derivation.","tokens_in":55063,"tokens_out":7261,"would_cite":true,"duration_ms":68021,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["52A20","15B48","81T12"],"pacs":[],"model":"deepseek-v4-flash","headline":"For three flavors, the forward-limit positivity cone $\\mathcal{C}_W$ has a complete classification: every extremal element is one of three explicit types, and the third type generates all inelastic bounds.","keywords":["positivity bounds","effective field theory","spectrahedron","extremal rays","convex cone","elastic bounds","inelastic bounds","three flavors"],"falsifier":"For a fixed basis of $V$, draw many positive semidefinite $S\\in\\mathcal{C}_W$ by taking random sums of tensor squares $Q_i\\otimes Q_i$ with $Q_i\\in\\mathrm{Sym}^2 V$ or $Q_i\\in\\Lambda^2 V$ and imposing $\\tau S=S$; then filter for four-dimensional kernel and test extremality through the kernel-inclusion criterion. Any extremal $S$ outside the three families of Theorem 4.1 would refute the classification.","tokens_in":53987,"feed_emoji":"🧮","tokens_out":8216,"duration_ms":83941,"temperature":0.7,"pith_summary":"Positivity bounds restrict the Wilson coefficients of any effective field theory that must come from a unitary, local, causal quantum theory. This paper solves the geometric core of those bounds for three particle flavors: it classifies every extremal ray of the forward-limit positivity cone $\\mathcal{C}_W$, the cone of positive semidefinite four-tensors with the symmetries of a crossing-symmetric amplitude. The classification has exactly three families, and the third family produces the inelastic bounds that cannot be reached from elastic scattering alone. The same theorem yields complete positivity bounds for three-flavor amplitudes, and it shows that amplitudes with $O(3)$, $\\mathbb{Z}_2^3$, or $SO(2)$ symmetry are fully constrained by elastic bounds alone.","feed_headline":"Three shapes generate every three-flavor positivity bound","feed_subtitle":"A complete theorem lists the cone's extremal rays; the new inelastic family goes beyond elastic-only arguments.","key_machinery":"The object doing the work is the linear spectrahedron $\\mathcal{C}_W = \\{S\\in W : S\\ge 0\\}$, where $W$ is the space of four-tensors in $\\mathrm{Sym}^2(\\mathrm{Sym}^2 V^*)\\oplus \\mathrm{Sym}^2(\\Lambda^2 V^*)$ that are invariant under the transposition $\\tau$. Extremality is tested through the minimal-face characterization: $S$ is extremal precisely when no non-zero $S'\\in\\mathcal{C}_W$ has kernel strictly containing $\\ker S$. The proof organizes $\\mathcal{C}_W$ by the rank of the curvature-type restriction $R'(S)$ and, for rank 1, by the dimension of $\\ker S\\cap (z\\vee V)$; the surviving case $\\operatorname{rank} R'(S)=1$, $\\operatorname{rank} R(S)=3$, and $\\ker S\\cap (z\\vee V)=\\{0\\}$ forces the explicit three-parameter kernel basis of Proposition 4.25, which is then converted into the closed form of Theorem 4.1.","core_discovery":"The paper's central claim is Theorem 4.1: when $\\dim V = 3$, an element $S\\in\\mathcal{C}_W$ is extremal if and only if it is one of (1) $S = \\alpha^4$ for $\\alpha\\in V^*\\setminus\\{0\\}$; (2) $S = (\\alpha_1\\vee\\alpha_2)^{\\otimes 2} + (\\alpha_1\\wedge\\alpha_2)^{\\otimes 2}$ for linearly independent $\\alpha_1,\\alpha_2$; or (3) $S = S_{\\mathrm{tot}} + 2(g^2+d^2-1-dh)\\big((\\alpha_2\\otimes\\alpha_3)^{\\otimes 2} + (\\alpha_3\\otimes\\alpha_2)^{\\otimes 2}\\big)$ for a basis $\\{\\alpha_1,\\alpha_2,\\alpha_3\\}$ and parameters satisfying $g^2 > 1-d^2+dh$, with $S_{\\mathrm{tot}}$ the displayed totally symmetric quartic. Types 1 and 2 are exactly the elastic bounds $M(\\alpha\\otimes\\beta,\\alpha\\otimes\\beta)\\ge 0$; type 3 is the inelastic family, so every inelastic bound for three flavors comes from this family. The proof also establishes that in the symmetric cases $O(3)$, $\\mathbb{Z}_2^3$, and $SO(2)$, every extremal ray of the invariant cone is the projection of an elastic extremal ray, so elastic bounds alone cut out the full invariant cone.","pith_inferences":["A natural stress test is to push the same rank-of-$R'$ strategy to $n=4$: nothing in the kernel-dimensionality arguments is obviously four-flavor-specific, and the growth of $W$ suggests additional inelastic families beyond the three.","The $O(3)/\\mathbb{Z}_2^3/SO(2)$ results suggest a pattern: when the symmetry group is large enough that projecting an inelastic extremal ray lowers its rank, elastic bounds become sufficient; finding the threshold symmetry group would be a clean follow-up.","The