{"id":"fb9d93a2-aba9-47ea-8a05-03e4d8dfb0c1","arxiv_id":"2506.19411","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"In any 1-h-minimal valued field of mixed characteristic (0,p), every transcendental definable curve has at most O_epsilon(H^epsilon) rational points of height at most H.","lead":"Over p-adic-like valued fields, a new theorem proves that transcendental definable curves contain very few rational points of bounded height, the count growing slower than any positive power of the height. This is the first such bound valid in arbitrary mixed-characteristic valued fields, extending the classical Pila-Wilkie theorem from ordered fields.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 4.1's ball-counting proof is invalid as written; the missing rational-packing bound underpins the residue-field-independent H^epsilon count.","rationale":"The reader's conditional verdict identifies the same soft spot: Lemma 4.1 has a statement/proof mismatch and the ball-counting step is not fully justified. I agree that this is the most load-bearing concern because it is the exact mechanism that makes the counting independent of the residue field and value group, which is the advertised novelty over existing p-adic results. The unproved T^r closure assertion is also central, but it is a standard and almost certainly true fact, whereas the packing bound in Lemma 4.1 is both demonstrably misproved and nontrivial in arbitrary mixed-characteristic fields with infinite residue field and non-discrete value group. I do not think the theorem is false; the concern is that the written proof has a genuine gap at a critical junction. The proposed test (checking the packing bound in C_2 with an irrational-valued c) would settle whether Lemma 4.2's O(p^N log H) bound survives, and if so, the proof can likely be repaired by replacing the faulty residue-class argument with a p-adic packing argument for rational points. The verdict should remain conditional rather than accept or reject: the theorem is plausible and the gap is localized, but the proof as written is not complete.","tokens_in":8094,"tokens_out":52679,"duration_ms":529668,"concrete_test":"Verify the rational-packing statement needed by Lemma 4.1: for every mixed-characteristic valued field K, every c in O_K, every delta in |K^×|, and every N >= 0, show that any rational x_1,...,x_m in Q ∩ O_K with |x_i - c| = delta and |x_i - x_j| >= |p^N| delta satisfies m <= C p^N for an absolute constant C independent of K, c, delta, N. A direct check: work in Q_2 with K = C_2, c = sqrt(2), so delta = 2^{-1/2}, and test candidates x = 0, 4, 8, 12, ... for N = 0..10; confirm m <= 2^N (or a fixed multiple). If m > p^N+1 occurs, Lemma 4.2 as stated is false; if m <= C p^N, insert the missing proof and check whether the extra factor C is absorbed into the constant m in Lemma 4.4 without changing the H^epsilon exponent.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof's ball-counting step, Lemma 4.2, is the only place where the H^epsilon rate is obtained independently of the residue field and value group. It depends on Lemma 4.1. The proof of Lemma 4.1 is not valid as written: it says the rational numbers x_i - x_1 are distinct modulo p^{N+1}delta and |x_i-x_1| <= delta, 'hence' m <= #(Z/p^{N+1}Z) = p^{N+1}. But the lemma claims m <= p^N+1, and p^{N+1} does not imply p^N+1. More seriously, for arbitrary mixed-characteristic K (e.g. C_p), the quotient O_K/(p^{N+1}delta O_K) is not Z/p^{N+1}Z: if the residue field is infinite it is infinite, and if delta is not a p-adic norm of a rational, the classes of rational x_i are not counted by Z/p^{N+1}Z. What is needed is a packing bound for rational points in Q_p at distance delta from an arbitrary c in K, with pairwise separation |p^N|delta. No such bound is proved, and the step is load-bearing: without an O(p^N log H) bound on nonempty balls, Lemma 4.4 yields no H^epsilon curve count. The flaw is likely repairable, since the true bound should be C p^N with C depending only on p and the fractional part of v(delta), but the paper does not supply the argument.