{"id":"8d886d2e-dd5e-4db0-a182-8c486cf83b05","arxiv_id":"2411.15011","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A function on a finite abelian group obeying three algebraic identities is shown to be exactly a Jacobi sum of a finite field, up to a single trivial exception.","lead":"The paper proves that any function on a finite abelian group satisfying three simple identities must be a classical Jacobi sum from a finite field. This gives a clean axiomatic characterization of Jacobi sums, with possible connections to tensor categories.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 5.5's S(0) normalization is off by a factor of m; the printed proof of associativity is therefore incomplete, though the intended argument is repairable.","rationale":"The central claim is a classification theorem: axioms (A)-(C) force J to be a Jacobi sum of a finite field. I traced the proof and found that the genuinely load-bearing step is the verification that the reconstructed addition ⊕ is associative (Lemma 5.6), which depends entirely on Lemma 5.5. Lemma 5.5's computation of S(0) contains a normalization error: the orthogonality sum ∑_y α(cy)β(y) equals α(c)mδ(αβ), so with S(w) defined as in the text, S(0) should be α(c)δ(αβ), not α(c)δ(αβ)/m. The printed derivation then uses this incorrect value to claim A = ..., and the factor of 1/m does not disappear as written. Consequently, Lemma 5.5 is not proven by the text, and the proof of the field structure is incomplete. However, the error is localized and appears typographical: using the correct S(0) and the identity J(α,α^{-1}) = δ(α) − α(c)/m, the claimed expression for A follows exactly. I therefore do not see a route to a counterexample to the theorem; the concern is about rigor of the written proof, not truth. The other issues noted by the reader: the trivial-group exception is explicitly excluded and honestly discussed in Remark 1.3(a); the reliance on (C) is legitimate, since Lemma 4.3 follows from (C) via the substitution (a,b,c,d) = (1,α,βχ^{-1},χ) together with symmetry (A); and the multiplication-by-zero rule in Section 5 is an evident typo. Thus the reader's CONDITIONAL verdict is appropriate and unchanged.","tokens_in":8284,"tokens_out":33976,"duration_ms":270302,"concrete_test":"Recompute Lemma 5.5 from first principles: set S(0) = (1/m)∑_{x⊕y=0} α(x)β(y) = α(c)δ(αβ) (without the erroneous denominator), substitute into A = (1/m)∑_{w⊕z=1, z≠0} S(w)γ(z), split the w=0 and w≠0 contributions, and verify that A equals J(α,β)J(αβ,γ) − J(α,α^{-1})δ(αβ) + δ(α)δ(β) using the identity J(α,α^{-1}) = δ(α) − α(c)/m. If the equality holds, the lemma is true and the proof is repairable; if not, the field-addition reconstruction in the first case fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of the main theorem hinges on Lemma 5.5, which is used to prove associativity of the reconstructed addition ⊕ in the first case (Section 5). As printed, the computation of S(0) is off by a factor of m. The text defines S(w) = (1/m)∑_{x⊕y=w} α(x)β(y) and then states S(0) = (1/m)∑_{y∈\\hat{M}} α(cy)β(y) = α(c)δ(αβ)/m. But orthogonality of characters gives ∑_y α(cy)β(y) = α(c)·m·δ(αβ), so the correct value is S(0) = α(c)δ(αβ), not α(c)δ(αβ)/m. This error propagates: the subsequent identity S(0) = δ(α)δ(β) − J(α,α^{-1})δ(αβ), and the claimed conclusion A = J(α,β)J(αβ,γ) − J(β,β^{-1})δ(αβ) + δ(α)δ(β), do not follow from the displayed formulas because the factor of 1/m cannot be eliminated. Since Lemma 5.6 (associativity) invokes Lemma 5.5, the reconstruction of the field addition is not rigorously established by the written proof. With the correct S(0), the intended computation does go through using J(α,α^{-1}) = δ(α) − α(c)/m, so this appears to be a remediable normalization typo rather than a fatal gap. Nevertheless, the central theorem is not fully proven as printed until this step is corrected.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes three elementary axioms (A), (B), and (C) for a complex-valued function J on a finite abelian group M, and proves that any such 'Jacobi function' on a nontrivial M arises from a unique finite field structure on the set F = M-hat union {0}, with J equal to the Jacobi sum associated to that field. The proof reconstructs the addition law by Fourier analysis from the function J, then verifies the field axioms. The classical Jacobi sum is shown to satisfy the axioms, and the trivial group is identified as a genuine exception (the 'boolean semi-ring' case).","tokens_in":1432,"tokens_out":2548,"duration_ms":177232,"significance":"If the