{"id":"09a17eda-fe1e-41f1-a2cd-a98eaac6ad6f","arxiv_id":"2509.01849","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Finite imprimitive quaternionic reflection groups of rank two are fully classified using reflection systems, fixing an incomplete 1980 classification and adding previously missing groups.","lead":"This paper classifies the finite quaternionic reflection groups of rank two, correcting and simplifying a 1980 classification by Cohen. Readers interested in reflection groups, their symmetries, or their use in conformal field theory will find new groups, explicit presentations, and a new organizing tool called a reflection system.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Completeness of the classification turns on unshown finite enumerations of reflection systems for T, O, I; independent exhaustive verification is needed.","rationale":"The reader's weakest_assumption is the same point, and I agree. The finite subgroups of H and the dicyclic case are handled by explicit arguments in Section 5, while the polyhedral reflection systems are simply asserted. Because Theorem 6.1 is a classification, a single omitted reflection system is fatal to completeness; because the paper explicitly corrects and extends Cohen's list, the burden is on the enumeration. The provided Magma code is not enough, since it constructs the groups in the list rather than proving no other reflection systems exist. The deferred phrases in Section 2, Theorem 5.2, and Proposition 5.1 are further instances of the same unverified computational pattern, but they are secondary to the T/O/I enumeration. I would keep the paper's CONDITIONAL status: the argument is plausible and the dicyclic portion has real proof content, so rejection is not warranted; acceptance should wait for an independent exhaustive computation. Since the reader already returned CONDITIONAL for exactly this reason, my verdict adjustment is UNCHANGED.","tokens_in":25110,"tokens_out":12499,"duration_ms":139358,"concrete_test":"Independently enumerate the reflection systems of K=T, O, I by a backtracking/SAT search over subsets A containing 1: compute the closure under (a,b)->a*b^{-1}*a, retain those whose closure generates K, and reduce up to equivalence L~xL~Lx (x in L) and Aut(K). Compare systems and multiplicities with Examples 2.2, 4.3, and 4.5. Also verify each H in Table 1 satisfies H <| K, H subset L, LH=L, and recompute orders and reflection counts using Lemma 2.2. If the search returns exactly two T-systems, five O-systems, and four I-systems with the stated copy counts, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim of Theorem 6.1 is a complete classification, up to two explicit isomorphisms. Its completeness and uniqueness rest on finite computational assertions that are stated but not demonstrated. The most load-bearing is the inventory of reflection systems for the binary polyhedral groups: Example 4.3 gives five systems for O and Example 4.5 gives four for I, both introduced by 'Elementary calculations show', with no derivation or code. Earlier, Section 2 (after (2.14)) uses 'It turns out (from our calculations)' to rule out distinct canonical labels for the same K, and Theorem 5.2 defers the canonical-form check for ab odd with 'it can be shown'. A missing reflection system for T, O, or I would mean a missing line in Tables 1 and 2, so Theorem 6.1 would be incomplete; an invalid listed system would make it include a spurious group. This is a computational premise, not an internal inconsistency, and it is not covered by the explicit lemmas that handle the dicyclic case or by the provided Magma code, which constructs groups from indices rather than certifying the enumeration.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper gives an elementary classification of finite imprimitive irreducible rank-two quaternionic reflection groups. The main new tool is a \"reflection system\" L, a subset of a finite subgroup K of the unit quaternions containing 1 and closed under a∘b=ab^{-1}a. The author shows that any imprimitive group can be put in a canonical form G_K(L,H), with H a normal subgroup of K, and reduces the classification to enumerating reflection systems and then determining the admissible H. The cyclic, dicyclic, and binary polyhedral cases are treated; Tables 1 and 2 list the resulting groups. The classification theorem (Theorem 6.1) states that every such group has a unique canonical form except for two explicit isomorphism exceptions: G_O(L_O^{14},1) ≅ G_T(L_T^{12},C_2) and G(n,1,n,2) ≅ G(2n,2,n,1) for odd n. The paper also claims to correct and complete Cohen's 1980 table, including a previously missing group of order 192 with 22 reflections, and sketches a rank-n analogue.","tokens_in":25449,"tokens_out":7220,"duration_ms":78756,"significance":"If the completeness claims are correct, this is a substantial and useful contribution: it gives a transparent, structural classification of an important family, identifies gaps and double-counting in the existing literature, and provides explicit generators and reflection-orbit data for every group. The Section 5 dicyclic-group analysis is a genuine derivation: Lemma 5.2, Theorem 5.1, and Theorem 5.3 give real proofs, and the canonical-form lemma (Lemma 2.2) is clean and load-bearing. The paper is also careful to state exactly where it relies on calculations. However, the central classification claim is broader