{"id":"7c9bd92d-b21f-4d41-add5-2ba79b4a2173","arxiv_id":"2411.14820","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Local and global stabilization of SL(2) orbital integrals and the trace formula is extended to all characteristics, with a new Shalika-type asymptotic germ expansion in characteristic 2.","lead":"This paper extends the stabilization of orbital integrals and the trace formula for SL(2) to fields of any characteristic, with new phenomena in characteristic 2. A specialist would read it because it supplies a Shalika-type asymptotic expansion near the identity and a pre-stabilization step needed when Arthur's methods fail.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The global trace-formula half of the all-characteristic stabilization claim is deferred to the unpublished preprint [LLe]; until its function-field truncation and spectral results are verified, the characteristic-2 unipotent contribution and the main theorem are conditional.","rationale":"I read the paper as an extension of the Labesse-Langlands program, with the characteristic-2 phenomena explicitly isolated. The local geometric transfer (§2) and spectral transfer (§3) are largely explicit, and I find no clear internal contradiction in their stated proofs. The global stabilization, however, is not self-contained: Section 4's function-field trace formula is explicitly imported from an unpublished preprint, and Section 7.3 asserts a verbatim extension of [LL] rather than proving it. The reader's CONDITIONAL verdict is the right calibration: the concern is one of external dependency and unverified proof load, not a demonstrated false step. I would not reject the paper; I would keep it conditional until the [LLe] dependency is replaced by a publicly verified statement or the relevant function-field theorems are proved in the paper itself. The concrete check above would settle whether the characteristic-2 unipotent contribution, and hence the all-characteristic stabilization claim, is actually supported.","tokens_in":50637,"tokens_out":27499,"duration_ms":299462,"concrete_test":"Audit [LLe] (and, where invoked, [LW]) for the exact statements used here: (i) convergence of the truncated geometric kernel over function fields, (ii) the PolExp expansion of J^X_hyp and J^X_unip at X=0, (iii) the function-field analogue of Lemmas 4.5.1-4.5.2 for the spectral side, and (iv) compatibility with strict truncations in §4.1. Then re-run the §5.4 computation of J^X_unip(f,κ,z) for F=F_2(t) with a non-trivial κ and a locally constant f; if the formula in Lemma 5.4.1, including the 1/2 factor from a↦a^2 and the remark on 'réseaux isomorphes à Z ou 2Z', does not match the direct Fourier-inversion definition, the characteristic-2 pre-stabilization fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 7.3 says the proofs of the inner-form statements are those of [LL] 'dont les arguments s'étendent verbatim', and Section 4 repeatedly sends the function-field trace formula to [LLe] (see §4.2: 'on renvoie à [LW] et [LLe] pour des preuves détaillées valables pour les groupes généraux en toute caractéristique'; §4.5: 'Le cas des corps de fonctions ... (voir [LLe])'). The central claim is the stabilization of the trace formula in every characteristic; the global trace formula over function fields is therefore load-bearing, and it is not proved in this note. This matters specifically for characteristic 2: §5.4 derives the unipotent contribution from a pre-stabilization whose convergence and normalization use [LLe], and the note itself says Arthur's unipotent measure arguments fail in characteristic 2. If [LLe] does not contain, or its arguments do not cover, the characteristic-2 SL(2) case with the strict truncations used here, the announced all-characteristic stabilization is unsupported. The blanket sentence in §7.3 cannot repair this because [LL] is a characteristic-zero reference and the characteristic-2 germ/measure phenomena are explicitly new.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper announces, for SL(2) and its inner forms, a stabilization of local orbital integrals and of the trace formula over local and global fields of arbitrary characteristic, with new phenomena in characteristic 2. In characteristic 2 the paper claims that the usual Shalika germ expansion near the identity must be replaced by a κ-germ expansion, since the set of rational unipotent conjugacy classes is uncountable, and that the unipotent contribution to the global trace formula requires a pre-stabilization because Arthur's unipotent measure techniques fail. Sections 2 and 3 give a fairly detailed treatment of local geometric transfer, the fundamental