Finding 59:29 in the Monster

Aluna Rizzoli; Anthony Pisani; Heiko Dietrich; Melissa Lee

REVIEW 3 major objections 6 minor 15 references

Finding 59:29 in the Monster

T0 review · 3 major / 6 minor · reviewed 2026-08-03 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The paper supplies explicit generators for the Monster group's final maximal subgroup 59:29, completing the explicit list for all maximal subgroups.

desk verdict A short, credible computational capstone for the Monster maximal subgroup classification, with the main theorem resting on an explicit mmgroup computation. read the letter →

arxiv 2607.28900 v1 pith:EA4UCUI7 submitted 2026-07-30 math.GR

Heiko Dietrich , Melissa Lee , Anthony Pisani , Aluna Rizzoli This is my paper · ORCID

classification math.GR MSC 20D0820-0820E28

keywords Monstergroupmaximalsubgroups59:29explicitgeneratorsfingerprintingcomputationaltheoryPSL2(59)sporadicsimplegroups

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper closes the last gap in the explicit construction of maximal subgroups of the Monster, the largest sporadic simple group. It exhibits two elements y and n, of orders 59 and 29, satisfying y^n = y^3, and verifies computationally that they generate a copy of the maximal subgroup 59:29. Because earlier work had already established uniqueness and maximality of this 59:29 class, the explicit generators complete the program of giving concrete generators for every maximal subgroup. The paper also uses the construction to give a short new proof that the Monster has no subgroup isomorphic to PSL2(59), correcting a previously published claim.

What carries the argument

The engine of the search is a chain of fingerprint reductions. A fingerprint for a 2A-axis (a distinguished fixed vector of a 2A-involution in the 196,884-dimensional representation) is a tuple of inner products that is invariant under the desired conjugation; matching fingerprints give candidate witnesses. Candidate pairs are refined by pushing a conjugacy problem down the subgroup chain M > 2.B > 2^{1+23}.Co2 > 2^{1+23}, using matching fingerprints at each level and a stabiliser-chain argument until a brute-force search in the 2^{24}-element group 2^{1+23} completes the job. Birthday-paradox estimates govern how many random axes are needed. The decisive identity is y^n = y^3, defining the

What would settle it

Run the code in Figure 1 with an independent implementation of the Monster: if the orders and the relation y^n=y^3 do not hold, the theorem collapses. Equivalently, exhibit an involution s in the supposed PSL2(59) with n^s=n^{-1} and |ys|=3, which the calculation in Figure 1 says cannot occur among the 29 listed inverters.

Watch

Extended reading notes

Core claim

The central result, Theorem 1, states that the elements y and n defined by the computation code in Figure 1 generate a maximal subgroup 59:29 of the Monster; specifically, they satisfy the presentation <y,n | y^59, n^29, y^n=y^3>. The proof is the computation that evaluates these relations as true in the software's implementation of the Monster and its 196,884-dimensional representation, together with the earlier classification showing uniqueness of the 59:29 class and its maximality. Corollary 2 then shows the Monster has no subgroup PSL2(59): working with the centraliser of n, of order 87, the authors enumerate all 29 involutions inverting n and verify computationally that none sends y to

Load-bearing premise

The computation in Figure 1 is trustworthy only if the software package the authors use correctly implements the Monster and its 196,884-dimensional representation; the earlier classification's uniqueness and maximality statements carry the rest.

Editorial extensions

If this is right

Every maximal subgroup of the Monster now possesses explicit generator words, not just an abstract classification.
The explicit 59:29 can be used in further computations inside the Monster, for example in testing subgroup structure or character-theoretic assertions.
The Monster's non-containment of PSL2(59) now has a short computational proof, replacing a longer refutation and correcting a published claim.
The chain-reduction fingerprinting method provides a template for constructing subgroups in other large finite groups where direct random search is infeasible.
The generators are certified only up to the correctness of the computation package, so independent verification in another implementation would firmly settle the explicit construction.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The same fingerprint-and-chain technique could plausibly be adapted to other sporadic groups or to constructing different local subgroups in the Monster, not just 59:29.
The correction of the PSL2(59) claim may have downstream consequences for papers that relied on that earlier result, although the authors do not survey them.
Because the proof hinges on a single software evaluation, a fully formal or independent computer proof would make the construction a theorem in the strongest sense; the authors note their computations do not require trusting AI-generated reasoning.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Colleague,

This paper closes the last open explicit-generator case in the Monster maximal subgroup classification. That is the thing to know: Dietrich and coauthors provide explicit mmgroup words for y and n that generate 59:29, and verify the presentation y^59, n^29, y^n = y^3 with a short, re-runnable Python snippet. It is a real computational capstone, not a repackaging.

What is genuinely new: the construction of the generators themselves, and a multi-step fingerprint descent through K = 2·B, H = 2^{1+23}.Co2, Q = 2^{1+23} that is more involved than the Baby Monster predecessor. The approach is clearly explained, and the authors are upfront that uniqueness and maximality of the 59:29 class are cited from their earlier classification papers [3,4]. That is acceptable—those are the papers that did that heavy lifting.

