{"id":"963bfff2-3ebb-4029-b3d1-33a43e7caf24","arxiv_id":"1908.10354","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For non-even p, every minimizer of the p-frame energy on the sphere has support with empty interior, and for potentials with finitely many positive Gegenbauer coefficients a discrete minimizer always exists.","lead":"The paper studies how charges on a sphere arrange themselves to minimize interaction energy. It proves optimal arrangements cannot contain a solid region for p-frame energies, and that many smooth energies have finite-point optimizers.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 4.3's Laplacian identity (4.12) has an off-by-one dimension term; as printed the operator D(k) does not satisfy (4.15), so Theorem 1.3 lacks a valid written proof, though the error is locally repairable.","rationale":"The reader's weakest_assumption is exactly the load-bearing defect I find: the Laplace-Beltrami computation in equation (4.12) is off by one, and since Proposition 4.3's differential operator is built on that identity, the proof of Theorem 1.3 is not valid as printed. I checked that the corrected coefficient changes the constants in (4.13)-(4.15) but preserves the sign dichotomy required for the contradiction: for the corrected operator D(k)=Δ∏_{r=0}^{k-1}(Δ+(p-2r)(p+d-2r-2)), the final expression is ∏_{j=0}^{2k}(p-j) t^{p-2k-2}[(p-2k-1)-(p+d-2k-2)t²], which is strictly negative for p∈(2k,2k+1] and strictly positive for p∈(2k-1,2k), exactly as needed. This makes the error localized and repairable rather than fatal. The rest of the paper, especially Theorem 3.3 and Theorem 5.1, appears sound and unaffected. I found no circular reasoning, no fitted parameters, no missing data, and no overclaiming beyond the printed computational slip. Since the reader already assigned CONDITIONAL, my stress test does not change the verdict; it confirms that the condition is the correction of the Laplacian identity and the associated constants in Proposition 4.3.","tokens_in":21302,"tokens_out":21064,"duration_ms":198418,"concrete_test":"Verify (4.12) on S^1 (d=2): with t=cos θ, Δ t^p = d²/dθ² (cos^p θ) = p(p-1)t^{p-2}(1-t²) - p t^p = p(p-1)t^{p-2} - p² t^p, not p(p+1)t^p. Then recompute (4.15) with the corrected constants c_r=(p-2r)(p+d-2r-2) in D(k)=Δ∏_{r=0}^{k-1}(Δ+c_r), and check the sign of the bracket (p-2k-1)-(p+d-2k-2)t² for t∈[δ,1] for p in (2k-1,2k) and (2k,2k+1]. If the signs match the paper's classification, Theorem 1.3 is recovered with a repaired proof.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Proposition 4.3 is essential: Theorem 1.3 follows by combining Propositions 4.2 and 4.3, which are mutually contradictory. The proof of Proposition 4.3 constructs D(k) using equation (4.12), which states Δ_x <x,y>^p = p(p-1)<x,y>^{p-2} - p(p+d-1)<x,y>^p. On S^{d-1}, the Laplace-Beltrami operator applied to a zonal function of t=<x,y> is Δ = (1-t²)∂_t² - (d-1)t ∂_t, so the correct identity is p(p-1)t^{p-2} - p(p+d-2)t^p. Every downstream constant in (4.13)-(4.15) is shifted by one (e.g., p+d-1 should be p+d-2, p+d-3 should be p+d-4, and so on). Consequently, the operator D(k) as printed does not obey the displayed formula (4.15) for the true Laplacian, and the strict-sign condition (ii) used to obtain the contradiction (4.10) is not established by the written argument. The defect is localized rather than conceptual: replacing p+d-1 by p+d-2 throughout and re-running the induction gives D(k)t^p = ∏_{j=0}^{2k}(p-j) t^{p-2k-2} [(p-2k-1)-(p+d-2k-2)t²], whose sign on [δ,1] is strictly negative for p∈(2k,2k+1] and strictly positive for p∈(2k-1,2k), exactly as needed. Thus the central claim remains plausible, but the proof as printed is invalid at its key computational step.