lemma in the unramified case, local spectral transfer, and the Weil-representation construction of L-packets. Sections 4–6 present the non-invariant trace formula, its pre-stabilization, and the stabilization leading to Theorem 6.5.2. Section 7 treats quaternion-division-algebra inner forms.","tokens_in":50848,"tokens_out":5666,"duration_ms":63181,"significance":"If the announced results are correct, the paper provides a complete stabilization package for SL(2) in all characteristics and, for the first time, a Shalika-type germ expansion in characteristic 2, formulated via κ-germs and integrals over the uncountable set of unipotent classes. The proposed pre-stabilization of the unipotent contribution is a genuinely new device, since Arthur's measure-theoretic treatment is characteristic-zero specific. The local transfer and fundamental lemma in §§2.2–2.4 are worked out explicitly, and the spectral transfer in §3.5 is supported by concrete character computations. These local parts are valuable even independently of the global stabilization. However, the global half of the announcement depends in essential places on an unpublished preprint [LLe] and on a blanket assertion that the arguments of [LL], a characteristic-zero paper, extend verbatim to all characteristics; this dependency is load-bearing and not resolved by the text as it stands.","major_comments":[{"comment":"The global trace formula for function fields is not proved in this manuscript. In §4.2 the convergence of the hyperbolic and unipotent terms is dispensed with by the sentence 'on renvoie à [LW] et [LLe] pour des preuves détaillées valables pour les groupes généraux en toute caractéristique'; in §4.5 the spectral side for function fields is again delegated to [LLe], with 'La formulation pour les corps de fonctions est laissée au lecteur (cf. [LLe])'. Since [LLe] is listed as a 'Prépublication' and is not available here, the central claim of a stabilization of the trace formula in every characteristic is conditional, not proved. This matters concretely for characteristic 2: §5.4 derives the unipotent contribution from the pre-stabilization whose convergence and normalization use the Poisson formula and [LLe], and §4.1 notes that the strict truncation inequalities used here change the answer for function fields. Please either include a self-contained proof for SL(2) over function fields or state the global theorems as conditional on [LLe].","section":"§4.2, §4.5, Theorem 6.5.2"},{"comment":"The blanket sentence 'On renvoie pour la preuve des énoncés ci-dessus à [LL] dont les arguments s'étendent verbatim au cas des corps locaux et globaux de caractéristique quelconque' cannot carry the load assigned to it. [LL] is a characteristic-zero reference, while the paper itself states in the Introduction and in §5.4 that characteristic 2 has new unipotent phenomena for which Arthur's techniques fail and a new pre-stabilization is required. A verbatim extension of [LL] therefore cannot cover the characteristic-2 germ expansion and the characteristic-2 unipotent contribution unless the relevant lemmas are separately listed and checked. The manuscript should identify exactly which lemmas of [LL] are characteristic-independent and give the proofs or references for the characteristic-2 replacements, rather than asking the reader to accept a global 'verbatim' transfer.","section":"§7.3"},{"comment":"The characteristic-2 germ expansion is the paper's advertised new local result, but it is not stated as a theorem with hypotheses, measure normalizations, and proof. In characteristic p≠2, Fourier inversion over the finite group U is clear; in characteristic 2, U ≃ Q_F is uncountable compact, and the displayed formula O(t,f)=Γ_1(t)f(1)+∑_κ Γ_κ(ν,t)O_κ(ν,f) needs a precise interpretation of the sum over the (now possibly uncountable) dual, a proof that the κ-orbital integrals O_κ(ν,f) are well-defined, and a justification of the Fourier-inversion step. The heuristic replacement of sums by integrals announced at the start of §2.4 is not the same as a proof. Since this expansion is the basis for the claimed new Shalika-type germs and for the comparison to the classical expansion in p≠2, this gap is load-bearing.","section":"§2.4"},{"comment":"The endoscopic contribution to the T-stable trace formula is reduced to the identity J_geom(f,ε_{E/F}) = (1/2)∑_{γ∈T_{E/F}(F)} f^E(γ). This reduction uses in an essential way the measure comparisons between local Tamagawa measures, the global transfer constructed from §§2.3–2.4, and Lemma 5.4.2. The argument is plausible, but the convergence of the global sums over γ∈T_{E/F}(F) and z∈Z(F), as well as the passage from