{"id":"614fb8e4-7b58-46c7-8231-645ae72d33c7","arxiv_id":"2411.18102","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Finite groups are classified by the number of conjugacy classes of nontrivial non-self-normalizing subgroups for small counts, with bounds on nilpotency class and derived length in general.","lead":"This paper studies finite groups by counting, up to conjugacy, the nontrivial subgroups that are not equal to their own normalizer. It classifies groups with 0, 1, 2, or 3 such classes and gives structural bounds for larger counts.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The D4 dichotomy rests on two unverified computational assertions in Section 8: the D-values for A6, L2(7), L2(8) and the nonexistence of order-48 groups with nonabelian Sylow 2 equal to G'.","rationale":"The reader's weakest-assumption analysis already identified the unverified computational assertions in Section 8 as the main risk, and my stress-testing concurs. Theorems 8.2 and 8.3 are the central claims, and both end in finite case eliminations that are not reproducible from the text. Theorem 8.2 uses a 'direct check' to assign D-values to A6, L2(7), and L2(8), while Theorem 8.3 uses a GAP assertion about groups of order 48. These are load-bearing because they rule out all alternatives to A5 and SL2(3) in D4; without them the dichotomy reduces to an unverified finite list. I found no internal logical contradiction in the surrounding arguments, and the small cases I checked (D0, D1, the Frobenius counts in Lemma 5.1, and the D2/D3 classifications) are consistent. The appropriate remedy is not rejection but a condition: release the computational checks. Since the reader already marked the paper CONDITIONAL for this reason, the verdict should remain unchanged.","tokens_in":20837,"tokens_out":6366,"duration_ms":53005,"concrete_test":"Write a short GAP or Magma script and run it once. (1) For G in [A5, A6, L2(7), L2(8), L2(17), L3(3), U3(3), U4(2)], compute D(G) by iterating over ConjugacyClassesSubgroups and counting classes whose representative H satisfies 1 < H < G and N_G(H) != H; verify the values 4, 11, 8, 6 for the first four and that the last four do not lie in D4. (2) Enumerate AllSmallGroups(48) and test whether any group g has a nonabelian Sylow 2-subgroup coinciding with DerivedSubgroup(g); the count should be 0. If the script and output are supplied as a supplement, the Section 8 proof is no longer conditional on private computation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central D4 dichotomy (Theorems 8.2 and 8.3) is logically reduced to a short list of finite cases, but the decisive elimination of those cases is asserted rather than demonstrated. In Theorem 8.2, after citing Herzog's list of simple groups with three prime divisors, the proof says 'the last four cannot lie in D4 because of the orders of their Sylow subgroups, whereas a direct check rules out all the others except A5' and reports A6 in D11, L2(7) in D8, L2(8) in D6. No code, transcript, or counting formula is supplied, and these D-values are exactly what separates A5 from the other groups in the list. In Theorem 8.3, the final case |G| = p^3 q, p = 3, q = 2, uses 'it can be checked via GAP that there do not exist groups of order 48 with a nonabelian Sylow 2-subgroup coinciding with the derived subgroup' to force |N| = 2 and hence G isomorphic to SL2(3). This is a nontrivial finite enumeration over all groups of order 48. If either computational assertion is wrong, the uniqueness of A5 or of SL2(3) in D4 fails and the dichotomy would need re-examination. The paper's own text flags the GAP check as an external computation, so this is a genuine gap in reproducibility, not a notational issue.