{"id":"ec621e16-e711-4d33-9007-e72824ba7cb4","arxiv_id":"2508.18685","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Tight spherical 5-designs of minimal type exist exactly when a specific Q-polynomial coherent configuration exists, and they cannot exist in infinitely many dimensions including 119 and 527.","lead":"This paper studies spherical 5-designs of minimal type, point sets on a sphere where one fixed vector sees only the values zero, plus one, and minus one. It proves structural equivalences for such designs, connects them to equiangular tight frames, and rules out their existence in dimensions 119 and 527.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.1(ii) states the wrong angle set for X1; the derived-code inner products are {a, -(√(d+2)+3)/(d-1)}, not {±a}, so condition (ii) as written is inconsistent with (i).","rationale":"The reader's weakest assumption correctly identifies the verification of condition (B2) and the Q-polynomial property in Theorem 3.1 as fragile, but the more decisive problem is an explicit algebraic error in the statement itself. The angle-set formula is a one-line calculation from Definition 3.1, so no numerical search or external software is needed to see it. Even if one generously treats the displayed ±a as a typo, the theorem as written is false and the reconstruction in (ii)⇒(i) does not go through: with the stated A(X1), the reconstructed set is not a tight spherical 5-design. This is load-bearing because Theorem 3.1 is the paper's main equivalence and feeds Corollary 3.1 and the ETF/strongly regular graph applications. The nonexistence results in Theorem 3.2 and the density argument in Theorem 3.3 are logically separate, but the central advertised characterization is invalid as stated. A corrected statement may be salvageable, but the submitted text requires revision before the central claim can be accepted.","tokens_in":21195,"tokens_out":36974,"duration_ms":318382,"concrete_test":"Re-derive the inner products in L_{α,1}(D) for ⟨y,z⟩=±1/√(d+2): verify whether the second value equals −a or −(√(d+2)+3)/(d−1). Then test d=23 by computing the two within-eX1 inner products obtained from the displayed A(X1) and from the corrected set; only the corrected set yields {±1/√(d+2)}. If the corrected set occurs, verify that the eigenmatrices Q^{(1,1)} and Q^{(1,3)} in Theorem 3.1(iii) are consistent with it.","verdict_should_be":"REJECT","load_bearing_attack":"In Theorem 3.1(ii) the angle set for X1 is declared to be {±a} with a=(√(d+2)-3)/(d-1). Applying Definition 3.1 to a tight 5-design D with ⟨α,x⟩∈{0,±1}, two vectors y,z∈L_{α,1}(D) come from y,z∈D with ⟨α,y⟩=⟨α,z⟩=1 and ⟨y,z⟩=ε/√(d+2), ε=±1. Then ⟨L_{α,1}(y),L_{α,1}(z)⟩=[(d+2)⟨y,z⟩−3]/(d−1), giving a for ε=+1 and −(√(d+2)+3)/(d−1) for ε=−1. Thus the second derived angle is not −a. Since X3=−X1, we must have A(X1,X3)⊇−A(X1)∪{−1}; the set {±c,−1} listed in (ii) is compatible only with the corrected A(X1)={a,−c}, not with the displayed {±a}. Moreover, in the reconstruction (ii)⇒(i), using A(X1)={±a} gives within-eX1 inner products 3/(d+2)±(d−1)/(d+2)a = ±1/√(d+2) for the first sign only; the second sign yields (6−√(d+2))/(d+2). For d=23 these are respectively 1/5 and 1/25 instead of ±1/5, so the reconstructed D is not a maximal ETF and not a tight 5-design. The main equivalence therefore fails as stated unless the angle set is corrected, and the Q-polynomial eigenmatrices in (iii) must be checked against the corrected set.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies spherical 5-designs in R^d that admit a vector α with ⟨α,x⟩∈{0,±1} for every design point ('minimal type'). For tight spherical 5-designs, it claims an equivalence (Theorem 3.1) between existence of such a design, existence of three spherical 3-designs X1,X2,X3 in S^{d-2} with prescribed sizes and angle sets, and existence of a Q-polynomial coherent configuration of a specified type; the paper then derives from this an ETF(d-1,(d-1)(d+1)/3) and a strongly regular graph (Corollary 3.1). A second theorem (Theorem 3.2) gives arithmetic conditions under which no tight spherical 5-design of minimal type exists, yielding infinitude and the examples d=119 and 527. The final section gives an analogous equivalence for antipodal spherical 4-distance 5-designs (Theorem 4.1) and proves that tight spherical 7-designs are never of minimal type (Theorem 4.2).","tokens_in":21446,"tokens_out":14905,"duration_ms":128864,"significance":"If the structural theorem were correct, it would give a new characterization of minimal-type tight 5-designs and a sufficient mechanism for deriving ETFs with parameters (d-1,(d-1)(d+1)/3) and associated strongly regular graphs, speaking directly to Conjecture 1.1. The lattice-theoretic proof of Theorem 3.2 is a solid piece of work and delivers genuine nonexistence in infinitely many dimensions, including 119 and 527; Theorem 4.2 is a short, correct contradiction, and the square-sieve argument in Theorem 3.3 establishes infinitude of the exceptional dimensions. The paper is therefore potentially valuable, but the central equivalence is currently not correctly stated: the angle set of X1 in Theorem 3.1(ii) is wrong, and the proof of the Q-polynomial part of (iii) is largely omitted. These issues must be resolved before the structural claims can be accepted.","major_comments":[{"comment":"The displayed angle set A(X1)=A(X3)={±a} with a=(√(d+2)-3)/(d-1) is inconsistent with Definition 3.1. For y,z∈D with ⟨α,y⟩=⟨α,z⟩=1 and ⟨y,z⟩=±1/√(d+2), the definition gives ⟨L_{α,1}(y),L_{α,1}(z)⟩=((d+2)⟨y,z⟩-3)/(d-1), which equals a for ⟨y,z⟩=+1/√(d+2) and -(√(d+2)+3)/(d-1) for ⟨y,z⟩=-1/√(d+2); it is not -a. Moreover, since X3=-X1, consistency forces A(X1,X3)={-a,(√(d+2)+3)/(d-1),-1}, whereas the theorem lists {±(√(d+2)+3)/(d-1),-1}. The reconstruction (ii)⇒(i) fails with {±a}: within eX1 the second sign gives 3/(d+2)-(d-1)a/(d+2), which for d=23 equals 1/25 rather than -1/5. The statement and proof must be corrected to A(X1)={a,-(√(d+2)+3)/(d-1)} and the matching A(X1,X3), and the eigenmatrices in (iii) must be recomputed with the corrected set.","section":"§3.1, Theorem 3.1(ii)"},{"comment":"The verification that the matrices E^{(i,j)}_ℓ form a basis satisfying (B1)-(B4) and the Q-polynomial property is not carried out. The text says 'Using a similar analysis in [Suda22], one can show that condition (B2) holds' and 'Then we can check that the Q-polynomial property holds', but no calculation is shown and no theorem of [Suda22] is quoted that directly applies to this three-fiber configuration. Because (iii) is one of the three equivalent conditions in a main theorem, this is a load-bearing omission; the authors should either provide the full verification or state and prove a lemma that covers this configuration.","section":"§3.1, proof of (ii)⇒(iii)"},{"comment":"The strongly regular graph conclusion inherits the error in Theorem 3.1(ii). With A(X1)={±a} the two squared inner products of X1 are equal, so the cited [BGOY15, Proposition 3.2] (which requires a^2≠b^2) does not apply and the parameters in (12) do not follow. With the corrected set A(X1)={a,-(√(d+2)+3)/(d-1)} the two absolute values are distinct, but the srg parameters must then be recomputed from formula (13); the verification that the resulting parameters are exactly (12) is missing.","section":"Corollary 3.1(ii)"}],"minor_comments":[{"comment":"The displayed valency formulas contain unbalanced parentheses (for example in the expressions for p^{1,1,2} and p^{1,1,3}), and the word 'integal' should be 'integral'.","section":"Remark 4.1"},{"comment":"The typesetting of the Q-polynomial eigenmatrices is garbled; as printed it is difficult to determine the row and column structure. Please use clearly delimited matrix environments.","section":"Theorem 3.1(iii)"},{"comment":"The assertion that 'through exhaustive enumeration, one can verify' that no vector in S has the required inner products is stated without a reproducible certificate; please supply the code or a concise counting/linear-algebra argument.","section":"Example 4.2"},{"comment":"The formula for N(x,i,j) appears to omit an error term: it should be N(x,i,j)=x/(24i^2j^2)+O(1) when (i,j)=1 and 0 otherwise, before the Möbius summation.","section":"Theorem 3.3"},{"comment":"Reference [NV00] is dated (2020) in the bibliography, but the cited Journal de théorie des nombres de Bordeaux volume 12(2) is from 2000; please correct the year.