local transfer to global transfer with Tamagawa-normalized measures, is only sketched. A precise statement of the measure normalizations used for the global κ-orbital integrals in §5.2 and §5.4 is needed before Theorem 6.5.2 can be regarded as proved.","section":"§6.2–6.3"}],"minor_comments":[{"comment":"The displayed definition 'q_F = e = lim_{n→∞}(1+1/n)^n = 2,718...' is confusing: for number fields q_F=e is a convention, but the notation q_F clashes with the later cardinality q=p^f for function fields. Please clarify the convention in one sentence and use separate notation if possible.","section":"§1.1"},{"comment":"The symbol U is used both for the unipotent radical of the Borel subgroup (§1.2) and for the set of unipotent conjugacy classes (§2.4). This makes statements such as 'U ≃ Q_F' ambiguous on first reading; a different letter, for example 𝒰 or 𝒞, would help.","section":"§2.4"},{"comment":"The term PolExp is used without definition at its first occurrence. Since the paper explicitly discusses function fields and truncation parameters taking rational values, PolExp should be defined, or a reference to [LLe] should be given at the first use.","section":"§4.2"},{"comment":"The proof of Proposition 4.3.1 is only sketched by a displayed identity and the phrase 'Il suffit alors d'observer que...'. For a paper claiming proofs in all characteristics, the role of the strict truncation inequalities in the function-field case should be spelled out here, especially because §4.1 says the strict versus non-strict cutoffs matter for function fields.","section":"§4.3"},{"comment":"The entries for [LL] and [H] lack full bibliographic data: [LL] is given without page numbers, and [H] is given with the title 'Differential Geometry and Symmetric Spaces' although the standard reference is Helgason's 'Differential Geometry, Lie Groups, and Symmetric Spaces'. Please correct and complete the references.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The core difficulty is that the paper's advertised global and characteristic-2 results are not self-contained: the function-field trace formula is delegated to the unpublished [LLe], and the characteristic-2 replacement for Arthur's unipotent techniques is asserted rather than proved. The local endoscopic transfer, fundamental lemma, and local spectral transfer sections are strong and could stand as a valuable contribution if the global claims are either proved in the manuscript or made explicitly conditional. I would advise the editor to ask for the missing proofs or a precise statement of dependence on [LLe] before publication, and to verify that [LLe] indeed covers the SL(2) function-field case with the strict truncations used here."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a serious, genuinely new extension of Labesse–Langlands to all characteristics, and the characteristic-2 parts are real. The main theorem is nonetheless conditional until the function-field trace formula in [LLe] is available.\n\nThe local transfer, fundamental lemma, and spectral transfer for SL(2) are proved or carefully reduced; the Weil representation construction of L-packets is clean. The characteristic-2 germ expansion is new: the uncountably many unipotent conjugacy classes make Shalika sums meaningless, and replacing them by kappa-germs with Fourier inversion is a natural and substantial step. The pre-stabilization of the unipotent contribution is also genuinely new, since Arthur's measure argument fails in char 2.\n\nSoft spots. Section 7.3 says the inner-form proofs extend verbally from [LL]. [LL] is characteristic zero; the paper itself notes that char-2 unipotent phenomena are new, so a blanket 'verbatim' cannot carry that part. The same issue appears in Sections 4.2 and 4.5, where the global trace formula for function fields is deferred to [LLe], an unpublished preprint. The announced all-characteristic stabilization of the trace formula depends on the function-field truncation, spectral decomposition, and polynomial-exponential asymptotics from [LLe]; until that preprint is checked, the global theorem is conditional. Section 5.4 derives the char-2 unipotent contribution with a pre-stabilization whose convergence and normalization also uses [LLe]. The asymptotic near identity is obtained via stabilization; the paper presents the germ argument as a derivation rather than a fully formal theorem. None of this shows the math is wrong; it is a load-bearing citation gap. Minor: the constant choices differ from [LL] and [Lan]; the paper notes this, and it is a convention issue, but it means readers cannot quote results without checking normalizations.