{"id":"f8a55d5a-6dda-4b3a-b225-be6846111c96","arxiv_id":"1908.06928","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Over infinite fields, every trace on the rational points of a quasi-simple algebraic group is a convex combination of the trivial character and characters supported on the center.","lead":"This paper proves that for the group of rational points of a simple algebraic group over an infinite field, every symmetric positive definite function is a trivial combination of the constant function and functions living on the center. The result classifies the possible invariant random subgroups and ergodic actions of these groups, which are central objects in group theory and operator algebras.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 2.2's uncountable spectral summation is not justified and can fail for diffuse measures, so the proof of the mixing lemma is incomplete.","rationale":"The reader's verdict of CONDITIONAL is appropriate, and the reader's rationale already flags the uncountable Borel partition summation in Proposition 2.2. However, the reader's formal weakest assumption points to the external classification [PT16, Theorem 2.4]. That external theorem is likely correct for arbitrary infinite fields and is quoted as such; the more fragile, internal step is the identity in equation (2), which is not a general property of projection-valued measures for uncountable partitions. The failure is concrete: diffuse spectral measures give E({x}) = 0 for all x even though the singletons partition the space, so the SOT sum can vanish while E(X) is the identity. This does not disprove Proposition 2.2, because a different partition (e.g., a countable one obtained from a compactness or Rokhlin argument) may restore the proof, but the manuscript does not supply such an argument. Since the proposition is used multiple times in the proof of Theorem A, the conditional status is justified. The paper's main theorem is plausible and the external dependency on [PT16] is legitimate, so I do not recommend rejection or a stronger verdict than CONDITIONAL; the requested revision is to make the summation in Proposition 2.2 rigorous, either by restricting to countable partitions or by proving the SOT identity for the specific spectral measure.","tokens_in":17642,"tokens_out":63779,"duration_ms":708665,"concrete_test":"Run the argument of Proposition 2.2 for the concrete case G = Z ⋉ Z with U = Z, H = Z acting by an irrational rotation h, and π the regular representation on ℓ^2(G). The spectral measure is Lebesgue measure on T. Take the Borel partition of T into singletons, which satisfies Lemma 2.3. Equation (2) claims ∑ E({x}) = I, but E({x}) = 0 for every x, so the equality is false; this isolates exactly the unjustified step. Then check whether the proof can be repaired by choosing a countable partition of T with B_i ∩ h(B_i) = ∅ (for instance two half-circles), verifying that (2) holds by countable additivity and that the rest of the mixing argument survives. If no countable Borel partition exists in the general setting of Prop 2.2, the mixing lemma and hence Theorem A remain unsupported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of the crucial mixing result, Proposition 2.2, rests on the assertion in equation (2) that for the Borel partition (B_i)_{i∈I} of X = \\hat{U}\\setminus\\{1\\} obtained from Lemma 2.3, one has ∑_{i∈I} E(B_i) = E(X) in the strong operator topology, and subsequently on expanding inner products into double sums over I. This is not a general property of projection-valued measures. In fact, it fails for diffuse spectral measures: for U = Z with π the regular representation on L^2(T) and E(B) the multiplication operator by 1_B, the partition of T into singletons satisfies the condition B_i ∩ h(B_i) = ∅ for any fixed-point-free rotation h, but E({x}) = 0 for every x, so the left side of (2) is 0 while E(T) = I. Lemma 2.3 only supplies some Borel partition; it does not ensure a countable one, nor that the uncountable sum of E(B_i) equals E(X). Since the mixing conclusion of Proposition 2.2 is used repeatedly in Steps 1, 3, and 4 of Section 4.1 to prove Theorem A, the central theorem is not established as written. The reliance on [PT16, Theorem 2.4] is a real external dependency, but the internal summation gap is more immediate and more damaging to the proof.