{"id":"6e15a4b6-1eb1-482b-94aa-3ab1c794773d","arxiv_id":"2607.16331","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For every Takeuchi class I hypergeometric datum with 0<a,b,c<1, E(a,b,c) equals the set of Hauptmodul values at points of Q(√−d)∩H (explicit d and t), and E={0} when c≥1.","lead":"This paper claims to pin down exactly which algebraic inputs z make certain hypergeometric functions 2F1(a,b,c;z) again algebraic: for all arithmetic 'triangle-group' cases in Takeuchi's class I, they should be exactly the images of complex-multiplication points under explicit modular functions. A generalist should care because it converts a transcendence question into a checkable list of algebraic special values, with explicit numbers such as 2F1(1/4,1/4,1/2;9) = (2−i)/(2√2)","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Tables 3/4 do not implement Theorem 1: the (2,3,∞) cases are absent and one row lists a (6,3,4) datum under (2,4,∞).","rationale":"The reader's weakest assumption was Lemma 9, and I agree that the converse's period-lattice step is under-derived: the inference from an algebraic hypergeometric value to D(z)∈Q(√−d) needs a Q̄-linear form of Wüstholz's theorem and a constructed isogeny, not just the dimension statement in Theorem 7. However, this is a standard consequence of Wolfart–Wüstholz and is plausibly repairable by a more careful citation; the M=8 Milne computation also appears internally plausible. The table defects are more immediately fatal to the paper's advertised deliverable: an exhaustive classification of all nine class I groups whose main table omits the entire (2,3,∞) family and includes a datum with signature (6,3,4) is not a valid reference, however sound the surrounding machinery may be. If the omission is intentional because Archinard already treated PSL2(Z), the paper must say so explicitly; as written, Theorem 1 claims Table 4 determines all cases. I therefore keep the reader's CONDITIONAL verdict, but the required condition must include regenerating and verifying Tables 3 and 4.","tokens_in":22276,"tokens_out":30896,"duration_ms":305404,"concrete_test":"Write a script that, for each (a,b,c) listed in Tables 3 and 4, computes e1=1/|1−c| (∞ if 0), e2=1/|c−a−b|, e3=1/|a−b|, sorts them, and compares the signature with the nine signatures in Figure 1. It will flag (1/8,3/8,5/6) as (6,3,4) rather than (2,4,∞), and it will report no row with signature (2,3,∞)—e.g., (1/12,1/12,1/2) is absent. If the script confirms both facts, the paper must add the missing (2,3,∞) rows and remove or reclassify the (6,3,4) row before Theorem 1's exhaustive claim can stand.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The load-bearing claim is that Table 4 explicitly determines E(a,b,c) for every class I datum. Two table-level defects make that claim unsupported as written. First, the (2,3,∞) row is missing. The datum (1/12,1/12,1/2) satisfies 0<a,b,c<1 and gives {|1−c|,|c−a−b|,|a−b|} = {1/2,1/3,0}, i.e. the monodromy triangle group (2,3,∞) = PSL2(Z), a class I group. The same holds for (5/12,5/12,1/2) and (1/12,5/12,1/2), and for the c=2/3 family. None appears in Table 4, nor do the corresponding c≥1 cases appear in Table 3. Theorem 1 therefore does not determine all class I exceptional sets as stated. Second, Table 4 lists (1/8,3/8,5/6) under (2,4,∞), but its exponent triple is (1/6,1/3,1/4), so its Schwarz triangle group is (6,3,4), a cocompact group not in class I. This is a direct computation from the paper's own definitions, not a subtle interpretation. If the main table both omits an entire group and includes an extraneous row, the central classification cannot be used as a reference until Tables 3/4 are regenerated and checked.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims to determine the exceptional sets E(a,b,c) = { z ∈ Qbar : 2F1(a,b,c;z) ∈ Qbar } for every hypergeometric datum whose monodromy group is an arithmetic triangle group in Takeuchi's class I. The method combines hypergeometric–modular identities with transcendence results (Wüstholz, Schneider): algebraic values of the Hauptmodul at CM points give the forward inclusion, and a period-lattice comparison via Wüstholz's theorem is used for the converse. The main theorem asserts that for 0<a,b,c<1 the exceptional set is exactly {t(τ_d) : τ_d ∈ Q(√−d)∩H} for an explicit Hauptmodul t and d∈{1,2,3}, while for c≥1 it asserts E(a,b,c)={0}. Explicit tables (Tables 3, 4, 5) are provided.","tokens_in":22494,"tokens_out":8552,"duration_ms":84830,"significance":"If correct, the result would give the first complete explicit determination of exceptional sets for all nine non-compact arithmetic triangle groups commensurable with PSL2(Z), going