proof is made fully rigorous, this is an appealing and self-contained converse: three simple identities characterize Jacobi sums among all functions on finite abelian groups. The Fourier-analytic reconstruction of the field addition is elegant, and the paper explicitly verifies that classical Jacobi sums satisfy the axioms, making the characterization concrete. The result also connects to the author's work on oligomorphic groups and tensor categories. The proof is not machine-checked, but the intended argument is conceptually sound; however, two local but load-bearing slips in Section 5 need correction before the theorem is fully established.","major_comments":[{"comment":"The computation of S(0) is off by a factor of m. The text defines S(0) = (1/m) times the sum over y in M-hat of alpha(cy) beta(y) and then states S(0) = alpha(c) delta(alpha beta) divided by m. But orthogonality gives the sum over y of alpha(cy) beta(y) equal to alpha(c) times m times delta(alpha beta), so the correct value is S(0) = alpha(c) delta(alpha beta). Consequently the displayed identity 'S(0) = delta(alpha) delta(beta) - J(alpha, alpha^{-1}) delta(alpha beta)' is false as written; the right-hand side equals alpha(c) delta(alpha beta) divided by m, not alpha(c) delta(alpha beta). The subsequent derivation of A = J(alpha,beta) J(alpha beta,gamma) - J(beta,beta^{-1}) delta(alpha beta) + delta(alpha) delta(beta) is nonetheless the correct final result, because the w=0 contribution to A is (1/m) S(0), not S(0). This is a repairable normalization typo, but as printed the proof of Lemma 5.5 is internally inconsistent, and since Lemma 5.6 relies on Lemma 5.5 to prove associativity, the central theorem is not fully proven as written until this step is corrected.","section":"Section 5, Lemma 5.5"},{"comment":"The text extends multiplication to F by declaring '0 · x = x · 0 = x'. In a field with 0 as the additive identity, 0 must be absorbing for multiplication, i.e., 0 · x = 0. The printed definition makes 0 a second multiplicative identity, which is impossible for a nontrivial group M-hat and is inconsistent with Lemma 5.3(c): for x = 0, the element c · 0 would equal c under the printed definition, but then c ⊕ 0 = c, not the additive identity. The intended definition is almost certainly 0 · x = x · 0 = 0; with that correction Lemma 5.3(c) holds (0 ⊕ 0 = 0) and the classical case matches the usual field structure. This typo affects the statement of Proposition 5.1 and must be fixed.","section":"Section 5, definition of multiplication on F"}],"minor_comments":[{"comment":"The uniqueness assertion in Theorem 1.2 is not explicitly proved. It would be helpful to state that the set S and function i in Proposition 4.1 are uniquely determined by J (Fourier inversion is injective), and hence the constructed operation is forced, so the field structure is unique.","section":"Theorem 1.2"},{"comment":"After correcting the S(0) normalization, the line 'S(0) = delta(alpha) delta(beta) - J(alpha,alpha^{-1}) delta(alpha beta)' should be rewritten as '(1/m) S(0) = ...' to avoid confusion between the normalized inner sum and its contribution to A.","section":"Section 5, Lemma 5.5"},{"comment":"The symbol 0 is used both for the additive identity produced by Lemma 6.6 and for an element of the group M-hat in the proof of Lemma 6.7; this is confusing. Using a different symbol for the additive identity would improve clarity.","section":"Section 6, Lemmas 6.6 and 6.7"},{"comment":"The abstract says the characterization is 'very nearly' complete, and Remark 1.3(a) explains the trivial-group exception. It would be helpful to state this exception explicitly in the abstract or in the theorem statement itself.","section":"Introduction, Remark 1.3(a)"}],"recommendation":"major_revision","confidential_remarks":"The two errors in Section 5 are clearly local typos rather than fundamental gaps: the intended proof of Lemma 5.5 works after restoring the missing factor of m, and the multiplication on F should have 0 as absorbing. Both are easily corrected without changing the structure of the paper. I therefore recommend major revision rather than rejection. The paper is otherwise well written and the result is attractive; once the Section 5 slips are fixed, it should be publishable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Snowden proves that three elementary properties (symmetry, the multiplicative identity (B), and the orthogonality relation (C)) characterize classical Jacobi sums on any non-trivial finite abelian group. The converse is new; Ireland-Rosen, Nica, and the Harman–Snowden tensor-category papers don't have it. The proof is self-contained and uses Fourier inversion on the group to extract the set of pairs (x,y) with x+y=1 from J, then reconstructs the addition law of the field. That's a nice structural argument, and the classical case satisfies axioms by straightforward computation—no circularity, no fitted parameters.