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses an interesting and timely question, and the nonexistence machinery in Sections 3.2 and 4 is sound in broad outline. The main barrier is the incorrect angle set in Theorem 3.1(ii), which also invalidates Corollary 3.1(ii) as stated. This is fixable in revision, but it is substantive rather than cosmetic. I would also ask the editor to ensure that the omitted verification of (B2) and the Q-polynomial property, attributed to [Suda22] (one of the authors' own papers), is supplied in full or tied to a precise stated theorem; an equivocation here would leave the central equivalence unsupported."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know before spending time on this: the headline theorem has a concrete error. For X1 = L_{α,1}(D), the derived-code inner products are (ε√(d+2)−3)/(d−1) for ε = ±1, so the angle set is {a, −c} with a = (√(d+2)−3)/(d−1), c = (√(d+2)+3)/(d−1), not the printed {±a}. Since X3 = −X1, the cross-angle set A(X1,X3) is {−1, −a, c}, not {±c, −1}. Condition (ii) of Theorem 3.1 is therefore inconsistent with (i), and the main equivalence as stated is false. Corollary 3.1(ii) inherits the problem because the SRG parameters are computed from the wrong angles.\n\nThe paper is not without value. Theorem 3.2's lattice argument is clean: under the stated arithmetic hypotheses, Λ* = 1/2 Λ+, the lattice Γ is even, and the contradiction with ⟨α,α⟩ = (2m+1)/3 not integral is convincing. That gives a new infinite family of dimensions, including 119 and 527, where minimal type is impossible. The idea of encoding minimal type as a Q-polynomial coherent configuration is natural and worth pursuing.\n\nThe other soft spots are secondary but real. The proof of Theorem 3.1 leaves condition (B2) and the Q-polynomial property to 'similar analysis' and 'one can check'; given the angle-set mistake, those assertions cannot be trusted as written. Example 4.2's exhaustive enumeration is not reproducible without code or data, and the valency formulas in Remark 4.1 are stated without derivation.\n\nWho is this for? Specialists in spherical designs, equiangular tight frames, and algebraic combinatorics. The nonexistence result is a genuine contribution, and the framework might yield more once corrected. But the central theorem cannot be used in its current form. If a referee can repair the angle set and the eigenmatrices, this could become a solid paper; as submitted it is not.\n\nMy recommendation: do not desk-reject outright—Theorem 3.2 and the coherent-configuration connection deserve referee time—but send with a warning that the main equivalence needs substantive correction. I would expect major revision.","headline":"The main equivalence has a wrong angle set and is false as printed, but the lattice nonexistence theorem is sound and the paper deserves a careful revision.","tokens_in":22117,"tokens_out":8568,"would_cite":false,"duration_ms":77576,"reading_group":"maybe","serious_thinker":"no","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05B30","05E30"],"pacs":[],"model":"deepseek-v4-flash","headline":"A tight spherical 5-design of minimal type exists exactly when a three-layer coherent structure exists, and arithmetic rules out infinitely many dimensions.","keywords":["spherical 5-design","minimal type","tight spherical design","coherent configuration","Q-polynomial","equiangular tight frame","strongly perfect lattice","Levenstein bound"],"falsifier":"Take the known tight spherical 5-design of minimal type in $\\mathbb{R}^{23}$ and compute the three derived sets $X_1,X_2,X_3$ of Theorem 3.1(ii): if their sizes or angle sets differ from the stated ones, the equivalence fails. Alternatively, exhibit any configuration in $\\mathbb{R}^{119}$ with the sizes and angle sets of Theorem 3.1(ii); by the equivalence it would yield a minimal-type tight 5-design in a dimension that Theorem 3.2 declares impossible.","tokens_in":20853,"feed_emoji":"🎯","tokens_out":12004,"duration_ms":110749,"temperature":0.7,"pith_summary":"This paper studies spherical 5-designs of minimal type: finite point sets on the unit sphere whose averages match the sphere integrals for every polynomial of degree at most five, with the extra restriction that some fixed direction evaluates to only $0,\\pm 1$ on the points. The central result is an equivalence: for dimension $d>7$, a tight spherical 5-design is of minimal type exactly when its points decompose into three spherical 3-designs with prescribed sizes and pairwise angle sets, equivalently when the whole configuration is a specific $Q$-polynomial coherent configuration. From that structure the