\n\nBottom line: for someone working in endoscopy or trace formulas in positive characteristic, this is worth reading seriously. The local results and the char-2 germ idea are the real content. It deserves peer review; a referee should demand that the [LLe] dependence be spelled out and, ideally, that [LLe] be posted or the needed statements proved.","headline":"Genuinely new characteristic-2 germ expansion and pre-stabilization for SL(2), but the global all-characteristic theorem leans on an unpublished preprint and a blanket verbal extension.","tokens_in":51418,"tokens_out":1680,"would_cite":true,"duration_ms":17691,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["11F70","11F72","22E35","22E50"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that the stabilization of local orbital integrals and of the trace formula for SL(2) holds over fields of every characteristic, and that in characteristic 2 the asymptotic germ expansion is new, with sums over unipotent…","keywords":["SL(2)","endoscopy","Shalika germs","kappa-germs","trace formula","characteristic 2","orbital integrals","stable conjugacy"],"falsifier":"The decisive test is to compute both sides of the identity $J_{\\mathrm{geom}}(f,\\varepsilon_{E/F})=J_{\\mathrm{spec}}(f,\\varepsilon_{E/F})$ from Section 6.2 for a characteristic-2 function field and a test function supported near the unipotent element $\\nu$, using the explicit transfer $f^E$ of Theorem 2.3.3; a mismatch in the coefficient of $f^E(1)$ would show the asserted $\\kappa$-germ expansion is incorrect.","tokens_in":50352,"feed_emoji":"🧮","tokens_out":13300,"duration_ms":114851,"temperature":0.7,"pith_summary":"This paper establishes the stabilization of local orbital integrals and of the global trace formula for SL(2) over local and global fields of arbitrary characteristic. Its new claim is that in characteristic 2 the asymptotic expansion near the identity cannot be written as the finite Shalika germ sum over unipotent conjugacy classes, because those classes form an uncountable compact set; replacing the sum by an integral over this set gives a $\\kappa$-germ expansion that is equivalent, up to Fourier transform, to the standard expansion in other characteristics. On the global side, the fine unipotent contribution to the trace formula is obtained by a pre-stabilization step, since the classical unipotent-variety measure technique is not available in characteristic 2. The paper's aim is to complete the SL(2) case of endoscopy uniformly in all characteristics and to provide the previously missing germ expansion in characteristic 2.","feed_headline":"SL(2) stabilization and germ expansion now work in all characteristics","feed_subtitle":"The trace formula and Shalika germs work in characteristic 2, with sums over unipotent classes replaced by integrals.","key_machinery":"The central mechanism is the endoscopic transfer from $G=\\mathrm{SL}(2)$ to its endoscopic data: the group itself and the norm-one tori $T_{E/F}$ attached to separable quadratic extensions $E/F$. The transfer factor $$\\Delta_E(t,t')=\\$\\lambda$(E/F,\\psi)^{-1}\\varepsilon_{E/F}\\bigl((\\gamma-\\bar{\\gamma})/(\\tau-\\bar{\\tau})\\bigr)|\\gamma-\\bar{\\gamma}|$$ converts $\\kappa$-orbital integrals $O_\\kappa(t,f)$ into smooth functions $f^E$ on the endoscopic torus, and the $\\kappa$-germs $\\Gamma_\\kappa(\\nu,t)$ obtained from the inverse transfer factor carry the asymptotic expansion near the identity. In characteristic $2$ the set of regular unipotent classes is isomorphic to the compact uncountable group $Q_F$, so Fourier inversion over its dual turns the Shalika sum into an integral; the same $\\kappa$-germ formalism remains meaningful in every characteristic. Globally, a pre-stabilization step extends every term of the trace formula to a function on $\\tilde G(\\mathbb{A}_F)$ and decomposes it under the compact quotient $Q_F$, replacing the unipotent-variety measure used in characteristic zero.","core_discovery":"The paper's central claim is that, for $G=\\mathrm{SL}(2)$, local orbital integrals admit an asymptotic expansion near the identity in every characteristic, obtained through endoscopic transfer. When the residual characteristic is not $2$, this expansion is equivalent, after Fourier transform on the finite group $Q_F\\simeq F^\\times/(F^\\times)^2$, to the classical Shalika germ expansion labelled by unipotent conjugacy classes. In characteristic $2$, the regular unipotent conjugacy classes form an uncountable compact set, so the expansion integrates $\\kappa$-germs over $Q_F$ rather than summing over classes; this is new. In parallel, the global trace formula is stabilized: a pre-stabilization step