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies continuous energy minimization on the sphere. Its main results are: (i) Theorem 3.3, a quantitative existence theorem for discrete minimizers whenever the kernel has only finitely many positive Gegenbauer coefficients; (ii) Theorem 1.3, stating that for the p-frame energy f(t)=|t|^p with p>0 and p not an even integer, every minimizer has support with empty interior, partially confirming Conjecture 1.1; (iii) Theorem 5.1, an analogous empty-interior statement for real-analytic kernels that are not positive definite up to a constant, with discreteness on S^1; and (iv) Proposition 7.2, showing that local minimizers of positive-definite kernels are global. The paper is largely clearly written, with self-contained treatments of the moment-theoretic tools in Section 3 and classical reduction arguments elsewhere.","tokens_in":21657,"tokens_out":25008,"duration_ms":227827,"significance":"If the proof of Theorem 1.3 is repaired, this is a meaningful step toward Conjecture 1.1 and a substantive contribution to the study of attractive-repulsive potentials on spheres. Theorem 3.3 is a clean, quantitative extension of earlier weighted-design results, and Theorem 5.1 gives a useful dichotomy for analytic kernels. The paper also contains a pleasant observation on local versus global minimizers for positive-definite kernels. The main techniques are standard but are applied with care, and the manuscript is honest about the limits of its results. The one substantial defect is in the proof of Proposition 4.3, where the Laplacian identity is misstated; the error is localized and repairable, but as printed the proof of Theorem 1.3 is incomplete.","major_comments":[{"comment":"Equation (4.12) is incorrect. Since the Laplace–Beltrami operator on S^{d-1} acts on a zonal function of t=<x,y> as (1-t^2)∂_t^2-(d-1)t∂_t, the correct identity is Δ_x <x,y>^p = p(p-1)<x,y>^{p-2} - p(p+d-2)<x,y>^p, not p(p+d-1)<x,y>^p. This error propagates into (4.13)–(4.15): for example, the operator in (4.13) should use the factor p(p+d-2), and the bracket should read (p-3)-(p+d-4)t^2 rather than (p-3)-(p+d-3)t^2. As printed, the operator D(k) does not satisfy the displayed identity (4.15), and the strict-sign property (ii) used in the contradiction (4.10) is not established. The defect is localized: replacing p+d-1 by p+d-2 throughout and re-running the recurrence yields the bracket (p-2k-1)-(p+d-2k-2)t^2, whose sign on (δ,1] is exactly as claimed in the final paragraph of the proof. Nevertheless, the written proof of Proposition 4.3, and hence of Theorem 1.3, is incomplete as it stands.","section":"§4, Eq. (4.12)"},{"comment":"The definition of D(k) in (4.14) is not well specified as printed. The displayed product has unbalanced parentheses, and for k≥2 the factors do not match the iterative reduction that (4.15) claims. For instance, after the corrected Laplacian, the step from k=1 to k=2 requires a factor Δ+(p-2)(p+d-4), whereas the displayed first factor for k=2 contains the extra product p(p-1). I recommend giving an explicit recurrence, for example D(k)=(Δ+α_{k-1})D(k-1) with α_r=(p-2r)(p+d-2r-2) and D(0)=Δ, and verifying (4.15) from that recurrence. This would make the construction reproducible and would resolve the ambiguity in the present notation.","section":"§4, Eq. (4.14)"}],"minor_comments":[{"comment":"There is a typo in the phrase \"Propostion 4.3\" near the end of the proof; it should read \"Proposition 4.3.\"","section":"§4, proof of Proposition 4.3"},{"comment":"In the sentence beginning \"We now analyze the coefficient of εp\", the exponent should be typeset as ε^p for consistency with the surrounding equations.","section":"§4, after Eq. (4.8)"},{"comment":"The sentence \"Without loss of generality, we shall assume that 0∈N_+(f)\" would benefit from an explicit justification: adding a sufficiently large constant to f shifts only the n=0 Gegenbauer coefficient and does not change the minimization problem. As written, the WLOG is clear to an expert but is stated too abruptly.","section":"§3.2, proof of Theorem 3.3"},{"comment":"In the S^1 case, the proof shows that the support cannot have an accumulation point; since S^1 is compact, it would be helpful to state explicitly that compactness implies the support is finite, not merely discrete.","section":"§5, Theorem 5.1"}],"recommendation":"major_revision","confidential_remarks":"This is a strong paper and the central claim of Theorem 1.3 is plausible, with a local, repairable computational error in §4. I would not reject. The main request is to correct the Laplacian identity and give a clean recursive definition of D(k), then recheck the sign analysis. The companion preprint [BGM+] is cited for context and conjecture; that is acceptable, but the authors may want to update the reference if it has appeared by the time of revision. The paper fits the journal scope well."