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces and studies the family D_n of finite groups with exactly n conjugacy classes of nontrivial subgroups that are not self-normalizing. The main structural results are: groups in D_n for n ≤ 3 are solvable with derived length at most 2 (Proposition 3.3); A5 is the unique nonsolvable group in D4 (Theorem 8.2); SL2(3) is the unique solvable group in D4 with derived length greater than 2, all other solvable groups in D4 having derived length at most 2 (Theorem 8.3). The paper also proves a lower bound for D(H × K) with an equality criterion (Lemma 4.2), bounds on the nilpotency class and derived length of nilpotent groups in D_n (Theorems 4.1 and 4.3), a classification of D0–D3 (Propositions 3.1, 3.4, 7.1, 7.2), and a general bound on the derived length of solvable groups in D_n (Theorem 8.4). The methods are elementary group theory: Sylow theory, Burnside's normal p-complement theorem, Frobenius group structure, Maschke's theorem, and Z-group structure.","tokens_in":21063,"tokens_out":23015,"duration_ms":188176,"significance":"If the D4 classification is correct, it is a clean and complete answer for the first nonsolvable value of n, and the paper provides several reusable tools: the factor-group lemma (Lemma 3.2), the product inequality (Lemma 4.2), and the Frobenius-group counts (Section 5). The general derived-length bound (Theorem 8.4) and the nilpotency-class bounds are reasonable first steps toward a broader theory. The paper is well-written and the non-computational arguments are coherent, relying on standard theorems. The manuscript also honestly states its computational steps and points to related work by Jones. The main weakness is that the two decisive finite computations in Section 8 are asserted without reproducible code or a complete hand-checkable derivation, which means the central D4 dichotomy is not fully demonstrated as written.","major_comments":[{"comment":"The proof eliminates all simple groups in Herzog's list except A5 by asserting that 'a direct check rules out all the others except A5' and reporting D(A6)=11, D(L2(7))=8, D(L2(8))=6, while the last four groups are dismissed 'because of the orders of their Sylow subgroups.' These assertions are load-bearing: without them the theorem only shows that a nonsolvable group in D4 must be almost simple with socle on this list. Since neither the D-values nor the Sylow-order argument are proved or documented, this step is not reproducible. Please provide the GAP/Magma code and output, or a hand-checkable counting argument, for each of the eight groups in the list, including the four groups excluded by the orders of their Sylow subgroups.","section":"Section 8, Theorem 8.2"},{"comment":"The proof uses the assertion 'it can be checked via GAP that there do not exist groups of order 48 with a nonabelian Sylow 2-subgroup coinciding with the derived subgroup' to conclude |N|=2 and hence G≅SL2(3). This is a nontrivial finite enumeration over all groups of order 48 and is essential to the uniqueness of SL2(3) in D4 with derived length 3. The paper does not include the GAP code, the exact search predicate, or the output. Please provide these, or give a mathematical proof that no such group exists.","section":"Section 8, Theorem 8.3"}],"minor_comments":[{"comment":"In the sentence 'the nilpotency class of D(P_j) is 1', the symbol D(P_j) appears to be a typo for P_j; the statement should concern the nilpotency class of the Sylow subgroup P_j.","section":"Theorem 4.3 proof"},{"comment":"The reference 'By Proposition 8.2, it follows that G is solvable' should be 'By Theorem 8.2', since Proposition 8.2 does not exist.","section":"Theorem 8.3 proof"},{"comment":"In the proof of the forward direction for the Frobenius group C_pr ⋊ C_q, the text says G has three normal subgroups 'of orders p, q, pr respectively'; the subgroup of order q is the Frobenius complement and is not normal in a nontrivial Frobenius group, so this should read 'of orders p, r, pr respectively'.","section":"Proposition 7.2, case (6)"},{"comment":"The displayed formulas for D(G) in the statements of Lemmas 5.2 and 5.3 have ambiguous fraction formatting; for example, 'D p+1 q +1' should be typeset as D_{(p+1)/q+1} to avoid confusion between (p+1)/q+1 and (p+1)/(q+1).","section":"Section 5, Lemmas 5.2 and 5.3"}],"recommendation":"major_revision","confidential_remarks":"The central results appear mathematically plausible and the paper is a good fit for this journal. The only obstruction to acceptance is the lack of reproducible evidence for the two finite