produces fine geometric terms, and endoscopy writes both sides as sums over the endoscopic data, namely $\\mathrm{SL}(2)$ itself and the norm-one tori of separable quadratic extensions. No step excludes characteristic $2$.","pith_inferences":["The same pre-stabilization idea may be the right route for other reductive groups in bad characteristic whenever the unipotent variety does not support the classical measure; the paper does not develop that generalization.","The $\\kappa$-germ formalism suggests that Shalika's expansion in equal-characteristic local fields should be formulated as an integral over a compact parameter space rather than a sum over conjugacy classes, with structure constants read off the transfer factor.","A direct extension would be to compute the Fourier transform of the $\\kappa$-germs over $Q_F$ for a ramified quadratic extension in characteristic 2 and compare with Shalika germs from the residue field; the paper does not tabulate this comparison."],"forward_implications":["Local orbital integrals for SL(2) have a $\\kappa$-germ expansion near the identity with no restriction on the characteristic.","In characteristic 2, the germ expansion takes the integral form of Shalika's expansion: sums over unipotent classes become integrals over the compact uncountable group $Q_F$.","The non-invariant trace formula for SL(2) can be written as a sum over endoscopic data in any characteristic, with the torus data contributing stable terms and SL(2) itself contributing the T-stable terms.","For inner forms given by norm-one units of quaternion algebras, the same stabilization holds, including the case where an L-packet component has local intertwining algebra $M(2,\\mathbb{C})$ instead of $\\mathbb{C}^4$.","The fundamental lemma, namely the transfer of the characteristic function of the hyperspecial maximal compact subgroup, is valid in all characteristics."],"supporting_citations":[{"why":"Supplies the transfer, stabilization, and germ framework that Section 7.3 extends verbatim to arbitrary characteristic.","marker":"[LL]"},{"why":"Provides the local integrability of characters for SL(n) and inner forms in all characteristics, used for spectral transfer.","marker":"[Le]"},{"why":"Provides the truncated trace formula over function fields of positive characteristic, used for the geometric side.","marker":"[LLe]"},{"why":"Provides GL(2) local integrability and the explicit construction of L-packets through the oscillator representation.","marker":"[JL]"},{"why":"Supplies the intertwining-operator asymptotics used to identify signs in the spectral transfer.","marker":"[Lan]"},{"why":"Names the unipotent-variety measure that the paper replaces by pre-stabilization in characteristic 2.","marker":"[A3]"}],"fun_headline_variants":["SL(2) germs and stabilization now hold in every characteristic","Characteristic 2 no longer blocks SL(2) germ expansion","SL(2) trace formula stabilized, with new germs in char 2","Germ integrals replace sums in char-2 SL(2) stabilization"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise, stated in Section 7.3, is that the arguments and lemmas of [LL] extend verbatim to local and global fields of arbitrary characteristic, together with the cited results [Le] and [LLe] for character integrability and the truncated trace formula over function fields; if any of these fail, the announced stabilization and the characteristic-2 germ expansion are not established.","fun_headline_variants_meta":{"raw":{"variants":["SL(2) germs and stabilization now hold in every characteristic","Characteristic 2 no longer blocks SL(2) germ expansion","SL(2) trace formula stabilized, with new germs in char 2","Germ integrals replace sums in char-2 SL(2) stabilization"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000639,"raw_usage":{"total_tokens":2889,"prompt_tokens":835,"completion_tokens":2054,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":451,"completion_tokens_details":{"reasoning_tokens":1987}},"tokens_in":451,"tokens_out":2054,"duration_ms":22593,"temperature":1.0,"reasoning_tokens":1987,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T14:51:03.495069+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"The decisive test is to compute both sides of the identity $J_{\\mathrm{geom}}(f,\\varepsilon_{E/F})=J_{\\mathrm{spec}}(f,\\varepsilon_{E/F})$ from Section 6.2 for a characteristic-2 function field and a test function supported near the unipotent element $\\nu$, using the explicit transfer $f^E$ of Theorem 2.3.3; a mismatch in the coefficient of $f^E(1)$ would show the asserted $\\kappa$-germ expansion is incorrect.","supporting_citations":[],"review_version":1}