\n\nThe main theorem holds for m≥2, and the trivial group exception is explicitly acknowledged: there is a non-classical solution with J(1,1)=1, which is the 'boolean semiring' artifact. That is the right level of completeness.\n\nNow the soft spots. Lemma 5.5 contains a normalization error. With S(w) = (1/m)∑_{x⊕y=w} α(x)β(y), orthogonality gives S(0) = α(c)δ(αβ), not α(c)δ(αβ)/m as printed. The following line, S(0) = δ(α)δ(β) − J(α,α^{-1})δ(αβ), is actually the value of (1/m)S(0) after substituting J(α,α^{-1}) = δ(α) − α(c)/m. So the printed chain has the factor of m in the wrong place, and the associativity proof (Lemma 5.6) depends on it. But the intended computation goes through cleanly once S(0) is corrected: you get exactly A = J(α,β)J(αβ,γ) − J(α,α^{-1})δ(αβ) + δ(α)δ(β), and B analogously, so (B) gives equality. This is a repairable typo, not a fatal gap. There is also a slip in the definition of multiplication on F in Section 5: it says 0·x = x, which would make 0 an identity rather than an annihilator; the correct field axiom is 0·x = 0. The proof never actually uses that rule, so it is harmless.\n\nI agree with the reader's take. The theorem is within-subfield, but it is clean, new, and the proof strategy is worth knowing. The paper deserves a serious referee; after fixing the two typos it should be publishable. I would cite it if I worked on Jacobi sums or on the tensor-category side.","headline":"A novel and substantially correct characterization of Jacobi sums; the proof has minor normalization typos that should be fixed but are not fatal.","tokens_in":9131,"tokens_out":11989,"would_cite":true,"duration_ms":90097,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["11T24","11T30"],"pacs":[],"model":"deepseek-v4-flash","headline":"Three elementary axioms force every Jacobi-like pairing to come from a finite field.","keywords":["Jacobi sums","finite fields","character sums","Fourier inversion","convolution identity","characterization theorem","finite abelian groups","Pontryagin duality"],"falsifier":"Search for a counterexample among small cyclic groups: write $J$ as the Fourier sum in Proposition 4.1 with unknown support $S$ and bijection $i$, impose (A)–(C) as polynomial equations, and check whether any solution with $|M|\\ge2$ yields an $(S,i)$ that does not define a field.","tokens_in":8035,"feed_emoji":"➕","tokens_out":8905,"duration_ms":84371,"temperature":0.7,"pith_summary":"Jacobi sums are exponential-sum averages over pairs of nonzero field elements adding to one, and this paper sets out to show that their essential structure is purely axiomatic. The main theorem states that if a function on a nontrivial finite abelian group obeys three elementary identities—symmetry, a two-variable product relation, and a convolution relation—then the group is automatically the dual of the multiplicative group of a finite field, and the function is the Jacobi sum of that field. In other words, the concrete formulas of classical number theory are forced by abstract combinatorial conditions, with a unique field structure recovered from the function itself. The only degeneracy is the trivial group, where one extra solution survives and corresponds to a boolean semi-ring rather than a field.","feed_headline":"Three axioms force every Jacobi-like pairing to come from a field","feed_subtitle":"In any nontrivial finite abelian group, three simple identities reconstruct the missing field structure.","key_machinery":"The engine of the proof is the convolution identity $Q_\\alpha * Q_\\beta = Q_{\\alpha\\beta}$ for $Q_\\alpha(\\beta)=J(\\alpha\\beta^{-1},\\beta)$, which is exactly condition (C) rewritten. This identity turns the Fourier transform $\\widehat Q_\\alpha$ into a character of $M$ supported on a common set $S$, so $J$ can be expressed as a Fourier sum over $S$ with a bijection $i:S\\to\\widehat M$; the pair $(S,i)$ then encodes the prospective field addition. The explicit addition law $x\\oplus y=x\\,i(x/y)^{-1}$ (with the edge case $x\\oplus y=0$ when $x=cy$) is the bridge from the abstract axioms to a concrete field structure.","core_discovery":"On the paper's own terms, the central object is the field structure hidden inside a Jacobi