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves a character rigidity theorem for quasi-simple algebraic groups over infinite fields. Specifically, if k is infinite and G is quasi-simple over k, every normalized positive-definite class function on G(k) is a convex combination of the function identically equal to 1 on G(k)^+ and a function supported on the center Z(G)(k). The main corollary is that the only characters of PG(k)^+ are the trivial character and delta_e. From this the author derives rigidity of factor representations, a classification of ergodic invariant random subgroups when k is countable, and a dichotomy for measure-preserving ergodic actions of G(k) on probability spaces when k is a global field. The proof strategy is a Howe-Moore-type analysis: a general mixing lemma for traces on semidirect products (Proposition 2.2) is combined with root-space structure, the Bruhat decomposition, Tits' simplicity theorem, and the Peterson-Thom classification of characters of SL2(k).","tokens_in":17899,"tokens_out":18285,"duration_ms":192998,"significance":"If the result stands, it is a substantial and natural extension of Tits' simplicity theorem and of earlier character-rigidity results of Kirillov, Ovchinnikov, and Peterson-Thom. The paper gives a clean dichotomy for all quasi-simple groups over infinite fields, with consequences for IRS rigidity and for ergodic actions of S-arithmetic groups that go beyond the previously known lattice cases. The proof is largely self-contained on the group-theoretic side and does not rely on fitted parameters or circular assumptions. However, the proof of the central mixing proposition contains a genuine analytical gap involving uncountable spectral sums, and the argument depends in an essential way on the external Peterson-Thom classification. The main theorem is likely true, but it is not established as written.","major_comments":[{"comment":"Equation (2) asserts that sum_{i in I} E(B_i) = E(X) in the strong operator topology for the partition produced by Lemma 2.3. Lemma 2.3 only guarantees a pairwise disjoint Borel partition; it does not guarantee that I is countable, and for projection-valued measures uncountable summation over a partition is not generally valid. For example, for the regular representation of Z on L^2(T), with E(B) equal to multiplication by 1_B and the partition of T\\{1} into singletons, every E(B_i) is zero while E(X)=I. Thus the expansion of the inner product into the double sum over I is unjustified. Since Proposition 2.2 is applied in Steps 1, 3, and 4 of Section 4.1, Theorem A is not established as written. The gap is local and likely repairable: one should prove that the partition can be chosen countable (or finite) in the relevant compact dual, or otherwise justify the uncountable sum by a stronger property of the projection-valued measure.","section":"Section 2, Proposition 2.2, Eq. (2)"},{"comment":"The proof relies on [PT16, Theorem 2.4] for SL2(k) over an arbitrary infinite field and applies it to L = rho_alpha(SL2(k)) for every non-multipliable root alpha. The theorem should be stated precisely, including its hypotheses on the field and on the center, and the deduction that pi|_L is mixing on the orthogonal complement of H^L should be expanded; as written, the single sentence does not make clear why the finite-center-supported component cannot contribute non-mixing finite-dimensional pieces. This is a load-bearing external dependency and should be verified explicitly.","section":"Section 4.1, first step"}],"minor_comments":[{"comment":"The phrase 'such that such that' is duplicated in the proof of Proposition 2.2; please remove the repetition.","section":"Section 2, Proposition 2.2"},{"comment":"The expression 'h in H /i⋉tegerdivideF' appears to be a corrupted version of 'h in H \\ F'; please correct the typesetting.","section":"Section 2, Proposition 2.2 proof"},{"comment":"In the decomposition phi = phi_1 + phi_2, the functions phi_i are not normalized; the statement of Theorem A speaks of a convex combination, so the proof should explicitly renormalize by ||xi_i||^2. This is a minor presentation issue because the intended normalization is clear.","section":"Section 4.1, beginning of proof of Theorem A"}],"recommendation":"major_revision","confidential_remarks":"The spectral-summation gap in Proposition 2.2 is the main obstacle to accepting the proof. It appears fixable by strengthening Lemma 2.3 to produce a countable wandering partition, but the author must supply the argument. The external dependence on [PT16] should also be checked against the precise statement of that theorem; if the classification is known only in a narrower setting, the main theorem would need additional hypotheses."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The main theorem is new and worth taking seriously: character rigidity for all quasi-simple groups over infinite fields, not just SL_n or split Chevalley groups. The treatment of relative root systems and centers of unipotent radicals is a substantial piece of algebraic-group work, and the corollaries on invariant random subgroups and ergodic actions are natural and valuable. The reliance on Peterson–Thom's classification for SL_2(k) is legitimate and clearly flagged; that is an external dependency, not a hidden assumption.