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves Theorem 1.2: for any 1-h-minimal valued field K of mixed characteristic (0,p), every transcendental definable curve C⊂K^n has at most O_{C,\\varepsilon}(H^\\varepsilon) rational points of height at most H. The proof strategy is to project C to a planar curve (Lemma 3.2), decompose it into finitely many graphs of functions satisfying strong derivative and Taylor bounds (Lemma 3.3), use scaling maps to produce T^r maps (Lemma 3.5), apply a Bombieri–Pila type determinant method to cover the rational points by few algebraic curves (Lemmas 4.3–4.4), and finally control the number of relevant balls with Lemma 4.2. The claimed bound is uniform in the residue field and value group, and applies for example to C_p with an analytic structure.","tokens_in":8369,"tokens_out":21436,"duration_ms":222569,"significance":"If the missing technical steps are supplied, this is a significant result: it extends the Pila–Wilkie counting philosophy and the Cluckers–Comte–Loeser p-adic results to arbitrary 1-h-minimal fields of mixed characteristic, and it is the first bound of this kind in such general valued fields. The proof strategy is attractive because it avoids r-th power maps and instead uses scaling maps, following the spirit of Binyamini–Novikov–Zak. The paper is not fully self-contained, relying on the 1-h-minimal machinery from [6,7] and on a projection lemma from [3], but this is made explicit and is acceptable for a research note. The explicit dependency of the constants on p and the curve is a useful additional feature.","major_comments":[{"comment":"The proof of Lemma 4.1 is not valid as written. The assumption gives that the xi-c are distinct modulo p^{N+1}δ, and from this the proof concludes that the rational numbers xi-x1 are distinct modulo p^{N+1}δ and hence that m≤#(Z/p^{N+1}Z)=p^{N+1}. This does not imply the claimed bound m≤p^N+1, and more importantly the counting of distinct classes by Z/p^{N+1}Z is unjustified for arbitrary K of mixed characteristic: if the residue field is infinite, the quotient O_K/(p^{N+1}δ O_K) is infinite, and the classes of rational numbers inside a ball around c need not be parametrized by Z/p^{N+1}Z. What is needed is a packing argument showing that the rational points in B(c,δ) lie in a single ball over Q_p and that their separation condition leaves at most O(p^N) of them; no such argument is supplied. Since Lemma 4.2 and hence the H^ε ball count in Lemma 4.4 depend on this lemma, the main proof has a load-bearing gap at this point. The gap appears repairable, but it must be fixed explicitly.","section":"Section 4, Lemma 4.1"},{"comment":"The assertion that 'Hence also s_a^i g^j is T^r for any i,j∈N' is made without proof and is not automatic. The class T^r as defined in Definition 2.3 is not closed under products with the stated constant 1: for a product of two T^r maps, the l-th derivative can involve l+1 terms, and the bound |(fg)^{(l)}|≤|l!| is not generally preserved. This matters because Lemma 4.3 uses the T^r bound |β_{i,j,k}|≤1 to obtain the determinant estimate, and Lemma 4.4 then applies the estimate to the functions s_a^i g^j. The author should either prove that these specific products satisfy the T^r bounds, or prove a variant of Lemma 4.3 with an extra constant factor and check that such a constant is harmless in the final count.","section":"Section 4, paragraph before Lemma 4.4"},{"comment":"The ball-counting constants are stated inconsistently: Lemma 4.1 gives m≤p^N+1, Lemma 4.2 states a bound with a factor that is printed as 2Mp^N+1, and the proof of Lemma 4.4 later refers to 2Mp^{N+2}(1+2log_p H). These notational discrepancies should be harmonized. More substantively, Lemma 4.2 also asserts without proof that the number of possible values |c-x| for rationals x of height at most H is at most 2+4log_p(H); this estimate should be justified explicitly, since it is part of the logarithmic factor used in the final count.","section":"Section 4, Lemma 4.2 and Lemma 4.4"}],"minor_comments":[{"comment":"The proof that f∘s_a inherits Taylor approximation of order r-1 is stated in one sentence; a short verification using the chain rule and the definition of an r-parametrizing map would make the argument easier to check.","section":"Section 3, Lemma 3.5"},{"comment":"The claim that π restricted to C is finite-to-one 'otherwise C would contain a subset {a}×B' uses the 1-h-minimal fact that every infinite definable subset