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. First, this is a genuinely substantial paper: it proves empty interior for p-frame minimizers with non-even p, gives a new existence result for discrete minimizers when the potential has finitely many positive Gegenbauer coefficients, and proves an analytic-kernel dichotomy. Second, the central proof has a concrete off-by-one error in the Laplacian computation that, as printed, invalidates Proposition 4.3; the error is local and repairable, but a referee should require the fix.\n\nWhat's new: Theorem 1.3 partially confirms the p-frame discreteness conjecture; Theorem 3.3 uses Karr's extreme point theorem to get discrete minimizers; Theorem 5.1 gives a clean dichotomy for analytic non-positive-definite kernels; Section 7's local-to-global observation for positive definite kernels is a nice bonus. The tools are appropriate, and the paper is honest about what remains open.\n\nThe soft spot is in Section 4, Proposition 4.3. The paper states Δ_x <x,y>^p = p(p-1)<x,y>^{p-2} - p(p+d-1)<x,y>^p. On S^{d-1}, for a zonal function of t=<x,y>, Δ = (1-t²)∂_t² - (d-1)t ∂_t, so the coefficient should be p+d-2, not p+d-1. The operator constants in (4.13)-(4.15) inherit the shift. As printed, D(k) does not satisfy (4.15), so the strict sign condition producing the contradiction (4.10) is not established. The stress-test note is right: replacing p+d-1 by p+d-2 throughout and rerunning the induction gives the required sign for p in the appropriate intervals. So the theorem is likely true, but the written proof of Proposition 4.3 is not valid as it stands.\n\nEverything else holds up: the Vandermonde moment vanishing in Proposition 4.2, the convexity argument in Theorem 3.3, the analytic continuation in Theorem 5.1, and the local-to-global argument in Section 7. No circularity, no fitted parameters, no overclaiming. The counterexample in Section 5 correctly explains why analyticity matters.\n\nWho this is for: anyone working on spherical energy minimization, frames, or design theory. It deserves a serious referee. My recommendation: engage with it, but send it back for a computational fix before acceptance.","headline":"A substantial, likely correct paper on discreteness of minimizing measures on spheres, but Proposition 4.3 contains a concrete off-by-one Laplacian error that needs a computational fix before the proof is valid.","tokens_in":22224,"tokens_out":2139,"would_cite":true,"duration_ms":19806,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["52A40","31E05","58C35","90C26"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that for every $p>0$ that is not an even integer, any minimizer of the $p$-frame energy on $S^{d-1}$ has support with empty interior, and that many potentials with finitely many positive Gegenbauer coefficients admit…","keywords":["p-frame energy","energy minimization on spheres","Gegenbauer polynomials","spherical harmonics","positive definite kernels","discrete minimizers","spherical designs","empty-interior support"],"falsifier":"Recompute $\\Delta_x\\langle x,y\\rangle^p$ on $S^{d-1}$; the standard value is $p(p-1)\\langle x,y\\rangle^{p-2}-p(p+d-2)\\langle x,y\\rangle^p$. Insert this into the operator $D^{(k)}$ and evaluate the expression in (4.15) at, say, $d=3$, $p=\\tfrac{5}{2}$, $t=\\langle x,y\\rangle=0.9$: for $p\\in(2,3]$ the quantity must be strictly negative. If instead it is positive or zero, Proposition 4.3 fails at its decisive step and Theorem 1.3 is not established by this argument.","tokens_in":21084,"feed_emoji":"🔵","tokens_out":18436,"duration_ms":162735,"temperature":0.7,"pith_summary":"The paper establishes that certain energy-minimizing probability measures on the sphere cannot be spread out. For the $p$-frame energy $I_f(\\mu)=\\int\\int |\\langle