computations in Section 8; I am confident the authors can supply GAP/Magma scripts and output or a hand-checkable proof in an appendix. Once those are added, I would support acceptance. The paper is clearly written and the D4 classification, together with the product and nilpotency bounds, constitutes a solid contribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things about this paper. First, it is a real piece of finite group theory: it introduces a new counting invariant D(G), classifies D0 through D3 completely, proves that A5 is the only nonsolvable group in D4, and shows SL2(3) is the only solvable D4-group of derived length 3. Second, the D4 dichotomy is the headline result, and it is probably right, but it rests on a couple of finite computational assertions that are stated without code or transcripts.\n\nWhat the paper does well: the counting arguments are coherent and mostly checkable by hand. Lemma 3.2 (the quotient bound) is a nice tool, the Frobenius counts in Section 5 are explicit and I spot-checked several of them, and the nilpotent bounds via the direct-product inequality are clean. The authors are also honest about the limits of their methods: they say the D2/D3 lists are already complicated, they flag the GAP check explicitly, and they do not oversell the logarithmic bound in Theorem 8.4, which they admit is much larger than what computations suggest. Citation practice looks fair; the Herzog list and Glasby's theorem are standard, and the self-citation to the related non-self-centralizing paper is appropriate.\n\nSoft spots, in proportion: the main one is Section 8. In Theorem 8.2, after reducing to Herzog's eight simple groups, the proof says a direct check rules out all but A5 and reports D(A6)=11, D(L2(7))=8, D(L2(8))=6. Those values are load-bearing: they separate A5 from the rest. In Theorem 8.3, the final case |G|=48 uses a GAP assertion that no group of order 48 has a nonabelian Sylow 2-subgroup equal to the derived subgroup. Neither check comes with code, a transcript, or a counting formula. This is a reproducibility gap, not a flaw I can point to in the logic—I verified several of the smaller counts myself and found no error—but it should be fixed before the paper is final. Minor: Proposition 3.3 asserts S4 lies in D7 without proof, and the phrase direct check appears a couple of times where a short argument would be better. Both are easily patched.\n\nOverall: the central argument holds up on reading. The classifications are new, the reductions are careful, and I do not see a logical hole. The paper is for finite group theorists who work on subgroup lattices, normalizers, or solvability criteria. It deserves a serious referee; the referee should ask for the computational evidence to be made available, ideally as a short appendix or supplementary file. I would send it to peer review with that condition.","headline":"Genuinely new small-n classification of a subgroup-counting invariant, with a likely-correct D4 dichotomy that needs reproducible computational checks.","tokens_in":21663,"tokens_out":1241,"would_cite":true,"duration_ms":13821,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["20E07","20E34","20D10","20D15"],"pacs":[],"model":"deepseek-v4-flash","headline":"Counting conjugacy classes of non-self-normalizing subgroups yields a complete description of $D_0$ through $D_4$, with the only nonsolvable group in $D_4$ being $A_5$.","keywords":["non self-normalizing subgroups","conjugacy classes","solvable groups","derived length","nilpotency class","Frobenius groups","finite simple groups","D4 classification"],"falsifier":"Re-run the two finite searches that Theorem 8.2 and Theorem 8.3 depend on: compute $D(G)$ for every simple group whose order has exactly three prime divisors, and enumerate all groups of order 48 to see whether any has a nonabelian Sylow 2-subgroup coinciding with its derived subgroup. A single counterexample in either search would overturn the claimed uniqueness.","tokens_in":20599,"feed_emoji":"🎯","tokens_out":11072,"duration_ms":82709,"temperature":0.7,"pith_summary":"Finite groups can be sorted by a single number: $D(G)$, the number of conjugacy