than the proved lemmas: the completeness of the reflection-system inventories for T, O, and I, and several uniqueness/no-isomorphism assertions, are presented as finite computations without a certified enumeration or reproducible check. The Magma code in Section 6 constructs groups from indices but does not certify that the lists of reflection systems are exhaustive. Thus the paper's main theorem is defensible but not yet fully supported.","major_comments":[{"comment":"The classification theorem depends critically on the assertions that the binary octahedral and icosahedral groups have exactly five and four reflection systems, respectively, and that each polyhedral K has at most one reflection system of a given size. These are introduced with \"Elementary calculations show\" (Examples 4.3 and 4.5) and \"It turns out (from our calculations)\" (after (2.14)). No derivation, pseudocode, or certified computation is supplied. A missing reflection system would remove a group from Tables 1 and 2; an invalid listed system would add a spurious group. This is a computational premise, not covered by the lemmas in Sections 2 and 5. Please provide a verifiable certificate: for example, an exhaustive closure computation over conjugacy classes of subsets, or a proof using the automorphism description (2.10) together with the explicit finite subgroup data in the appendix.","section":"Examples 4.3 and 4.5; after (2.14)"},{"comment":"Uniqueness of the canonical form for a fixed K is justified by the sentence \"It turns out (from our calculations) that this is always the case.\" This is load-bearing for the \"uniquely in canonical form\" part of Theorem 6.1. If, for some K, there were two inequivalent reflection systems of the same size or two distinct normal subgroups of the same order, then the order/reflection-count conditions in (2.14) would not distinguish the corresponding groups, and the uniqueness claim would fail. The normal-subgroup part for the listed K is easy to check, but the reflection-system uniqueness is exactly the unshown enumeration. Please supply the calculation or a complete argument.","section":"Section 2, case (i) after (2.14)"},{"comment":"The proof that the higher-order dicyclic groups G_{D_n}(L^{(n)}_{(a,b)}, C_{2n/ab}) are in canonical form for odd ab is deferred: the text says \"in can be shown\" (presumably \"it can be shown\") and gives no argument. This is load-bearing because it determines which H can be added to a reflection system without introducing new nondiagonal reflections; if the claim failed for some (a,b), Table 2 would contain invalid lines. The promised \"considering all the diagonal matrices\" case check should be written out or replaced by a complete, checkable computation. Also, the displayed \"C_{2n/ab}=⟨ω^{2b}⟩\" appears inconsistent with Table 2's \"C_{2n/ab}=⟨ω^{ab}⟩\" and with the convention C_r=⟨ω^{2n/r}⟩; please correct the theorem statement.","section":"Theorem 5.2 proof"},{"comment":"The exclusion of isomorphisms between polyhedral-group reflection groups and dicyclic-group reflection groups is argued by reducing to finitely many orders (48, 96, 192, 384, 480, 768) and then \"simply examin[ing] the reflection structure of each group, or their isomorphism class.\" Example 5.4 lists SmallGroup identifiers for the relevant orders, but it does not state the isomorphism tests performed or provide a script that certifies absence of further coincidences. This is also load-bearing for Theorem 6.1's uniqueness claim. Please make the finite check explicit and reproducible, and clarify the meaning of the \"*\" entries in Example 5.4 (which pairs are asserted to be isomorphic).","section":"Proposition 5.1 and Example 5.4"}],"minor_comments":[{"comment":"Typos and wording: Table 2 heading \"imprimtive\" should be \"imprimitive\"; Corollary 5.1 has \"descriminant\" for \"discriminant\"; Theorem 5.2 has \"in can be shown\" for \"it can be shown\"; Example 4.1 has \"occurences\" for \"occurrences.\"","section":"Various"},{"comment":"The title \"N =▽ SCFTs\" appears to contain a placeholder or rendering error; please check that the symbol is correct.","section":"Reference [DZ24]"},{"comment":"The edge label \"G(n,a,b,2n/ab)\" in Figure 3 applies only when ab is odd; the caption or figure should state this explicitly, as the surrounding text does.","section":"Figure 3"},{"comment":"The asterisk notation in the order-48 table is introduced only as \"with * indicating an isomorphism,\" but it is not immediately clear which starred entries are being identified with which. Please define the pairing explicitly.","section":"Example 5.4"}],"recommendation":"major_revision","confidential_remarks":"I agree with the reader's stress-test concern: the heart of the paper is sound, but the completeness of the classification rests on finite enumerations that are asserted rather than demonstrated. The manuscript can be fixed within its own scope by adding a certified exhaustive computation or a complete proof for the polyhedral reflection-system lists, the Section 2 uniqueness assertion, the deferred canonical-form check in Theorem 5.2, and the Proposition 5.1 isomorphism check. I saw no evidence of circularity: the derivations do not assume Theorem 6.1. The self-citations to [Wal24], [BW25], and [Wal25] appear peripheral rather than load-bearing."