x,y\\rangle|^p\\,d\\mu(x)\\,d\\mu(y)$ with $p>0$ and $p\\notin 2\\mathbb{N}$, it proves that the support of any minimizer has empty interior: no minimizer can contain a spherical cap. This is a step toward the conjecture that every such minimizer is actually a finite discrete measure. The paper also proves that whenever a continuous potential has only finitely many positive Gegenbauer coefficients, a discrete minimizer always exists, with support size bounded by the dimension of the corresponding spherical-harmonic spaces; polynomial and real-analytic potentials are treated as special cases.","feed_headline":"Spherical p-frame minimizers have empty-interior support","feed_subtitle":"For every non-even p, the support of an optimal measure on a sphere contains no open set.","key_machinery":"The mechanism behind Theorem 1.3 is a family of differential operators $D^{(k)}$ on the sphere, built from the Laplace–Beltrami operator $\\Delta$ by a product of linear factors in $\\Delta$ designed so that, for $p$ in the interval $(2k-1,2k+1]\\setminus 2\\mathbb{N}$, applying $D^{(k)}$ to $\\langle x,y\\rangle^p$ has a fixed sign on the relevant region, while $D^{(k)}$ kills constants; combining these facts with the equilibrium identity $F_\\mu\\equiv I_f(\\mu)$ on $\\operatorname{supp}\\mu$ yields $0=\\int D^{(k)}_x\\langle x,y\\rangle^p\\,d\\mu(y)<0$, a contradiction. The discrete-minimizer theorem is carried instead by Karr's extreme-point theorem: among measures with prescribed spherical-harmonic moments, an extreme point has support of cardinality at most the number of constraints, and convexity of the negative-coefficient part of the energy pushes a minimizer onto an extreme point.","core_discovery":"On the paper's own terms, the central discovery is Theorem 1.3: fix $p>0$, $p\\notin 2\\mathbb{N}$, $f(t)=|t|^p$. If $\\mu\\in\\mathcal{P}(S^{d-1})$ minimizes $I_f$, then $(\\operatorname{supp}\\mu)^\\circ=\\varnothing$. The proof forces a contradiction from any interior point $z$: Proposition 4.2 shows $\\operatorname{supp}\\mu\\cap z^\\perp=\\varnothing$ by constructing, from points on a great circle through $z$, a matrix $[|\\langle x_i,x_j\\rangle|^p]$ that is not positive semidefinite, violating positive definiteness on the support; Proposition 4.3 shows $\\operatorname{supp}\\mu\\cap z^\\perp\\neq\\varnothing$ by applying a differential operator $D^{(k)}$ constructed from the Laplace–Beltrami operator, which annihilates constants and has a strict sign on $\\langle x,y\\rangle^p$, contradicting the constancy of the potential $F_\\mu$ on the support. A second discovery, Theorem 3.3, is that for any $f$ with finitely many positive Gegenbauer coefficients there exists a discrete minimizer whose support has at most $\\sum_{n\\in N_+(f)\\cup\\{0\\}}\\dim H_d^n$ points.","pith_inferences":["A natural next step suggested by the proof would be to adapt the differential-operator sign argument to other kernels with conjectured discretness, such as $\\arccos|t|$ or the causal-variational kernel; the obstacle is obtaining a tractable formula for $\\Delta_x f(\\langle x,y\\rangle)$.","The extreme-point theorem is constructive in principle: since a minimizer with at most $\\sum\\dim H_d^n$ support points exists, one could search for it by convex optimization over the moment polytope, which may give a practical algorithm for weighted designs with prescribed spherical-harmonic frequencies.","To reach Conjecture 1.1 one would need to rule out supports that are empty-interior but still infinite, such as Cantor-type sets; iterating the sign-contradiction on nested neighborhoods is a natural but nontrivial next step.","Because $|\\langle x,y\\rangle|^p$ behaves like a quadratic cusp $|x-y|^2$ at short distances, the paper's support restriction may be a spherical analogue of known results for mildly repulsive potentials at the endpoint case; testing the same question for $|x-y|^2$ interactions on $\\mathbb{R}^d$ would connect the two literatures."],"forward_implications":["For every non-even $p>0$, no minimizer of the $p$-frame energy can have an absolutely continuous part concentrated on an open region; in particular