{"id":"431f785f-a26c-43af-824a-f57dcd1ad45f","arxiv_id":"2411.14960","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"For q-bounded algebraic extensions of F_p(t), rings of S-integral functions are first-order definable, and if the constant field is infinite, the first-order theory of the field is undecidable.","lead":"This paper shows that for a class of infinite extensions of the function field F_p(t), called q-bounded extensions, rings of integral functions can be described by first-order formulas inside the field, and the field's full theory is undecidable when its constant field is infinite. The work extends classic definability and decidability results from number fields to function fields and strengthens overlapping results by another group.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Remark 7.4 defines V_{K,S} incorrectly for non-Galois extensions; Corollary 9.3 relies on this via Corollary 8.4, so the proof needs repair (a Jacobson-radical definition fixes it).","rationale":"The reader's weakest assumption is the application of Demeyer's theorem to tuple sets in Theorem 10.1. That concern is valid and matters for Corollary 10.2 ('every F_p[w] is definable'), but it does not affect Corollary 9.3, which only needs one definable polynomial ring O_{K,u} and does not invoke Theorem 10.1. The central undecidability claim for non-Galois q-bounded fields instead depends on Corollary 8.4's second case and hence on the definability of V_{K,S}. The paper's Remark 7.4(2) gives an incorrect characterization of V_{K,p_K} for rings with several maximal ideals, which is exactly what can occur when a prime splits in a non-Galois extension. Since the paper present the main theorem without the Galois hypothesis, this is a load-bearing gap in the proof as written. However, the gap is readily fixable by defining V via the first-order definable Jacobson radical of the definable ring O_{p_K,K}, so the underlying mathematics appears sound and the appropriate verdict remains conditional rather than rejection. The concrete test isolates the failure of the formula in a split-prime example and verifies the repair.","tokens_in":38714,"tokens_out":31161,"duration_ms":284717,"concrete_test":"Let K=F_p(t) and let K be an algebraic extension in which a fixed prime p of K splits completely into two distinct primes p1,p2, while K is still q-bounded at p (for example, adjoin roots of X^2 - c for suitable constants to force splitting). Choose u in K with v_{p1}(u-1)>0 and v_{p2}(u-1)=0. Then u is a unit with u-1 non-unit in O_{p_K,K}, so Remark 7.4(2) puts u in its purported V_{K,p}, yet u is not in the actual V_{K,p} because v_{p2}(u-1)=0. This verifies the defect. Separately, check that replacing the definition with the Jacobson-radical formula, x in J(R) iff for all y in R, 1-xy is a unit, restores the proof of Corollary 8.4's second case and hence of Corollary 9.3.","verdict_should_be":"UNCHANGED","load_bearing_attack":"For the non-Galois case of Corollary 8.4, the definability of V_{K,S} is asserted in Corollary 7.5(a) using Remark 7.4(2), which identifies V_{K,p_K} with {u in U_{p_K,K} : (u-1) not in U_{p_K,K}}. In the integral closure O_{p_K,K} of a valuation ring, this condition means that u-1 has positive valuation at some prime over p_K, not at all primes over p_K as required by Notation 1.7. The two sets differ exactly when p_K has more than one extension to K and u-1 is positive at only one of them. Uniform q-boundedness (Definition 6.11) does not prevent a prime from splitting completely, so the stated formula can fail in the non-Galois situation used by Corollary 8.4's second case. Corollary 8.5 then uses this case to produce the definable O_{K,u} on which Corollary 9.3 depends. A correct definition exists: for R=O_{p_K,K}, the Jacobson radical J(R) is first-order definable by x in J(R) iff for all y in R, 1-xy is a unit in R, and V_{K,p_K} consists of units u with u-1 in J(R). Thus the main theorem is repairable, but the proof as written has a genuine gap.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a notion of q-boundedness for algebraic extensions of F_p(t), analogous to earlier work of the first author over Q, and uses norm equations together with the Hasse Norm Principle to obtain first-order definitions of integral closures of valuation rings and of S-integer rings in such extensions. The main consequences are: for a q-bounded extension K/F_p(t), the integral closures of infinitely many polynomial rings F_p[u] are definable in K; if the constant field is infinite, the first-order theory of K is undecidable and every F_p[w] is definable with parameters; and if some prime has finitely many factors in K, the first-order theory is undecidable. The paper also contains a construction of such extensions with finite constant field and a transfer theorem stating that definability of one F_p-polynomial ring implies definability of all of them.","tokens_in":38986,"tokens_out":10668,"duration_ms":101691,"significance":"If the gaps described below are repaired, this is a substantial advance in the model theory of infinite algebraic extensions of global function fields. The definability results are broader than those in the concurrent work by Martinez-Ranero, Salcedo, and Utreras, and the undecidability results