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a real correction to Cohen's classification. It proves there is a previously unknown order-192 imprimitive quaternionic reflection group with index [6,1,3,4], and it finds an infinite family of isomorphisms that Cohen's framework excluded. The dicyclic portion is the strongest part; the polyhedral portion is the part I would press on.\n\nWaldron's reflection-system formalism is a good organizing device. The closure condition ab^{-1}a in L comes directly from conjugating reflections, and Lemma 2.2 gives clean formulas |G|=2|H||K| and 2|H|+|L|-2 reflections. The classification of D_n is genuinely derived: Lemma 5.2 determines the L(n)_{a,b} with distinct sizes, Theorem 5.1 gives base groups, and Theorem 5.3 finds the only infinite index pair [n,1,n,2] and [2n,2,n,1] for n odd. That is real work. The order-192 group [6,1,3,4] and the row counts in Example 5.3 are concrete and consistent with the stated framework.\n\nNow the soft spots. Examples 4.3 and 4.5 list the reflection systems for O and I—five and four respectively—with only “Elementary calculations show”, no derivation, no code, and no certificate. The uniqueness point after (2.14) sits on “It turns out (from our calculations)”. Part of Theorem 5.2's canonical-form check is deferred with “it can be shown”. None of these is an internal inconsistency, and I do not think the central argument collapses. But Theorem 6.1 is a completeness claim: if one reflection system for T, O, or I is missing, Tables 1 and 2 miss a group. The Magma code in Section 6 constructs groups from the index set Λ_n; it does not certify that the polyhedral reflection-system inventory is exhaustive, nor that the “no further isomorphisms” statements in Example 4.5 and Proposition 5.1 are machine-checked. This is a computational premise that should be made reproducible.\n\nCitation pattern is fine. The self-citations [Wal24], [BW25], [Wal25] are peripheral. No fitted-parameter smell.\n\nWho gets value: anyone working on quaternionic or complex reflection groups, finite collineation groups, symplectic reflection groups, and the CFT applications cited in the introduction. It deserves a serious referee. My recommendation: send it out, and in review ask for the polyhedral enumeration to be either proved or moved into transparent code with a certificate. If that is supplied, this should become the standard reference. My own verdict is conditional-accept as it stands.","headline":"A real correction to Cohen's classification, with the dicyclic part solidly argued but the polyhedral reflection-system lists asserted rather than demonstrated; refereeing should press on the enumeration.","tokens_in":25830,"tokens_out":3486,"would_cite":true,"duration_ms":37049,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05B30","15B33","20C25","20G20","51M05","51M20","15B57","51E99"],"pacs":[],"model":"deepseek-v4-flash","headline":"Every imprimitive rank-two quaternionic reflection group has a canonical form, with two known duplicates; the paper gives the complete corrected list.","keywords":["imprimitive quaternionic reflection groups","rank two","reflection systems","binary polyhedral groups","dicyclic groups","finite subgroups of unit quaternions","classification","monomial matrices"],"falsifier":"Run an exhaustive computer search over all subsets of the binary tetrahedral, octahedral, and icosahedral groups that contain 1 and are closed under ab^{-1}a; if any subset appears that is not one of the systems listed in Examples 4.3 and 4.5, the classification is incomplete. Similarly, finding one unlisted isomorphism between groups from Tables 1 and 2 would refute the uniqueness claim.","tokens_in":25001,"feed_emoji":"🧮","tokens_out":6925,"duration_ms":74816,"temperature":0.7,"pith_summary":"The paper aims to give a complete, elementary classification of the finite imprimitive irreducible quaternionic reflection groups of rank two, meaning finite groups generated by reflections, i.e. linear maps that fix a one-dimensional subspace of a two-dimensional quaternionic space. It shows that each such group is determined by a 'reflection system': a subset L of a finite subgroup K of the unit quaternions that contains 1, generates K, and is closed under the binary operation a∘b=ab^{-1}a. The main theorem lists all these groups in canonical form in two tables and proves the form is unique except for two explicit isomorphisms. The classification corrects and extends Cohen's earlier list: for example, there are four non-isomorphic imprimitive groups of order 192 with 22 reflections, one of which had not been previously identified. The point of the elementary approach is that questions about reflection groups become finite algebraic computations over the classical finite quaternion groups: cyclic, dicyclic, and binary tetrahedral, octahedral, and icosahedral.","feed_headline":"One rule classifies all rank-two quaternionic reflection groups","feed_subtitle":"A closure condition on quaternion subgroups gives the full corrected list, including a missed order-192 group.","key_machinery":"The central object is the reflection system: for a finite subgroup K of the unit quaternions, a reflection system is a subset L with 1∈L, K=⟨L⟩, and closure under a∘b=ab^{-1}a. Closure encodes the algebra of products and conjugates of the off-diagonal reflection matrices; products of such reflections naturally produce diagonal entries from L, and the operation a∘b keeps L closed under conjugation-like