beyond Archinard's PSL2(Z) case and connecting the existing transcendence criteria to explicit modular parametrizations. The paper's concrete computations are a strength: Example 3's linear algebra (c1=2, c2=i−1, α=i) checks out, Table 5's first row evaluates correctly from the stated identity at τ=i, and the c=1 cases are consistent with Schneider's theorem on K(k)/π. The explicit CM evaluations are also potentially useful for special L-value computations. However, the completeness and correctness of the central tables are currently not established, and the converse direction contains a significant unproved lattice-inclusion step. These are load-bearing issues, so the paper cannot be accepted in its present form.","major_comments":[{"comment":"The class I group (2,3,∞) = PSL2(Z) is entirely absent from the tables. For example, the data (1/12,1/12,1/2), (5/12,5/12,1/2), and (1/12,5/12,1/2) satisfy 0<a,b,c<1 and give {|1−c|,|c−a−b|,|a−b|} = {1/2,1/3,0}, so their monodromy triangle group is (2,3,∞), which appears in Figure 1 as a class I group. Likewise the c≥1 datum (7/12,7/12,3/2) has the same exponent triple. Neither Table 4 nor Table 3 contains any (2,3,∞) row, and Table 2 has no modular identity for it. Since Theorem 1 is quantified over all class I data, the main claim is false as stated; the classification is incomplete until these cases are added and checked.","section":"§4, Tables 3 and 4; Theorem 1"},{"comment":"The datum (1/8,3/8,5/6) is listed under the triangle group (2,4,∞), but its own exponent triple is |1−c|=1/6, |c−a−b|=1/3, and |a−b|=1/4. Hence its Schwarz triangle group is (6,3,4), a cocompact group that is not one of the non-compact class I groups. This row is therefore extraneous, and its entry (Q(i√2), t=(J4−1)/J4) is not justified by the method of the paper. The tables must be regenerated from a complete enumeration of data whose exponent triple matches one of the nine class I groups.","section":"Table 4, row (1/8,3/8,5/6)"},{"comment":"The converse direction of Theorem 1 depends on the assertion: 'Since Tz and T0 are defined over Q, there is a Q-linear transformation λ: C→C^{φ(N)} with the property λ(Λ(0)) ⊆ Λ(z).' This is not a consequence of Wüstholz's Theorem 7 as stated. That theorem yields isogenies between simple factors of abelian varieties, but converting an isogeny into a period-lattice inclusion requires identifying the relevant subvariety of Tz and specifying the complex-linear embedding of its period lattice into Λ(z). Moreover, the further step 'Due to the linearity of λ applied to τ_M, D_abc(z)∈Q(ζ_M)' is asserted rather than proved. For the M=8 cases the argument also relies on a CM-type computation (Φ={1,3}, H={1,3}) whose consequences for the period lattice are only sketched. Since Lemma 9 is the only mechanism for the 'only if' inclusion, this is a load-bearing gap that needs a complete proof.","section":"§4, Lemma 9"}],"minor_comments":[{"comment":"The statement reads '2F1(...) ∈ Ω_{−d}/Ω_{−d}·Q', which is simply Q. Presumably a ratio with a different Chowla–Selberg period (e.g., Ω_{−d}/Ω_{−x} as in equation (14)) was intended; please clarify.","section":"§5, Theorem 14"},{"comment":"Several typos and formatting ambiguities: 'Borechards forms' should be 'Borcherds forms'; 'Pochammer' should be 'Pochhammer'; in Example 3 the displayed formula '2τ+i−1/2i' lacks parentheses and should be '(2τ+i−1)/(2i)'; in Table 4 the entry split as 'R2' over 'R4−1' should be written as a single fraction such as R4/(R4−1).","section":"Throughout"},{"comment":"The comparison column cites the Hauptmoduln of [3, Appendix] but several entries use symbols (u, t3, t4^+, t6^+) that are not defined in the text. A brief definition or reference to the exact location in [3] would improve reproducibility.","section":"Table 1"}],"recommendation":"major_revision","confidential_remarks":"The manuscript has a plausible and interesting method, and the checked numerical examples lend it some credibility. However, the missing (2,3,∞) cases and the extraneous (6,3,4) row mean that the central tables are not a faithful implementation of Theorem 1, and Lemma 9's lattice-inclusion step is not proven. These are fixable in principle, so I recommend major revision rather than rejection. I would also ask the editor to verify that reference [12] (a privately communicated manuscript) is sufficiently accessible for the claims that rely on it."