are considerably stronger. The main technical engine, Proposition 4.7, is proven in full and correctly reduces norm equations in infinite extensions to finite subextensions; this is a genuinely useful contribution. The paper is largely self-contained in its use of standard tools (Weak Approximation, Hasse Norm Principle, class field theory), and it contains no fitted parameters or circular dependencies. The weak points are localized: a genuinely incorrect definition of V_{K,p_K} in the non-Galois case, and an under-justified application of Demeyer's theorem to tuples plus a missing coordinate in the displayed formula of Theorem 10.1. Both are repairable within the manuscript's scope.","major_comments":[{"comment":"The stated identification V_{K,p_K} = {u in U_{p_K,K} : (u-1) notin U_{p_K,K}} is not correct when p_K has more than one extension to K. For u in U_{p_K,K}, the condition (u-1) notin U_{p_K,K} means that there exists at least one prime v over p_K with v(u-1)>0, whereas Notation 1.7 defines V_{K,p_K} by v(u-1)>0 for all v over p_K. These two sets differ exactly when p_K splits and u-1 has positive valuation at one prime but valuation zero at another. Since Corollary 8.4's second case (uniform q-boundedness, without a Galois assumption) relies on Corollary 7.5(a), and Corollaries 8.5 and 9.3 use that case, the proof as written has a genuine gap. A correct first-order definition is available: for R=O_{p_K,K}, define its Jacobson radical by x in J(R) iff for all y in R, 1-xy is a unit in R, and then V_{K,p_K} = {u in U_{p_K,K} : u-1 in J(R)}. The manuscript should be revised accordingly; the Galois case is fine because then all primes over p_K are conjugate.","section":"Section 7, Remark 7.4(2) and Corollary 7.5(a)"},{"comment":"The proof applies Demeyer's theorem to the set B in (10.16), which is a subset of F_p[u]^{n+1}, but the cited theorem [Dem07] is stated for c.e. subsets of F_p[u], not for subsets of F_p[u]^{n+1}. The manuscript should either prove the needed tuple version via a Diophantine coding of tuples or explicitly state and prove a generalized version of Demeyer's theorem. Additionally, the displayed formula in the proof appears to drop the coordinate a_0: it uses p(a_1,...,a_{n-1},b,Y_1,...,Y_m)=0, while B is defined using (a_0,...,a_{n-1},b). As written, the formula does not define F_p[w] because the condition on a_0 is not enforced. This is load-bearing for Corollary 10.2 and should be corrected.","section":"Section 10, Theorem 10.1 and Eq. (10.16)"}],"minor_comments":[{"comment":"The notation F_p[u]^{n+1} in (10.16) should be written as F_p[u]^{n+1}; the superscript is not typeset correctly.","section":"Section 10, Eq. (10.16)"},{"comment":"The sentence 'Therefore, therefore the integral closure of F_p[u] is definable over K' contains a duplicated 'therefore' and should be rewritten.","section":"Section 8, Corollary 8.5"},{"comment":"Formula (8.14) is called a sentence but contains the free variable x; it is a formula defining a subset of K, not a sentence. The terminology should be adjusted.","section":"Section 8, Theorem 8.7"},{"comment":"The proof uses the same symbol L for the infinite field K(β_1,...,β_s) and for the finite extension ĤE(β_1,...,β_s); using two different symbols would improve readability and avoid confusion.","section":"Section 4, Proposition 4.7"}],"recommendation":"major_revision","confidential_remarks":"The two main issues are localized and appear fully repairable within the scope of the paper: the V_{K,p_K} definition can be fixed with the Jacobson radical, and the Demeyer application needs a tuple version and a corrected display. I would not recommend rejection, but the revised version should be checked carefully on these two points, since both are load-bearing for the announced corollaries. The overlap with MSU24 is handled fairly and the comparison is accurate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth knowing two things. First, this is the real article: q-boundedness for function fields is a genuinely useful analogue of Shlapentokh's number-field work, the proof machinery is detailed, and the definability results for valuation rings and rings of S-integers go beyond the overlapping MSU24 preprint. Second, there is a genuine gap in the non-Galois branch of the main undecidability theorem. Remark 7.4(2) defines V_{K,p_K} as {u in U_{p_K,K} : u-1 not in U_{p_K,K}}. In the integral closure of a valuation ring, 'u-1 not a unit' means v(u-1)>0 at some prime over p_K, whereas Notation 1.7 requires v(u-1)>0 at all primes over p_K. These sets differ when p_K has more than one extension in a non-Galois extension. The Galois case of Corollary 8.4 is fine, since Galois automorphisms move positivity at one prime to all primes. But Corollary 8.5, and hence Corollary 9.3, rely on the non-Galois/uniform branch. So the paper as written does not prove its headline undecidability theorem for arbitrary q-bounded K. The fix is standard: define V_{K,p_K} using the Jacobson radical of O_{p_K,K}, i.e. u-1 lies in J(R), equivalently for all y in R, 1-uy is a unit in R. This is first-order and works without Galois. I expect the theorem to survive, but the proof needs revision. What the paper does well: the transfer Proposition 4.7 is proven in full and does heavy lifting