\n\nThe stumbling block is in Proposition 2.2. The proof asserts that for a Borel partition (B_i) of X = Ĥ \\ {1}, one has Σ_i E(B_i) = E(X) in the strong operator topology. That is not a general property of projection-valued measures. It holds for countable partitions, but the partition from Lemma 2.3 is not guaranteed to be countable. The paper's argument breaks precisely at the uncountable summation. A concrete failure is the regular representation of Z on L²(T), where the singleton partition of T \\ {1} gives E({x}) = 0 pointwise but E(T \\ {1}) = I. Lemma 2.3 only produces some Borel partition; it does nothing to ensure the spectral sum adds up correctly.\n\nThis is not a cosmetic issue. Proposition 2.2 is the mixing result used in Steps 1, 3, and 4 of Section 4.1, so the proof of Theorem A is incomplete as written. That said, I think the gap is likely repairable: in the actual applications, U is the k-points of a root group, and its dual should admit a countable Borel partition separating points from their translates, or one can argue directly from the structure of the spectral measure. But the paper must supply that argument.\n\nIf Proposition 2.2 can be fixed, the rest of the proof looks solid. The later algebraic steps are careful, and the corollaries follow cleanly. The paper deserves a serious referee, but not acceptance in its current form. I would ask for a rewritten proof of Proposition 2.2 before recommending publication.\n\nFor researchers in character rigidity, IRS, and actions of algebraic groups, this is clearly relevant. I would bring it to a reading group and would cite it once the gap is closed.","headline":"Strong, genuinely new theorem in character rigidity; the proof has a real gap in Proposition 2.2 that needs patching before the result is established.","tokens_in":18397,"tokens_out":4836,"would_cite":true,"duration_ms":53578,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["20G05","22D10","22D25","22D40"],"pacs":[],"model":"deepseek-v4-flash","headline":"For any infinite field k and any quasi-simple k-group G, the group PG(k)+ has exactly two characters: the trivial character and the Dirac delta at the identity.","keywords":["character rigidity","traces","positive definite functions","quasi-simple algebraic groups","Tits simplicity theorem","invariant random subgroups","ergodic actions","root subgroups"],"falsifier":"Look for a counterexample over a non-perfect infinite field: verify whether the classification of characters of SL2(F_q((t))) described in [PT16, Theorem 2.4] is complete; any extra trace there would break the first step of the proof, and exhibiting any nontrivial character of PG(k)+ over such a field would refute the theorem.","tokens_in":17434,"feed_emoji":"♾️","tokens_out":11474,"duration_ms":103869,"temperature":0.7,"pith_summary":"The paper establishes a rigidity theorem for characters of groups of rational points of simple algebraic groups. For any infinite field k and any quasi-simple k-group G, every normalized central positive-definite function on the subgroup G(k)+ generated by unipotent radicals is a convex combination of the constant function and a function supported on the center. In particular the quotient PG(k)+ has exactly two characters, the trivial one and the Dirac delta. This extends Tits' classical simplicity theorem from normal subgroups to traces, and it yields sharp consequences for invariant random subgroups, finite factor representations, and ergodic actions.","feed_headline":"Only two characters remain for simple groups over infinite fields","feed_subtitle":"The rigidity forces every ergodic invariant random subgroup to be trivial or the whole group.","key_machinery":"The central mechanism is a mixing lemma (Proposition 2.2): if a group G has a normal subgroup G+, a subgroup H, and an abelian subgroup U normalized by H such that no nontrivial character of U is fixed by any element of H outside a finite set, then the restriction of any G+-invariant positive-definite function to H is mixing on the orthogonal complement of the U-invariant vectors. The proof applies this to pairs (H, U) = (H_lambda, U(alpha)(k)), where H_lambda is a one-parameter subgroup of the maximal k-split torus and U(alpha)(k) is a root subgroup, using the classification of characters of SL2(k) to obtain the no-fixed-character condition. Iterating through the root-space structure of the group, the Bruhat decomposition, and the centers of unipotent radicals, and invoking Tits' simplicity theorem at key steps, the argument eliminates all nonzero values of the trace outside the center.","core_discovery":"Theorem A states that if k is an infinite field, G is quasi-simple over k, and $\\varphi$ is a normalized function of positive type on G(k) that is invariant under conjugation by G(k)+, then $\\varphi$ is a convex combination of a function equal to 1 on G(k)+ and a function supported on Z(G)(k). The central consequence is character rigidity for the simple quotient PG(k)+: its only characters are the constant function 1 and the Dirac mass delta_e. Over countable k this