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read it. The paper does something genuinely useful: it extends Archinard's (2,3,∞) result to the other eight non-compact Takeuchi class I groups, and it produces explicit hypergeometric–modular identities plus concrete algebraic values. I checked several mechanical claims—Example 3's linear algebra, Table 5's first row, and the c≥1 triviality cases—and they are internally consistent. The identities in Table 2 and the explicit CM-evaluations in Table 5 are the kind of concrete output that makes this paper a reference if the classification holds.\n\nBut the central table does not currently support Theorem 1. The stress-test note is correct and easy to verify. The datum (1/12,1/12,1/2) gives {|1−c|,|c−a−b|,|a−b|} = {1/2,1/3,0}, i.e. the (2,3,∞) triangle group, which is class I. It appears nowhere in Table 4, and no c≥1 version appears in Table 3. So Theorem 1 as stated—covering every class I datum—is not implemented by the tables. Worse, Table 4 lists (1/8,3/8,5/6) under (2,4,∞), but its exponent triple is (1/6,1/3,1/4), which is the cocompact (6,3,4) triangle group, not class I at all. These are not subtle interpretative issues; they are direct consequences of the paper's own definitions.\n\nThe converse direction also has a real gap. Lemma 9 is the load-bearing step: it asserts that algebraicity forces an isogeny between simple factors of Tz and T0, then asserts a Q-linear λ with λ(Λ(0)) ⊆ Λ(z), and concludes D_abc(z) ∈ Q(√−d). The construction of λ is not given, and for M=8 the isogeny-type computation is assumed rather than justified. This is precisely the step that separates a conditional result from a complete one.\n\nThere are smaller issues: Theorem 7 states Wüstholz's theorem in a form over Q that is only valid over Q̄; completeness of Tables 3/4 is not argued; and Section 5 overclaims for class II, where Remark 15 admits that one datum is not determined. All of these are fixable in principle, but the tables must be regenerated and checked before the paper can serve as a reference.\n\nWho gets value from this? Specialists in hypergeometric transcendence and modular curves. The explicit identities and examples are the real payoff. I would not cite the classification as it stands, but the method and the modular identities deserve a serious referee. Send it to review, but expect major revision.","headline":"The explicit exceptional-set classification is real and worth refereeing, but Tables 3/4 have load-bearing defects and Lemma 9 is asserted rather than proved.","tokens_in":23150,"tokens_out":1946,"would_cite":false,"duration_ms":21970,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["33C05","11F03","11G15","11J89"],"pacs":[],"model":"deepseek-v4-flash","headline":"For every hypergeometric datum with 0<a,b,c<1 and class-I arithmetic triangle monodromy, the exceptional set equals the Hauptmodul image of imaginary quadratic points; if c≥1, it is just {0}.","keywords":["exceptional set","hypergeometric function","arithmetic triangle group","Hauptmodul","complex multiplication","periods","transcendence","modular forms"],"falsifier":"For the datum (1/8,1/8,1/2), verify the claimed isogeny type of the base abelian variety: if it is not isogenous to a square of an elliptic curve with CM by Q(√−2), the proof's conclusion D_abc(z)∈Q(√−2) collapses. Independently, scan algebraic z of small degree and height outside the listed t(τ_d) values and compute 2F1(a,b,c;z) to very high precision; any value matching an algebraic number to all computed digits would falsify the only-if direction.","tokens_in":21980,"feed_emoji":"🔢","tokens_out":8576,"duration_ms":84262,"temperature":0.7,"pith_summary":"This paper establishes a complete explicit description of the exceptional set E(a,b,c)—the algebraic inputs z for which the Gauss hypergeometric value 2F1(a,b,c;z) is again algebraic—for every datum with 0<a,b,c<1 whose monodromy group is an arithmetic triangle group in class I (the nine non-compact groups commensurable with PSL2(Z)). The description is: z is exceptional if and only if z=t(τ) for the associated Hauptmodul t evaluated at a point τ in Q(√−d)∩H, with d∈{1,2,3}; for data with c≥1 the set collapses to {0}. The paper gives explicit tables of t and d for every case, plus exact algebraic values such as 2F1(1/4,1/4,1/2;9)=(2−i)/(2√2). A reader should care because it turns a transcendence question—where a transcendental special function still lands in the algebraic numbers—into a finite, computable list for a natural infinite family, and it identifies the exact imaginary quadratic field forced by algebraicity.","feed_headline":"Algebraic inputs of nine hypergeometric functions fully listed","feed_subtitle":"For each class I case, algebraic values occur exactly at modular values on imaginary quadratic points.","key_machinery":"The work is carried by the inverse Schwarz map—a Hauptmodul t for the genus-zero triangle group—together with hypergeometric–modular identities