authors extract a smaller equiangular tight frame, giving a concrete sufficient route from maximal equiangular tight frames to frames one dimension down. They also prove an arithmetic nonexistence theorem: infinitely many dimensions, including $d=119$ and $d=527$, admit no tight spherical 5-design of minimal type. For antipodal spherical 4-distance 5-designs they give an analogous three-layer decomposition and use integrality of valencies to test known examples.","feed_headline":"One structure characterizes minimal-type tight 5-designs","feed_subtitle":"If true, these designs yield new equiangular tight frames and rule out dimensions 119 and 527.","key_machinery":"The load-bearing object is the derived decomposition of a tight design $D$ along a framing vector $\\alpha$. Given $\\alpha$ with $\\langle\\alpha,x\\rangle\\in\\{0,\\pm 1\\}$, the paper slices $D$ into rescaled layers $L_{\\alpha,\\beta}(D)$ for $\\beta=0,\\pm 1$, sitting in $S^{d-2}$; Theorem 2.1, a generalization of the classical derived-design lemma, makes each layer a spherical 3-design. The layers are then assembled into a coherent configuration -- a partition of all pairs into angle relations whose intersection numbers are well defined -- and the decisive object is the basis of idempotent matrices $E^{(i,j)}_\\ell$ built from characteristic matrices of the layers. The key check is that these matrices satisfy conditions (B1)-(B4) and the $Q$-polynomial property, which turns the design into a $Q$-polynomial coherent configuration of type $\\begin{pmatrix}3&2&3\\\\2&4&2\\\\3&2&3\\end{pmatrix}$ with explicitly listed second eigenmatrices. For nonexistence, the machinery is the rescaled lattice $\\Gamma=\\frac{1}{\\sqrt{2}}\\Lambda^+$ built from the maximal ETF half $X$; the quotient computation forces $\\Lambda^*=\\frac{1}{2}\\Lambda^+$, contradicting the rationality needed for minimal type.","core_discovery":"The paper's claim is that 'tight spherical 5-design of minimal type' is not an ad hoc condition but the shadow of a three-part combinatorial structure. Theorem 3.1 states that for $d>7$ the following exist simultaneously or not at all: a tight spherical 5-design in $S^{d-1}$ of minimal type; three spherical 3-designs $X_1,X_2,X_3\\subset S^{d-2}$ with $X_3=-X_1$, $X_2=-X_2$, the stated cardinalities $(d+1)(d+2)/6$, $2(d-1)(d+1)/3$, $(d+1)(d+2)/6$, and the stated angle sets (for instance $A(X_1)=A(X_3)=\\{(\\sqrt{d+2}-3)/(d-1),-(\\sqrt{d+2}-3)/(d-1)\\}$ and $A(X_2)=\\{1/\\sqrt{d+2},-1/\\sqrt{d+2},-1\\}$); and a $Q$-polynomial coherent configuration of the displayed type with the listed second eigenmatrices. A corollary is that half of the zero layer $L_{\\alpha,0}(D)$ is an equiangular tight frame with parameters $(d-1,(d-1)(d+1)/3)$, so a maximal ETF whose antipodal closure is a minimal-type design produces an ETF one dimension down; when $d=k^2-2$ with odd $k>3$, the same design also yields a strongly regular graph with explicit parameters. Theorem 3.2 then rules out minimal type for $d=(2m+1)^2-2$ whenever $m$ is odd, $m\\not\\equiv 1\\pmod 3$, $m(m+1)$ is free of odd prime square factors and $m+1$ is not a multiple of $8$; the proof runs through the lattice quotient $\\Gamma=\\frac{1}{\\sqrt{2}}\\Lambda^+$ and uses $\\Gamma^*/\\Gamma\\cong\\mathbb{Z}/2\\mathbb{Z}$ to force a contradiction with the required inner product of the framing vector.","pith_inferences":["The coherent-configuration formulation turns existence in open dimensions such as $d=223$, $287$, or $727$ into a finite algebraic search: the listed eigenmatrices determine all intersection numbers, so one could search for a $Q$-polynomial coherent configuration of the stated type before attempting any sphere-geometric construction.","The strong-regular-graph parameters of Corollary 3.1 offer a cheap test for the unknown rows of Table 1: if no strongly regular graph with those parameters exists in a candidate dimension, then no minimal-type tight spherical 5-design exists there, regardless of sphere geometry.","The lattice-duality obstruction behind Theorem 3.2 is probably not special to strength 5; analogous quotients should constrain minimal type for other tight designs, though the paper does not pursue that extension."],"forward_implications":["A minimal-type tight spherical 5-design in $\\mathbb{R}^{k^2-2}$ with odd $k>3$ forces an ETF with parameters $(k^2-3,(k^2-3)(k^2-1)/3)$ and a strongly regular graph with the parameters in Display (12) -- the first sufficient condition for one direction of the ETF equivalence conjecture.","Infinitely many allowed dimensions are nonexistence cases for minimal type: all $d=(2m+1)^2-2$ with $m$ odd, $m\\not\\equiv 1\\pmod 3$, $m(m+1)$ free of odd prime square factors, and $m+1\\not\\equiv 0\\pmod 8$, including $d=119$ and $d=527$.","For antipodal spherical 4-distance 5-designs, minimal type is equivalent to a three-layer slicing into spherical 3-designs of sizes $(d+2)n/(3d)$, $4(d-1)n/(3d)$, and $(d+2)n/(3d)$; integrality of the derived valencies becomes a testable obstruction, and known examples behave as Table 2 records.","Tight spherical 7-designs in $\\mathbb{R}^d$ with $d>1$ are never of minimal type, and the maximal-MUB design in $\\mathbb{R}^{16}$ is not minimal type."],"supporting_citations":[{"why":"supplies the Q-polynomial coherent configuration framework, the idempotent construction from characteristic matrices, and the 'similar analysis' used for condition (B2)","marker":"[Suda22]"},{"why":"provides the definition of spherical designs, characteristic matrices, the derived-code lemma Theorem 2.1 generalizes, and the absolute bound for tight designs","marker":"[DGS77]"},{"why":"records that tight spherical 5-designs are exactly antipodal closures of maximal ETFs and restricts their dimensions","marker":"[BB09b]"},{"why":"supplies the moment equations (9a)-(9b) used to derive the layer sizes and the lattice parameters for nonexistence","marker":"[BMV04]"},{"why":"provides Lemma 3.1 computing Gamma*/Gamma as Z/2Z, the key quotient used in Theorem 3.2","marker":"[NV13]"},{"why":"gives the minimal-type criterion for strongly perfect lattices and the relation <alpha,x> in {0,±1} linking lattices to spherical designs","marker":"[Ven01]"},{"why":"shows that two-distance tight frames produce strongly regular graphs, used in Corollary 3.1(ii)","marker":"[BGOY15]"},{"why":"shows unions of spherical designs form coherent configurations, the tool that starts the (ii) implies (iii) direction","marker":"[Suda10]"},{"why":"states the ETF equivalence conjecture whose (i) implies (ii) direction Corollary 3.1 addresses","marker":"[XXY21]"}],"fun_headline_variants":["Minimal-type tight 5-designs: a triple of designs characterizes them","Tight 5-designs of minimal type yield ETFs and rule out d=119, 527","Minimal-type 5-designs: existence tied to Q-polynomial structure","No minimal-type tight 5-designs for infinitely many dimensions","Spherical 5-designs: minimal type implies ETF construction"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The main equivalence rests on an unshown computational step: the matrices $E^{(i,j)}_\\ell$ defined in the proof of Theorem 3.1 are asserted, after 'a similar analysis' or 'we can check', to satisfy the multiplication, basis, and polynomial properties that define a $Q$-polynomial coherent configuration, and if that computation is wrong the equivalence collapses.","fun_headline_variants_meta":{"raw":{"variants":["Minimal-type tight 5-designs: a triple of designs characterizes them","Tight 5-designs of minimal type yield ETFs and rule out d=119, 527","Minimal-type 5-designs: existence tied to Q-polynomial structure","No minimal-type tight 5-designs for infinitely many dimensions","Spherical 5-designs: minimal type implies ETF construction"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001235,"raw_usage":{"total_tokens":5213,"prompt_tokens":1230,"completion_tokens":3983,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":846,"completion_tokens_details":{"reasoning_tokens":3883}},"tokens_in":846,"tokens_out":3983,"duration_ms":32159,"temperature":1.0,"reasoning_tokens":3883,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T16:54:21.617762+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the known tight spherical 5-design of minimal type in $\\mathbb{R}^{23}$ and compute the three derived sets $X_1,X_2,X_3$ of Theorem 3.1(ii): if their sizes or angle sets differ from the stated ones, the equivalence fails. Alternatively, exhibit any configuration in $\\mathbb{R}^{119}$ with the sizes and angle sets of Theorem 3.1(ii); by the equivalence it would yield a minimal-type tight 5-design in a dimension that Theorem 3.2 declares impossible.","supporting_citations":[],"review_version":2}