of O_K contains a ball; this fact should be cited or stated.","section":"Proof of Theorem 1.2"},{"comment":"The reduction to algebraic Skolem functions is justified by reference to [7, Prop. 3.2.3], but the sentence 'This is no loss in generality' would benefit from an explanation of why the rational-point count is preserved under the relevant extension.","section":"Section 2"},{"comment":"In the expansion of the determinant, the case ℓ_j=0 should be addressed explicitly: then column j equals the first column, so the corresponding determinant vanishes; this is needed for the conclusion that non-zero terms must have sum of ℓ_j at least e.","section":"Section 4, Lemma 4.3"}],"recommendation":"major_revision","confidential_remarks":"The main theorem is plausible and the strategy is promising, but the proof as submitted has a serious gap in the ball-counting lemma (Lemma 4.1) and an unproved T^r closure assertion before Lemma 4.4. Both appear fixable within the scope of the paper, so I recommend major revision rather than rejection. The paper also leans on several prior works of the author, including [3] and [7], and the editor may wish to confirm that the cited projection lemma (Lemma 3.2 from [3]) is available in exactly the 1-h-minimal mixed-characteristic setting used here."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read Vermeulen's note on counting rational points on transcendental curves in valued fields. The main theorem is exactly what the title says: for K of mixed characteristic (0,p) with 1-h-minimal structure, any transcendental definable curve has at most O(H^ε) rational points of height ≤ H. That's a real step beyond the p-adic results of Cluckers-Comte-Loeser and the equicharacteristic zero work, and the innovation is the scaling-map parametrization, replacing r-th power maps. This is the right idea and the paper is short and readable.\n\nWhere it holds up: the determinant method part (Lemma 4.3) is standard and fine; the reduction to planar curves and the parametrization lemmas are properly imported from prior work; the paper is transparent that it only handles curves and says plainly why higher dimensions get stuck.\n\nThe soft spot is Lemma 4.1 and its use in Lemma 4.2. The proof of Lemma 4.1 claims that distinctness modulo p^{N+1}δ forces m ≤ p^{N+1}, and then concludes m ≤ p^N+1. Those don't match. More importantly, for a field like C_p the residue ring O_K/(p^{N+1}δ O_K) is infinite, so the \"hence m ≤ #Z/p^{N+1}Z\" step needs an argument that the rational points occupy only the Z_p-image; the paper doesn't give it. Lemma 4.2 is the only place where the H^ε rate gets independent of the residue field and value group, so this is load-bearing, not cosmetic. I think the lemma is true and repairable—a packing argument for rationals in Q_p at distance δ from c with separation |p^N|δ should give C p^N with C depending on p and the fractional part of v(δ)—but it's not in the paper.\n\nAlso, Section 4 states without proof that s_a^i g^j is T^r whenever g is; that's presumably a routine check with the chain rule, but it deserves a line.\n\nNet: the theorem is likely right and important, and the paper deserves a serious referee. I'd send it out, but the referee should insist on a corrected Lemma 4.1/4.2 before acceptance. The rest is solid.","headline":"A genuinely new scaling-map trick gives a Pila-Wilkie bound for curves in arbitrary 1-h-minimal mixed-characteristic fields, but the proof has a load-bearing gap in the ball-counting step that needs fixing.","tokens_in":8937,"tokens_out":10579,"would_cite":true,"duration_ms":101854,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A transcendental definable curve in a 1-h-minimal valued field of mixed characteristic has at most $c_\\varepsilon H^\\varepsilon$ rational points of height at most H.","keywords":["rational points","transcendental curves","valued fields","1-h-minimality","height bounds","determinant method","parametrizations","mixed characteristic"],"falsifier":"Find a 1-h-minimal valued field of mixed characteristic and a definable transcendental curve whose rational-point count grows faster than every $H^\\varepsilon$, for example roughly $H^c$ with $c>0$; such a curve would refute Theorem 1.2. At the lemma level, a single pair of $T^r$ maps whose product or scaling fails to be $T^r$ would break the determinant step.","tokens_in":7854,"feed_emoji":"📐","tokens_out":11686,"duration_ms":104983,"temperature":0.7,"pith_summary":"This paper proves a counting theorem for rational points on transcendental curves in arbitrary valued fields of mixed characteristic that carry a 1-h-minimal structure: for every $\\varepsilon>0$, the number of rational points of height at most $H$ on such a curve is bounded by $O_\\varepsilon(H^\\varepsilon)$. The theorem extends to this general setting a bound previously known for $p$-adic fields with analytic structure, and it brings the non-archimedean world into line with the classical o-minimal counting result. The proof covers the rational points on the curve by few algebraic curves of large degree, combining the determinant method with parametrizations by scaling maps instead of $r$-th power maps.","feed_headline":"Transcendental curves in valued fields have at most H^ε rational points","feed_subtitle":"The bound extends a p-adic counting result to all 1-h-minimal mixed-characteristic fields, using scaling maps and determinants.","key_machinery":"The central object is the $r$-parametrizing map: a definable function $f:U \\subset \\mathcal{O}_K \\to \\mathcal{O}_K$ whose derivatives satisfy $|f^{(i)}(x)| \\le |M|^{-i}|x-c|^{-i}$ on balls prepared by a centre $c$ and integer $M$. Lemma 3.3 asserts that every definable curve is a finite union of graphs of such maps, and Lemma 3.5 shows that composing an $r$-parametrizing map with the scaling map $s_a(x)=(a-c)(1+p^M x)+c$ turns it into a $T^r$ map, meaning uniform Taylor approximation and derivative bounds. This scaling trick replaces the $r$-th power substitutions used in earlier $p$-adic work and feeds the determinant estimate of Lemma 4.3, which bounds a determinant of values of $T^r$ maps by the ball radius to the power $e=\\binom{r}{2}$. That estimate is what forces the rational points to lie on few algebraic curves.","core_discovery":"Let $K$ be a valued field of mixed characteristic $(0,p)$ equipped with a 1-h-minimal structure, and let $C \\subset K^n$ be a definable curve whose intersection with every algebraic curve is finite; call such a curve transcendental. Theorem 1.2 states that $\\#C(\\mathbb{Q},H) \\le c_\\varepsilon H^\\varepsilon$ for every $\\varepsilon>0$, where $C(\\mathbb{Q},H)$ is the set of rational points on $C$ with numerator and denominator bounded by $H$. The paper proves this by first projecting $C$ to a planar curve that is non-algebraic up to a high degree, then decomposing it into graphs of $r$-parametrizing functions, and then applying the determinant method to confine all rational points of bounded height to few algebraic curves. Since each algebraic curve meets a transcendental curve in only finitely many points, the sub-polynomial bound follows.","pith_inferences":["A concrete next test is the dimension-two version of Lemma 4.2: the paper identifies bounding the number of twisted boxes that contain rational points as the main obstacle, and a counterexample or a proof for restricted derivative growth would show whether the method extends to definable surfaces.","Since the proof avoids $r$-th power maps and is insensitive to the residue field and value group, it suggests that the sub-polynomial bound is a feature of the definable structure itself; testing a valued field where power maps are not definable would delimit exactly how general the mechanism is.","The reliance on algebraic Skolem functions could be probed directly: if a 1-h-minimal structure without definable choice still satisfies Lemma 3.3, the theorem would transfer unchanged, and finding such a structure would separate the choice assumption from the counting phenomenon."],"forward_implications":["If Theorem 1.2 is correct, every transcendental definable curve in a mixed-characteristic 1-h-minimal valued field has sub-polynomial rational point counts, matching the classical o-minimal bound in a new setting.","The form of the constant $c = M p^\\alpha$, with $M$ and $\\alpha$ independent of the field, gives a uniform-in-$p$ version and an extension to large positive characteristic once the characteristic is large enough for the curve.","A compactness argument turns the single-curve