The soft spots are minor but real. Theorem 1's proof is essentially the Python expression in Figure 1 evaluating True. That makes the correctness of mmgroup 1.0.8 load-bearing. The code is explicit and published, so the fix is independent execution, not more argument. A referee should run it. Corollary 2 has an unshown 'direct calculation' in PSL2(59); it is true (the stress-test note gives a concrete example), but the authors should display the two-line calculation instead of leaving the reader to dig. Also, Remark 18 discloses that the search pipeline was largely run autonomously by AI systems. The authors say the mathematics does not depend on trusting AI reasoning, and I take that at face value—the burden is still on the code, which is inspectable.

The character-theoretic part of Corollary 2 is argued properly, and the short nonexistence proof for PSL2(59) is a nice byproduct.

Who is this for? Computational group theorists and anyone depending on the completed maximal subgroup list. It is narrow, but it is exactly the missing brick. I would send it to a serious referee, with a request to execute the code and to ask for the explicit direct calculation.

Referee Report

3 major / 6 minor

Summary. Building on the completed classification of the maximal subgroups of the Monster, this paper supplies explicit mmgroup words y, n for the one remaining maximal subgroup, P = 59:29. Theorem 1 states that y and n satisfy the presentation ⟨y, n | y^59, n^29, y^n = y^3⟩ and that, by the uniqueness/maximality established in the earlier papers [3,4], they generate a maximal 59:29. The proof is the Python expression in Figure 1, which checks the two orders and the conjugation relation. Corollary 2 uses additional elements t, w (also defined in Figure 1) to determine all involutions inverting n and, via a stated but not displayed calculation in PSL2(59), concludes that the Monster has no subgroup isomorphic to PSL2(59). Sections 2 and 3 develop and justify a fingerprinting/birthday-paradox search, with reductions M > 2·B > 2^{1+23}.Co2 > 2^{1+23}, and give implementation details.

Significance. The central claim, if accepted, is a significant computational milestone: it completes the explicit generator list for all maximal subgroups of the Monster. The paper's approach is transparent and reproducible: the decisive check is an explicit Python snippet pinned to mmgroup 1.0.8, the probabilistic heuristics are clearly labelled as heuristics, and the group-theoretic reductions in §2.4–2.6 are stated as lemmas with proofs. The main limitations are the dependence on the correctness of mmgroup and on [3,4] for uniqueness/maximality, and the omitted 'direct calculation' in Corollary 2. These are fixable without changing the central approach.

major comments (3)

[Corollary 2, proof] The statement 'A direct calculation shows that PSL2(59) contains an involution s such that n^s = n^{-1} and |ys| = 3' is not proved or displayed. This is the load-bearing step that links the hypothetical H ≅ PSL2(59) to the list I = {t n^j}; without it the contradiction does not follow. Please insert the calculation explicitly, for example in PSL2(59) take n = diag(ζ, ζ^{-1}) with ζ of order 29, s = [[0,1],[-1,0]], y = [[1,1],[0,1]]; then n^s = n^{-1} and ys has order 3. Alternatively, give a precise reference. As written, Corollary 2 is not fully proved.
[Theorem 1 proof / Figure 1] The proof of Theorem 1 is a single mmgroup computation. The paper should state the exact environment and output, and explain why the verified relations imply a subgroup of order 1651. The code checks y^59 = n^29 = 1 and y^n = y^3; since 3 ≠ 1 mod 59, the conjugation action is nontrivial, so the generated subgroup is indeed 59:29. Including the output line and a hash/version pin of the mmgroup package would make the computational proof auditable; as written, a reader must rerun the code to confirm it evaluates True.
[Lemma 5, proof] The step 'As before, this forces c = 1' relies on the facts that C_M(u) = 2·B and N_M(⟨y⟩) = 59:29 have coprime orders. These facts are not stated. Since Lemma 5 underpins the birthday-paradox estimate in Remark 6, please add the explicit gcd argument so that the proof is self-contained.

minor comments (6)

[Corollary 2, proof] Typo: 'the relations 2 = 1 gives w^{2j} = 1' should read 'since s^2 = 1, ...'.
[Throughout] The notation y^n = y^3 for conjugation can be confused with n-th powering. Please add a sentence early in the introduction fixing the convention x^g = g^{-1} x g.
[Figure 1] The code comment says '# true' while the expression evaluates to a Python boolean; please print the literal output 'True' so that the transcript is unambiguous.
[Section 2.5] The assertion that C_K(\tilde z) = 2^{1+23}.Co2 is said to be 'routinely verified with GAP'. A brief indication of the computation (e.g., the relevant fusion map) would be helpful for reproducibility.
[Remark 18] The statement that correctness does not depend on trusting AI-generated reasoning is not verifiable from the text. It would be useful to specify which parts of the pipeline were independently inspected by the authors.
[Corollary 2, proof] The sentence 'the normaliser of ⟨y⟩ in H and in M is N_H(⟨y⟩) = ⟨y,n⟩' conflates the two groups. In H it is the normalizer inside PSL2(59); in M it follows from [15] together with the maximality result from [3,4]. Please separate these assertions.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the generators are explicit and verified by a direct computation; the cited classification results are external inputs, not definitions.