that express 2F1(a,b,c;t(τ)) as a modular form (an eta-quotient or theta-quotient times a linear factor in τ). On the transcendence side, Euler's integral represents 2F1 as a quotient of periods of the abelian varieties T_z and T_0 built from the curves in the integral; a theorem on periods of abelian varieties then implies that algebraic values can only occur when the Schwarz map lands in an imaginary quadratic field. Classical algebraicity of modular functions at CM points supplies the reverse inclusion. The tables of Hauptmoduln and fields are wha","core_discovery":"The central claim (Theorem 1) is that for any datum (a,b,c) with 0<a,b,c<1 whose hypergeometric monodromy is an arithmetic triangle group in class I, there is a Hauptmodul t and a positive integer d∈{1,2,3} such that z∈E(a,b,c) if and only if z=t(τ_d) for some τ_d∈Q(√−d)∩H. If c≥1, E(a,b,c)={0}. The forward direction uses modular identities and the classical fact that Hauptmoduln take algebraic values at CM points; the converse uses period-theoretic transcendence: Euler's integral realizes 2F1 as a quotient of periods on abelian varieties, and if that quotient is algebraic, the periods force the Schwarz map value D_abc(z) into the imaginary quadratic field Q(√−d). Tables 1–4 list t and d for","pith_inferences":["The same mechanism should determine exceptional sets for any datum in which the beta integral reduces to a period on an elliptic curve; the paper already applies this to two class II data, suggesting a broader template.","The explicit CM evaluations of Hauptmoduln are ready-made algebraic constants for Ramanujan-Sato type hypergeometric evaluations and for L-values of CM modular forms; the paper notes this link without pursuing it.","A numerical census of small-height algebraic z and high-precision 2F1 values could test completeness directly; an unexpected algebraic value outside the listed images would pinpoint a failure in the period-lattice step.","If the period-lattice step is eventually proved in full generality, one would expect every arithmetic triangle group to have an exceptional set of this form, with the field Q(√−d) read off from the period data."],"forward_implications":["For the nine class I data, checking whether a given algebraic z is exceptional is a finite computation: evaluate the listed Hauptmodul at the CM points of the stated imaginary quadratic field.","For all listed data with c≥1, no nonzero algebraic z yields an algebraic hypergeometric value; 0 is the only exceptional point.","Exact algebraic values like 2F1(1/4,1/4,1/2;9)=(2−i)/(2√2) exist at specific CM points and are listed in Table 5.","The complete characterization replaces mere density or infinitude statements with an if-and-only-if description of the entire exceptional locus for this family."],"fun_headline_variants":["Exceptional sets for class I hypergeometric functions fully described","Algebraic values of hypergeometric functions pinned to CM points","Exact algebraic inputs listed for nine hypergeometric functions","Hypergeometric exceptional sets: complete table for class I"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The unproved linchpin is the period-lattice inclusion step: once 2F1(a,b,c;z) is assumed algebraic, the paper asserts rather than constructs a Q-linear map carrying the base period lattice into the z-period lattice, and it is this assertion that forces the Schwarz value into Q(√−d).","fun_headline_variants_meta":{"raw":{"variants":["Exceptional sets for class I hypergeometric functions fully described","Algebraic values of hypergeometric functions pinned to CM points","Exact algebraic inputs listed for nine hypergeometric functions","Hypergeometric exceptional sets: complete table for class I"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000494,"raw_usage":{"total_tokens":2255,"prompt_tokens":729,"completion_tokens":1526,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":473,"completion_tokens_details":{"reasoning_tokens":1469}},"tokens_in":473,"tokens_out":1526,"duration_ms":12296,"temperature":1.0,"reasoning_tokens":1469,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T02:25:05.638302+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For the datum (1/8,1/8,1/2), verify the claimed isogeny type of the base abelian variety: if it is not isogenous to a square of an elliptic curve with CM by Q(√−2), the proof's conclusion D_abc(z)∈Q(√−2) collapses. Independently, scan algebraic z of small degree and height outside the listed t(τ_d) values and compute 2F1(a,b,c;z) to very high precision; any value matching an algebraic number to all computed digits would falsify the only-if direction.","supporting_citations":[],"review_version":1}