bound into a bound for definable families of transcendental curves with a constant independent of the parameter.","At the level of a single parametrizing map, the rational points of height at most $H$ lie on at most $O(H^\\varepsilon)$ algebraic curves of degree $d$, and each such algebraic curve contributes only finitely many points."],"supporting_citations":[{"why":"Supplies the classical o-minimal counting theorem being generalized.","marker":"[11]"},{"why":"Establishes the p-adic analytic counting result that Theorem 1.2 extends.","marker":"[4]"},{"why":"Provides the Jacobian property and Taylor approximation theorems underlying the parametrizing decomposition.","marker":"[7]"},{"why":"Gives the projection lemma reducing a transcendental curve to a planar curve non-algebraic up to degree d.","marker":"[3]"},{"why":"Introduces the determinant method used to cover rational points by few algebraic curves.","marker":"[2]"},{"why":"Provides uniform parametrizations and the T^r-map framework used in the determinant estimate.","marker":"[5]"},{"why":"Motivates the T^r notion and the higher-dimensional parametrization program.","marker":"[8]"}],"fun_headline_variants":["Sub-polynomial rational points on transcendental curves in valued fields","Valued fields: rational points on transcendental curves are sub-polynomial","New Pila-Wilkie type bound for transcendental curves in valued fields","Transcendental curves in valued fields: at most H^ε rational points","H^ε bound on rational points for transcendental curves in valued fields"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof depends on the full tame-geometry package on the valued field: definable curves decompose into finitely many graphs with controlled derivatives and Taylor approximations, and definable choice is available; if any part of that package fails, the counting argument stops.","fun_headline_variants_meta":{"raw":{"variants":["Sub-polynomial rational points on transcendental curves in valued fields","Valued fields: rational points on transcendental curves are sub-polynomial","New Pila-Wilkie type bound for transcendental curves in valued fields","Transcendental curves in valued fields: at most H^ε rational points","H^ε bound on rational points for transcendental curves in valued fields"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001068,"raw_usage":{"total_tokens":4406,"prompt_tokens":810,"completion_tokens":3596,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":426,"completion_tokens_details":{"reasoning_tokens":3503}},"tokens_in":426,"tokens_out":3596,"duration_ms":23525,"temperature":1.0,"reasoning_tokens":3503,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T18:33:28.352629+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a 1-h-minimal valued field of mixed characteristic and a definable transcendental curve whose rational-point count grows faster than every $H^\\varepsilon$, for example roughly $H^c$ with $c>0$; such a curve would refute Theorem 1.2. At the lemma level, a single pair of $T^r$ maps whose product or scaling fails to be $T^r$ would break the determinant step.","supporting_citations":[{"cited_title":"Pila and A","cited_arxiv_id":null,"evidence_quote":"Supplies the classical o-minimal counting theorem being generalized."},{"cited_title":"Cluckers, G","cited_arxiv_id":null,"evidence_quote":"Establishes the p-adic analytic counting result that Theorem 1.2 extends."},{"cited_title":"Cluckers, I","cited_arxiv_id":null,"evidence_quote":"Provides the Jacobian property and Taylor approximation theorems underlying the parametrizing decomposition."},{"cited_title":"Cantoral-Farf´ an, K","cited_arxiv_id":null,"evidence_quote":"Gives the projection lemma reducing a transcendental curve to a planar curve non-algebraic up to degree d."},{"cited_title":"Bombieri and J","cited_arxiv_id":null,"evidence_quote":"Introduces the determinant method used to cover rational points by few algebraic curves."},{"cited_title":"Cluckers, A","cited_arxiv_id":null,"evidence_quote":"Provides uniform parametrizations and the T^r-map framework used in the determinant estimate."},{"cited_title":"Cluckers, I","cited_arxiv_id":null,"evidence_quote":"Motivates the T^r notion and the higher-dimensional parametrization program."}],"review_version":2}