the normalized surface measure restricted to any spherical cap is never optimal.","For any polynomial potential with at least one negative Gegenbauer coefficient, all minimizers have empty-interior support, while a discrete minimizer with an explicit cardinality bound always exists.","For any real-analytic potential for which the uniform measure is not a minimizer, the support of every minimizer has empty interior; on the circle the support is finite.","Positive definiteness up to additive constants makes local and global minimizers coincide: any local minimizer of an even $p$-frame energy is global.","For potentials with finitely many positive Gegenbauer coefficients, there is always a discrete minimizer supported on at most $\\sum_{n\\in N_+(f)\\cup\\{0\\}}\\dim H_d^n$ points, generalizing the existence of weighted spherical designs."],"supporting_citations":[{"why":"Supplies the extreme-point theorem for moment-constrained measures, which gives the support cardinality bound in Theorem 3.3.","marker":"[K]"},{"why":"Gives Douglas's theorem that an extreme point's L1 span is the constraint span, used to prove Karr's theorem.","marker":"[D]"},{"why":"Establishes that the potential of a minimizer must be positive definite on the support, the engine of Proposition 4.2.","marker":"[Bj]"},{"why":"Provides the empty-interior support argument for causal variational principles that the proof of Theorem 1.3 adapts.","marker":"[FS]"},{"why":"Characterizes positive definite functions on spheres by nonnegative Gegenbauer coefficients, used to locate the sign contradiction.","marker":"[Sch]"},{"why":"Supplies the inverse Vandermonde formula used to build the test vector in Proposition 4.2.","marker":"[Kn]"},{"why":"Supplies the exponential-sum bound used to show the coefficient in Proposition 4.2 vanishes exactly at even integers.","marker":"[PS]"},{"why":"Gives the spherical-coordinate form of the Laplace–Beltrami operator used to construct the operators $D^{(k)}$.","marker":"[KMR]"},{"why":"States the discretness conjecture and tight-design minimization results that Theorem 1.3 is designed to approach.","marker":"[BGM+]"}],"fun_headline_variants":["Non-even p yields hollow support on spheres","Sphere minimizers shun open sets for non-even p","Discrete minimizers exist for broad energy class","Empty interior support for p-frame minimizers","p-frame energies: interior points eliminated"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise of the empty-interior theorem is a sign calculation: a differential operator built from the spherical Laplacian must make $|\\langle x,y\\rangle|^p$ strictly one-signed while killing constants; as printed, the key Laplacian formula has a dimension-dependent coefficient error ($p(p+d-1)$ where the standard computation gives $p(p+d-2)$), so the written proof must be repaired for the contradiction to be valid as stated.","fun_headline_variants_meta":{"raw":{"variants":["Non-even p yields hollow support on spheres","Sphere minimizers shun open sets for non-even p","Discrete minimizers exist for broad energy class","Empty interior support for p-frame minimizers","p-frame energies: interior points eliminated"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000217,"raw_usage":{"total_tokens":1418,"prompt_tokens":910,"completion_tokens":508,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":526,"completion_tokens_details":{"reasoning_tokens":439}},"tokens_in":526,"tokens_out":508,"duration_ms":5243,"temperature":1.0,"reasoning_tokens":439,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:49:18.434353+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute $\\Delta_x\\langle x,y\\rangle^p$ on $S^{d-1}$; the standard value is $p(p-1)\\langle x,y\\rangle^{p-2}-p(p+d-2)\\langle x,y\\rangle^p$. Insert this into the operator $D^{(k)}$ and evaluate the expression in (4.15) at, say, $d=3$, $p=\\tfrac{5}{2}$, $t=\\langle x,y\\rangle=0.9$: for $p\\in(2,3]$ the quantity must be strictly negative. If instead it is positive or zero, Proposition 4.3 fails at its decisive step and Theorem 1.3 is not established by this argument.","supporting_citations":[],"review_version":1}