forces the only ergodic invariant random subgroups of PG(k)+ to be the whole group and the trivial subgroup, and over global fields with trivial center it forces every ergodic measure-preserving action of G(k) either to factor through the abelianization or to be essentially free.","pith_inferences":["The countable assumption in the invariant-random-subgroup corollary is only needed to pass from traces to measures; one could try to remove it by an approximation argument, since the main theorem itself has no countability hypothesis.","Because the proof's first step uses the SL2(k) character classification at every root subgroup, a natural companion project is to verify that classification for non-perfect infinite fields such as F_q((t)); any failure there would become the first obstruction to the theorem's full generality.","The global-field corollary can be read as a dichotomy: if the Whitehead group G(k)/G(k)+ is abelian, then the only possible obstruction to freeness of an ergodic action is the abelianization, so nontrivial stabilizer behavior would have to be detected by that abelian quotient."],"forward_implications":["When k is countable, the only ergodic invariant random subgroups of PG(k)+ are the whole group PG(k)+ and the trivial subgroup {e}.","Every finite-factor representation of PG(k)+ is either trivial or equivalent to the regular representation; in particular, the group von Neumann algebra L(PG(k)+) is the only nontrivial finite-trace factor generated by PG(k)+.","For k a local or global field and G k-isotropic and almost k-simple, Char(G(k)) equals the dual of the abelianization G(k)^ab together with characters lifted from the center, so the character theory is fully controlled by these two quotients.","For k a global field and G with trivial center, every ergodic measure-preserving action of G(k) either factors through the abelianization of G(k) or is essentially free.","The rigidity extends to semisimple groups: every character of G(k)+ is a tensor product of characters of the almost-simple k-isotropic factors, pushed through a finite kernel."],"supporting_citations":[{"why":"Supplies Tits' simplicity theorem that every subgroup of G(k) normalized by G(k)+ is either central or contains G(k)+, used to force invariant subspaces to be trivial.","marker":"[Tit64]"},{"why":"Classifies the characters of SL2(k) for infinite fields k, giving the mixing property for each root subgroup in the first step of Section 4.1.","marker":"[PT16, Theorem 2.4]"},{"why":"Provides the Bruhat decomposition of G(k), used in the ninth step to reduce elements outside the centralizer of the split torus to a normal form.","marker":"[Bor91, Theorem 21.15]"},{"why":"Describes the unipotent radical as a direct product of root subgroups, used throughout to decompose elements and prove vanishing on unipotent subgroups.","marker":"[BT65, Proposition 3.11]"},{"why":"Identifies the center of the unipotent radical, needed in the fifth step to handle root subgroups contained in the center.","marker":"[LP11, Proposition 8.3]"},{"why":"Gives Zariski density of the one-parameter subgroups H_lambda in the split torus, making the mixing arguments work over arbitrary infinite fields.","marker":"[BT73, Corollaire 6.8]"}],"fun_headline_variants":["Character rigidity: only two characters for simple groups","Simple groups: characters are just 1 and delta_e","Rigidity forces simple groups to have only trivial characters","Only trivial characters survive in simple groups"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole argument assumes that a complete classification of characters of SL2(k) is known for every infinite field k, a result imported from [PT16, Theorem 2.4] and applied to every root subgroup; if that classification misses cases for some field, the proof cannot start.","fun_headline_variants_meta":{"raw":{"variants":["Character rigidity: only two characters for simple groups","Simple groups: characters are just 1 and delta_e","Rigidity forces simple groups to have only trivial characters","Only trivial characters survive in simple groups"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000328,"raw_usage":{"total_tokens":1815,"prompt_tokens":912,"completion_tokens":903,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":528,"completion_tokens_details":{"reasoning_tokens":842}},"tokens_in":528,"tokens_out":903,"duration_ms":8842,"temperature":1.0,"reasoning_tokens":842,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:33:05.978634+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Look for a counterexample over a non-perfect infinite field: verify whether the classification of characters of SL2(F_q((t))) described in [PT16, Theorem 2.4] is complete; any extra trace there would break the first step of the proof, and exhibiting any nontrivial character of PG(k)+ over such a field would refute the theorem.","supporting_citations":[],"review_version":1}