classes of nontrivial subgroups that are not self-normalizing. The paper studies the families $D_n$ of groups with $D(G)=n$ and asks how this one number constrains solvability, derived length, and nilpotency class. Its main result is a complete dichotomy for $n=4$: a nonsolvable group in $D_4$ must be the alternating group $A_5$, and every other group in $D_4$ is solvable, with derived length at most $2$ unless it is isomorphic to $SL_2(3)$. Along the way it classifies $D_0$ through $D_3$, proves those groups are solvable of derived length at most $2$, bounds nilpotency class and derived length for nilpotent groups in $D_n$, and obtains a general derived-length bound for solvable groups in $D_n$. If the paper is right, the single invariant $D(G)$ determines group structure much more tightly than the mere absence of self-normalizing subgroups would suggest.","feed_headline":"A5 is the only nonsolvable group in D4","feed_subtitle":"Counting non-self-normalizing subgroups: four classes leave only A5 and SL2(3) as exceptions.","key_machinery":"The engine of the paper is the invariant $D(G)$, the number of $G$-conjugacy classes of nontrivial subgroups $H$ with $N_G(H)\\neq H$. The load-bearing tool is the quotient lemma (Lemma 3.2): if $N$ is a nontrivial proper normal subgroup, then $D(G/N)\\le D(G)-1$, because every non-self-normalizing subgroup class of the quotient lifts to one of $G$ not properly contained in $N$. This lets the authors descend from a group to its quotients, reducing questions about $D_n$ to small cases. For the $D_4$ dichotomy, the final step combines this descent with the finite list of simple groups whose order is divisible by exactly three primes, and with a computer search over groups of order $48$. The Frobenius-group counts in Section 5 provide the explicit classifications of $D_2$ and $D_3$.","core_discovery":"Define $D(G)$ as the number of conjugacy classes of nontrivial subgroups $H$ of a finite group $G$ with $N_G(H)\\neq H$; $D_n$ is the family of groups with $D(G)=n$. The central claim is that $D_4$ is completely understood: by Theorem 8.2, if $G$ is nonsolvable and $G\\in D_4$, then $G\\cong A_5$; by Theorem 8.3, every other group in $D_4$ is solvable and either has derived length at most $2$ or is isomorphic to $SL_2(3)$. The same invariant yields explicit classifications for $n\\le 3$: all groups in $D_0,D_1,D_2,D_3$ are solvable with derived length at most $2$, with the exceptional examples listed as cyclic, Frobenius, and $A_4$ groups. Nilpotent groups in $D_n$ have nilpotency class at most $n/2$ and derived length at most $\\log_2(n/2)+1$, and solvable groups in $D_n$ have derived length at most $\\min(n-1,3\\log_2(n+1)+9)$. The paper also proves a product inequality $D(H\\times K)\\ge (D(H)+2)(D(K)+2)-2$, with equality exactly when $H$ and $K$ are nilpotent of coprime orders.","pith_inferences":["The $D_4$ dichotomy makes it plausible that $D_5$ contains only solvable groups; enumerating the finite groups of the relevant orders would test this directly.","The product inequality suggests that direct products of nilpotent groups of coprime orders realize many $D$-values exactly, so one could ask whether every sufficiently large integer occurs as $D(G)$ for some finite group.","The uniqueness of $A_5$ and $SL_2(3)$ currently rests on unscripted computer checks described in Section 8; formalizing those two finite searches would make the classification machine-checkable, and any error there would reopen the theorem.","The paper's own computational experiments indicate the bound $3\\log_2(n+1)+9$ is far from sharp; improving it, or showing that all $D_5$ groups have derived length at most $3$, would be the next quantitative step."],"forward_implications":["Every group with $D(G)\\le 3$ is solvable of derived length at most $2$, so small values of the invariant force solvability.","The only nonsolvable group with $D(G)=4$ is $A_5$; no other simple or almost-simple group has exactly four conjugacy classes of non-self-normalizing subgroups.","Every solvable group in $D_4$ has derived length at most $3$, and the only one reaching $3$ is $SL_2(3)$.","Nilpotent groups in $D_n$ have nilpotency class at most $n/2$ and derived length at most $\\log_2(n/2)+1$, so the count of non-self-normalizing subgroup classes bounds how deep the group is.","For solvable groups, membership in $D_n$ bounds derived length by $\\min(n-1,3\\log_2(n+1)+9)$, a constraint that grows only logarithmically in $n$."],"supporting_citations":[{"why":"Supplies the finite list of simple groups whose order is divisible by exactly three primes, used in the final reduction of Theorem 8.2.","marker":"[7]"},{"why":"The computational algebra system used for the order-48 nonexistence check that completes Theorem 8.3.","marker":"[4]"},{"why":"Provides foundational facts on normalizers of Sylow subgroups, nilpotent groups, Z-groups, and Frobenius groups used throughout.","marker":"[13]"},{"why":"Used in Theorem 8.3 for the structure of $G/N$ when $G/N$ is not abelian, and in Proposition 3.3 for Burnside's $p^a q^b$ theorem.","marker":"[12]"},{"why":"Gives the general derived-length bound for solvable groups in terms of composition length, used in Theorem 8.4.","marker":"[6]"},{"why":"Supplies the Correspondence Theorem used repeatedly to transfer conjugacy-class counts between groups and quotients.","marker":"[11]"}],"fun_headline_variants":["A5 and SL2(3) are the only D4 exceptions","D4: A5 nonsolvable, SL2(3) exceptional","For D(G)=4, A5 is the only nonsolvable","D4's odd ones: A5 and SL2(3)","A5 and SL2(3) break D4 patterns"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the finite computational assertions in Section 8 are correct: the paper reports a 'direct check' giving $A_6\\in D_{11}$, $L_2(7)\\in D_8$, $L_2(8)\\in D_6$ while leaving $A_5$ as the only simple candidate, and asserts that a computational search finds no group of order 48 with a nonabelian Sylow 2-subgroup equal to its derived subgroup. If either of these unscripted checks is wrong, the uniqueness of $A_5$ and $SL_2(3)$ in $D_4$ would need re-examination.","fun_headline_variants_meta":{"raw":{"variants":["A5 and SL2(3) are the only D4 exceptions","D4: A5 nonsolvable, SL2(3) exceptional","For D(G)=4, A5 is the only nonsolvable","D4's odd ones: A5 and SL2(3)","A5 and SL2(3) break D4 patterns"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000929,"raw_usage":{"total_tokens":4100,"prompt_tokens":1190,"completion_tokens":2910,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":806,"completion_tokens_details":{"reasoning_tokens":2816}},"tokens_in":806,"tokens_out":2910,"duration_ms":20865,"temperature":1.0,"reasoning_tokens":2816,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T11:30:55.597283+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-run the two finite searches that Theorem 8.2 and Theorem 8.3 depend on: compute $D(G)$ for every simple group whose order has exactly three prime divisors, and enumerate all groups of order 48 to see whether any has a nonabelian Sylow 2-subgroup coinciding with its derived subgroup. A single counterexample in either search would overturn the claimed uniqueness.","supporting_citations":[{"cited_title":"Herzog, On Finite Simple Groups of Order Divisible by Three Primes On ly, J","cited_arxiv_id":null,"evidence_quote":"Supplies the finite list of simple groups whose order is divisible by exactly three primes, used in the final reduction of Theorem 8.2."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The computational algebra system used for the order-48 nonexistence check that completes Theorem 8.3."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides foundational facts on normalizers of Sylow subgroups, nilpotent groups, Z-groups, and Frobenius groups used throughout."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Used in Theorem 8.3 for the structure of $G/N$ when $G/N$ is not abelian, and in Proposition 3.3 for Burnside's $p^a q^b$ theorem."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the general derived-length bound for solvable groups in terms of composition length, used in Theorem 8.4."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Correspondence Theorem used repeatedly to transfer conjugacy-class counts between groups and quotients."}],"review_version":1}