cleanly; the norm-equation lemmas are careful; the comparison with MSU24 is honest; and the undecidability claims are indeed much stronger. The idea that q-boundedness, rather than the more restrictive absolutely-q'-bounded prime condition, suffices for S-integer definability is the right generalization. Section 10's reduction from one polynomial ring to all polynomial rings is a nice argument. I agree with the reader's caveat there: the Demeyer theorem is applied to a tuple set, and the displayed formula appears to drop the coordinate a0. That is almost certainly a typo, and Demeyer's theorem should cover tuples, but the statement should be made explicit. Bottom line: this paper deserves a serious referee. The central definability machinery looks sound, the Galois version is solid, and the non-Galois gap is repairable by standard means. If I were the editor I would send it out, asking the referee to check the uniform-q-boundedness branch and the Demeyer citation carefully.","headline":"Genuinely strong paper with a real but repairable gap in the non-Galois case of the main undecidability theorem: Remark 7.4(2) mis-defines V_{K,p_K} when a prime splits, so Corollary 9.3 needs a fix before it is fully proved.","tokens_in":857,"tokens_out":1044,"would_cite":true,"duration_ms":53500,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["03B25","11U05","11R58","12L05"],"pacs":[],"model":"deepseek-v4-flash","headline":"A local degree bound makes infinite algebraic extensions of F_p(t) first-order undecidable when their constant field is infinite.","keywords":["first-order definability","undecidability","global function fields","q-bounded extensions","rings of S-integers","norm equations","Hasse norm principle","algebraic extensions of F_p(t)"],"falsifier":"For Corollary 9.3, the decisive observation would be a q-bounded algebraic extension of F_p(t) with infinite constant field whose first-order theory is decidable—no such field is known if the theorem is correct. For Theorem 10.1, one could examine the polynomial p(a_1,...,a_{n-1},b,Y_1,...,Y_m) in the displayed formula and test whether it defines the set B of tuples whose associated rational functions lie in F_p[w]; if the missing coordinate a_0 changes the membership condition, the proof of 'every F_p[w] is definable' would fail.","tokens_in":38490,"feed_emoji":"🧮","tokens_out":6491,"duration_ms":57299,"temperature":0.7,"pith_summary":"This paper tries to establish that a large class of infinite algebraic extensions of the rational function field F_p(t)—those satisfying a local property called q-boundedness—have undecidable first-order theories in the language of rings. The route is definability: under Galois or uniform q-boundedness assumptions, the ring of S-integral functions is first-order definable inside the field, and when the constant subfield is infinite this definability transfers undecidability from the ring to the whole field. The paper also proves that if one polynomial ring F_p[u] is definable in such a field, then every polynomial ring F_p[w] is definable, and it produces field extensions where a prescribed prime has only finitely many factors, yielding undecidability without the infinite-constant-field assumption. A sympathetic reader would care because this pushes the classical definability/decidability boundary for global fields into infinite algebraic extensions, giving new undecidable fields and identifying what makes them so.","feed_headline":"q-boundedness forces undecidability in infinite function fields","feed_subtitle":"A local bound on prime degrees defines rings of integral functions inside the field, transferring undecidability upward.","key_machinery":"The paper's central object is q-boundedness: for each prime path in the factor tree from F_p(t) to K, the order at the prime q of the ramification degree and relative degree along finite subextensions is uniformly bounded; locally, this means each tower of completions has q-bounded finite subextension degrees. The argument runs on norm equations over cyclic q-th root extensions, of the form $N_{L(\\sqrt[q]{a})/L}(y)=bx^q+b^q$, where L is a carefully chosen tower of q-th root extensions. q-boundedness supplies auxiliary elements a and b whose orders at primes remain controlled, so that a pole of x makes the norm equation unsolvable locally (by an inert prime), while integrality of x makes it solvable everywhere via the Hasse Norm Principle. The dynamic choice of L keeps the extension $L(\\sqrt[q]{a})/L$ unramified, so that local obstructions can be read off from orders of elements.","core_discovery":"The central claim, stated as Corollary 8.4, is that if K is a q-bounded algebraic extension of a finite extension of F_p(t), and either the extension is Galois or the primes in S are uniformly q-bounded, then the ring O_{K,S} of S-integral functions is first-order definable in K. From this the paper derives that when the algebraic closure of F_p in K is infinite, both O_K and K have undecidable first-order theories (Corollaries 9.2 and 9.3), and every polynomial ring F_p[w] with w non-constant is definable in K with parameters (Corollary 10.2). A separate theorem (9.4) removes the infinite-constant-field assumption when some prime of the base field has only finitely many factors in K. The