moves. Each reflection system L, together with a normal subgroup H, yields the canonical group G_K(L,H), so enumerating reflection systems replaces the classification of reflection groups by finite combinatorics. For the dicyclic groups the enumeration reduces to a parameter set Ω_n of coprime","core_discovery":"The central claim is Theorem 6.1: every finite imprimitive irreducible rank-two quaternionic reflection group can be written uniquely in the canonical form G_K(L,H), except for two cases that have two canonical forms: G_O(L_O^14,1) is isomorphic to G_T(L_T^12,C_2), and G(n,1,n,2) is isomorphic to G(2n,2,n,1) for odd n. Here K is one of the finite subgroups of the unit quaternions, L is a reflection system for K, and H is a normal subgroup satisfying H⊂L and LH=L; the canonical form consists of all matrices with a diagonal part from K, a coset factor determined by L, and the optional swap matrix. The proof enumerates reflection systems for each possible K, computes the base reflection group f","pith_inferences":["The paper's reliance on hand-enumerated reflection systems for the binary tetrahedral, octahedral, and icosahedral groups suggests a natural testable extension: an exhaustive machine search over all subsets of these groups closed under a∘b would independently certify completeness.","Because the operation a∘b=ab^{-1}a depends only on the underlying group multiplication, the reflection-system method should transfer to imprimitive reflection groups over other finite subgroups of the unit quaternions, or to analogous noncommutative coefficient rings, not just the quaternionic case treated here.","The infinite families of index pairs described in Corollary 5.1 point to many non-isomorphic groups that agree in order and reflection count; the reflection-orbit invariant is likely the practical distinguisher, and it may reveal further unknown groups at higher orders."],"forward_implications":["If the classification is correct, Tables 1 and 2 form the complete roster: every imprimitive irreducible rank-two quaternionic reflection group appears exactly once, except the two flagged duplicate labels.","The canonical form gives immediate structural data, including the order |G|=2|H||K| and the number of reflections 2|H|+|L|-2, without constructing the group.","Several previously unlisted groups enter the classification, such as the order-192 group with index [6,1,3,4], and the corrected list reconciles earlier counts of imprimitive versus primitive-conjugate groups.","For rank n≥3, the same setup yields the canonical form G_n(K,H), so the higher-rank imprimitive classification follows once the rank-two case is settled."],"supporting_citations":[{"why":"Baseline classification this paper extends and corrects; supplies the earlier Table I comparison and the flawed isomorphism lemma.","marker":"[Coh80]"},{"why":"Complex reflection group classification and the G(n,p,2) notation used as the model and comparison class.","marker":"[ST54]"},{"why":"Source for the list of finite subgroups of the unit quaternions used as candidate groups K.","marker":"[LT09]"},{"why":"Original Stringham enumeration of finite quaternion groups underlying the binary polyhedral subgroups.","marker":"[Str81]"},{"why":"Supplies the reflection-orbit invariant used to distinguish non-isomorphic groups with the same order and reflection count.","marker":"[Wal25]"},{"why":"Observation that certain imprimitive groups are conjugate to primitive complex reflection groups, used in the final isomorphism count.","marker":"[Tay25]"}],"fun_headline_variants":["Simple rule classifies all rank-two quaternionic groups","Canonical form corrects quaternionic reflection group list","Missing order-192 quaternionic groups now classified","Reflection systems yield elementary group classification","One closure condition fixes quaternionic group gaps"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The classification is complete only if the hand listings of reflection systems for the binary tetrahedral, octahedral, and icosahedral groups are complete; the paper states these listings come from 'elementary calculations' without giving an exhaustive search procedure.","fun_headline_variants_meta":{"raw":{"variants":["Simple rule classifies all rank-two quaternionic groups","Canonical form corrects quaternionic reflection group list","Missing order-192 quaternionic groups now classified","Reflection systems yield elementary group classification","One closure condition fixes quaternionic group gaps"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000443,"raw_usage":{"total_tokens":2013,"prompt_tokens":612,"completion_tokens":1401,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":356,"completion_tokens_details":{"reasoning_tokens":1326}},"tokens_in":356,"tokens_out":1401,"duration_ms":12433,"temperature":1.0,"reasoning_tokens":1326,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T12:07:09.957765+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run an exhaustive computer search over all subsets of the binary tetrahedral, octahedral, and icosahedral groups that contain 1 and are closed under ab^{-1}a; if any subset appears that is not one of the systems listed in Examples 4.3 and 4.5, the classification is incomplete. Similarly, finding one unlisted isomorphism between groups from Tables 1 and 2 would refute the uniqueness claim.","supporting_citations":[],"review_version":1}