explicit type-3 family gives a concrete target for numerical searches: sample the 21-dimensional three-flavor tensor space near tensors that annihilate an inelastic ray, since the paper notes uninformed random directions rarely expose the elastic/inelastic gap."],"forward_implications":["For three flavors, the complete set of positivity bounds is now explicit: elastic bounds from types 1 and 2 plus the inelastic inequalities from type 3, so no further extremal rays are missing.","Any amplitude with $O(3)$, $\\mathbb{Z}_2^3$, or $SO(2)$ symmetry is fully constrained by elastic bounds, so inelastic bounds add nothing in those symmetric sectors.","In chiral perturbation theory for pions, the positivity cone reduces to the elastic conditions $\\ell_2\\ge -\\ell_1$ and $\\ell_2\\ge 0$, reproducing the bounds known from earlier dispersion-relation analyses.","Without symmetry, the elastic cone is strictly smaller than the full cone, so there exist three-flavor Wilson-coefficient tensors that pass every elastic bound yet are excluded by an inelastic bound.","The paper recovers the known two-flavor result as a special case: there the third family degenerates and all extremal rays are elastic."],"supporting_citations":[{"why":"defines the cone $C = \\mathrm{conv}\\{Q^{\\otimes 2}+\\tau Q^{\\otimes 2}\\}$ and identifies its dual with $\\mathcal{C}_W$, supplying the central geometric translation of positivity.","marker":"[94]"},{"why":"introduced the dual-cone description $\\mathcal{C}_W=\\{S\\in W:S\\ge 0\\}$, solved $\\dim V=2$, and linked extremal rays to elastic and inelastic bounds; this paper extends that framework to $\\dim V=3$.","marker":"[98]"},{"why":"established that unitarity, locality, and causality imply positivity bounds on Wilson coefficients from dispersion relations, giving the physical meaning of the cone.","marker":"[4]"},{"why":"provides the convex-geometry foundations used here: faces, relative interiors, extremal representation, and Carathéodory's theorem.","marker":"[99]"},{"why":"supplies the theorem that a closed, line-free convex cone is determined by its extremal rays, which is why the classification yields all bounds.","marker":"[100]"},{"why":"gives the minimal-face and kernel characterization of spectrahedra used to test extremality throughout the proof.","marker":"[101]"},{"why":"prior chiral perturbation theory positivity bounds that the $O(3)$ application reproduces.","marker":"[51]"},{"why":"standard result on $O(3)$-invariant tensors used in the symmetric-amplitude section.","marker":"[103]"}],"fun_headline_variants":["Three families classify all three-flavor positivity bounds","Inelastic bounds join elastic to complete three-flavor cone","Three extremal families fully classify three-flavor bounds","Complete three-flavor positivity cone: three families suffice","Three shapes span all three-flavor positivity bounds"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the four-tensor extracted from forward two-to-two amplitudes lies in $\\mathcal{C}_W^*$ whenever the UV completion is unitary, local, and causal—an identification imported from earlier work and only sketched in Appendix A.1—together with the paper's explicit restriction to tree-level, quartic-in-momentum, non-truncated cones without full crossing constraints.","fun_headline_variants_meta":{"raw":{"variants":["Three families classify all three-flavor positivity bounds","Inelastic bounds join elastic to complete three-flavor cone","Three extremal families fully classify three-flavor bounds","Complete three-flavor positivity cone: three families suffice","Three shapes span all three-flavor positivity bounds"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000463,"raw_usage":{"total_tokens":2409,"prompt_tokens":1138,"completion_tokens":1271,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":754,"completion_tokens_details":{"reasoning_tokens":1196}},"tokens_in":754,"tokens_out":1271,"duration_ms":10839,"temperature":1.0,"reasoning_tokens":1196,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T16:58:04.284516+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a fixed basis of $V$, draw many positive semidefinite $S\\in\\mathcal{C}_W$ by taking random sums of tensor squares $Q_i\\otimes Q_i$ with $Q_i\\in\\mathrm{Sym}^2 V$ or $Q_i\\in\\Lambda^2 V$ and imposing $\\tau S=S$; then filter for four-dimensional kernel and test extremality through the kernel-inclusion criterion. Any extremal $S$ outside the three families of Theorem 4.1 would refute the classification.","supporting_citations":[],"review_version":2}