function. Condition (C) is used to define auxiliary functions $Q_\\alpha(\\beta)=J(\\alpha\\beta^{-1},\\beta)$, whose convolution identity $Q_\\alpha*Q_\\beta=Q_{\\alpha\\beta}$ makes the Fourier transform of $Q_\\alpha$ factor. From that factorization the paper reconstructs a subset $S \\subseteq \\widehat M$ and a bijection $i:S\\to\\widehat M$, so that $J(\\alpha,\\beta)=\\frac1m\\sum_{x\\in S}\\alpha(i(x))\\beta(i(x)x^{-1})$; two cases remain, and the first yields an explicit addition law on $F=\\widehat M\\sqcup\\{0\\}$, namely $x\\oplus y=0$ when $x=cy$ and $x\\oplus y=x\\,i(x/y)^{-1}$ otherwise. Associativity of $\\oplus$ is proved by comparing two triple sums whose equality follows from condition (B), and the resulting structure is a field whose Jacobi sum is exactly $J$. Thus every Jacobi function on a nontrivial finite abelian group is classical, with the trivial group as the single near-exception.","pith_inferences":["The same axiomatic method should apply to other exponential sums defined by algebraic equations over finite rings: if an analogous convolution identity holds, the Fourier support will again carry a ring structure, giving characterizations for Kloosterman-type sums.","The boolean semi-ring exception suggests a degeneration $q\\to 1$ in which finite fields collapse to the boolean semi-ring; one can test this by taking a one-parameter family of Jacobi sums and passing to the limit.","Condition (C) is strong enough that it may imply condition (B) in many settings; checking whether (B) is redundant would simplify the axiomatization to two identities."],"forward_implications":["If the theorem is correct, a Jacobi function is never exotic: every solution on a nontrivial finite abelian group is the ordinary Jacobi sum of some finite field, so the axioms describe the whole class.","The field addition is recoverable algorithmically from $J$: compute $\\widehat Q_\\alpha$, read off the support $S$ and the bijection $i$, then define $\\oplus$; no search is required.","Precomposing a Jacobi function with a group automorphism or postcomposing with a Galois automorphism of $\\mathbf{C}$ does not produce a new type beyond the classical family; the paper shows each such twist is the Jacobi sum for a new field structure on the same underlying set.","The trivial group is the only place where the characterization fails; the extra solution $J(1,1)=1$ is interpreted as the Jacobi sum of the boolean semi-ring, so the theorem is nearly complete."],"supporting_citations":[{"why":"Supplies the standard evaluation $J(\\alpha,\\alpha^{-1})$ used in the verification of condition (B) and in Proposition 2.1.","marker":"[IR]"}],"fun_headline_variants":["Three elementary properties force any Jacobi-like map to be a Jacobi sum","Three axioms reconstruct a finite field hidden in any Jacobi-like pairing","Jacobi sums are pinned down by three axioms on finite abelian groups","A finite field emerges from three simple properties of Jacobi sums","Three axioms force every Jacobi-like pairing to be classical"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument depends on condition (C), the convolution identity $Q_\\alpha*Q_\\beta=Q_{\\alpha\\beta}$: without it the Fourier reconstruction of the field addition has no starting point, and the nontriviality assumption $m\\ge2$ is also essential because the trivial group has the extra boolean solution.","fun_headline_variants_meta":{"raw":{"variants":["Three elementary properties force any Jacobi-like map to be a Jacobi sum","Three axioms reconstruct a finite field hidden in any Jacobi-like pairing","Jacobi sums are pinned down by three axioms on finite abelian groups","A finite field emerges from three simple properties of Jacobi sums","Three axioms force every Jacobi-like pairing to be classical"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001441,"raw_usage":{"total_tokens":5789,"prompt_tokens":912,"completion_tokens":4877,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":528,"completion_tokens_details":{"reasoning_tokens":4786}},"tokens_in":528,"tokens_out":4877,"duration_ms":35000,"temperature":1.0,"reasoning_tokens":4786,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T14:39:20.259023+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Search for a counterexample among small cyclic groups: write $J$ as the Fourier sum in Proposition 4.1 with unknown support $S$ and bijection $i$, impose (A)–(C) as polynomial equations, and check whether any solution with $|M|\\ge2$ yields an $(S,i)$ that does not define a field.","supporting_citations":[],"review_version":1}