intended force of the paper is that q-boundedness is the function-field counterpart of the number-field condition used earlier by the first author, and that it marks a reliable boundary between decidable and undecidable infinite algebraic extensions.","pith_inferences":["The main undecidability theorem (Corollary 9.3) does not pass through the all-polynomial-rings result; it uses only definability of one integral closure O_{K,u} and of F_p[t], so it would survive even if Theorem 10.1 needed repair.","The proof of Theorem 10.1 relies on Demeyer's theorem applied to a set of tuples rather than to a subset of F_p[u]; a skeptical reader should check whether the tuple version is actually supplied, since the displayed defining formula appears to omit the coordinate a_0.","q-boundedness is preserved under passing to finite extensions along the tower, so the same norm-equation template should define integral closures of valuation rings in any field whose local degree growth is q-bounded, suggesting the method may extend to fields where different primes use different primes q.","The construction at the end of the paper shows that the finitely-many-factors condition is easy to engineer by alternating inert and totally ramified primes, so the undecidability in Theorem 9.4 likely applies to many explicitly built towers, not just the examples listed."],"forward_implications":["For every q-bounded Galois extension K of F_p(t), the integral closure of F_p[t] in K is first-order definable in K, as a special case of Corollary 8.4.","If a q-bounded extension has an infinite constant field, then the first-order theory of the field K is undecidable (Corollary 9.3).","Under the same hypotheses, every polynomial ring F_p[w], for non-constant w in K, is first-order definable in K with parameters (Corollary 10.2).","If some prime of the base field has only finitely many factors in K, the first-order theory of K is undecidable without assuming the constant field is infinite (Theorem 9.4).","The integral closure of any valuation ring of a prime with uniformly q-bounded ramification and relative degree has a first-order definition in K; with absolutely bounded ramification the definition is existential (Theorems 7.2–7.3)."],"supporting_citations":[{"why":"Supplies the Hasse Norm Principle, which the paper uses to pass from local solvability of norm equations to global solvability.","marker":"[Tat67]"},{"why":"Provides the classical existential definability of valuation rings in global function fields that this paper adapts to infinite extensions.","marker":"[Rum80]"},{"why":"Introduces the q-boundedness idea in number fields and the norm-equation strategy that the paper transfers to function fields.","marker":"[Shl18]"},{"why":"Gives the existential undecidability of rings of S-integers over function fields used in Theorem 9.4 and Section 9.2.","marker":"[Shl92]"},{"why":"Asserts that every computably enumerable subset of F_p[u] is Diophantine, the external input on which Theorem 10.1 depends.","marker":"[Dem07]"},{"why":"Shows the first-order theory of F_p[t] is undecidable, which converts definability of F_p[t] into undecidability of larger rings.","marker":"[Den79]"}],"fun_headline_variants":["q-boundedness yields definable rings and undecidability","In infinite function fields, q-boundedness marks undecidability","Definable integral rings from a local bound on prime degrees","Hasse norm principle plus q-boundedness: undecidable fields"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof of Theorem 10.1 assumes that Demeyer's theorem—which the paper cites as saying every computably enumerable subset of F_p[u] is Diophantine—also covers the specific computable set B of (n+1)-tuples defined in (10.16); the paper does not prove this tuple version, and the displayed formula appears to drop the coordinate a_0, so that application is the load-bearing premise for the 'every polynomial ring is definable' conclusion.","fun_headline_variants_meta":{"raw":{"variants":["q-boundedness yields definable rings and undecidability","In infinite function fields, q-boundedness marks undecidability","Definable integral rings from a local bound on prime degrees","Hasse norm principle plus q-boundedness: undecidable fields"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000312,"raw_usage":{"total_tokens":1828,"prompt_tokens":1048,"completion_tokens":780,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":664,"completion_tokens_details":{"reasoning_tokens":716}},"tokens_in":664,"tokens_out":780,"duration_ms":8191,"temperature":1.0,"reasoning_tokens":716,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T14:42:02.971830+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For Corollary 9.3, the decisive observation would be a q-bounded algebraic extension of F_p(t) with infinite constant field whose first-order theory is decidable—no such field is known if the theorem is correct. For Theorem 10.1, one could examine the polynomial p(a_1,...,a_{n-1},b,Y_1,...,Y_m) in the displayed formula and test whether it defines the set B of tuples whose associated rational functions lie in F_p[w]; if the missing coordinate a_0 changes the membership condition, the proof of 'every F_p[w] is definable' would fail.","supporting_citations":[],"review_version":1}