full rationale

Theorem 1's claim has two parts. The existence of a subgroup 59:29 generated by the explicit words y and n is established by the mmgroup check in Figure 1: y.order() == 59, n.order() == 29, and y**n == y**3. This is a direct, rerunnable computation on explicit elements, not an inference from the theorem's conclusion; no parameter is fitted, and no quantity being predicted is used as input. The further claim that this subgroup is maximal and lies in the unique 59:29 class is imported from the authors' earlier classification papers [3,4]. Those are self-citations, but they are published, code-backed classification results used as external premises; the present paper does not redefine maximality in terms of its own construction. Corollary 2 similarly depends on the normalizer statement from Wilson [15], on character table data for the Monster, and on the displayed computation for t and w; the unshown 'direct calculation' in PSL2(59) is a minor exposition gap rather than a circular reduction. The fingerprinting search in Section 2 is heuristic and does not enter the proof itself: the proof is the verification in Figure 1, which is independent of the search narrative. No equation is defined in terms of the claimed result, no fitted input is relabelled as a prediction, and no conclusion is forced by a self-citation chain. The central derivation is self-contained against an external computational oracle (mmgroup), so the appropriate circularity score is 0.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

No new entities or fitted constants are introduced. The construction depends on prior classification results (some by the same authors) and on the correctness of the mmgroup software. The search sizes (e.g., 286,000,000 axes) are heuristic resource choices, not fitted parameters of the result.

assumptions (5)

domain assumption M has a unique conjugacy class of subgroups 59:29, and each such subgroup is maximal.
Used in the proof of Theorem 1; cited from the authors' earlier papers [3,4] and not re-proven here. This carries the 'maximal' part of the theorem.
domain assumption mmgroup version 1.0.8 correctly implements the Monster group and its 196,884-dimensional representation, so the Python expression in Figure 1 returning True is reliable evidence.
The entire proof of Theorem 1 reduces to executing the supplied code; if the software is incorrect, the theorem is unsupported.
domain assumption The character table of M implies |CM(n)| = 87.
Invoked in the proof of Corollary 2 to identify CM(n) as C29×C3; the data are standard but not independently verified in the paper.
domain assumption PSL2(59) contains an involution s with n^s = n^{-1} and |y^s| = 3.
Asserted as 'a direct calculation' in the proof of Corollary 2 without showing the calculation or giving a reference. This is a load-bearing step for the non-existence proof.
standard math Elements y and n of orders 59 and 29 satisfying y^n = y^3 generate a group of order 59·29.
Standard semidirect-product fact used implicitly in Theorem 1 and its proof.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Finding 59:29 in the Monster." pith.science (2026). https://pith.science/paper/EA4UCUI7

@misc{pith2026260728900,
  author       = {Pith},
  title        = {Pith review of: Finding 59:29 in the Monster},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EA4UCUI7}},
  note         = {Machine review of arXiv:2607.28900}
}

read the original abstract

The classification of the maximal subgroups of the Monster group has been completed recently, and explicit generators for each such subgroup (up to conjugacy) have been made available for the software mmgroup, with the exception of the maximal subgroup 59:29, see Dietrich et al. (Adv. Math., 2025; J Algebra, 2026). We provide explicit generators for this last maximal subgroup and comment on the extensive search that led to finding them. Our method is similar to, but significantly more involved than Bray et al.'s (London Math. Soc. J. Comput. Math., 2016) approach for constructing 47:23 in the Baby Monster. Our result allows us to provide a new short proof that the Monster does not have a subgroup PSL2(59), correcting a result of Holmes and Wilson (J. London Math. Soc., 2004).

Figures

Figures reproduced from arXiv: 2607.28900 by the authors.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

15 extracted references · 1 linked inside Pith

[1]

J. N. Bray, R. A. Parker, R. A. Wilson. Finding47:23in the Baby Monster. London Math. Soc. J. Comput. Math., 19:229–234, 2016

2016
[2]

Breuer, CTBLLIB— aGAPpackage, Version 1.3.11, 2025

T. Breuer, CTBLLIB— aGAPpackage, Version 1.3.11, 2025. https://www.math.rwth-aachen.de/homes/Thomas.Breuer/ctbllib/

2025
[3]

Dietrich, M

H. Dietrich, M. Lee, T. Popiel. The maximal subgroups of the Monster. Adv. Math. 469:110214, 2025

2025
[4]

Dietrich, M

H. Dietrich, M. Lee, A. Pisani, T. Popiel. Explicit construction of the maximal subgroups of the Monster. J. Algebra 698:862–895, 2026. Accompanying code:https://github.com/melissa-maths/MonsterSubgroups

2026
[5]

GAP — Groups, algorithms and programming.https://gap-system.org

The GAP Group. GAP — Groups, algorithms and programming.https://gap-system.org
[6]

R. L. Griess, Jr. The Friendly Giant. Invent. Math., 69:1–102, 1982

1982
[7]

H ¨ohn, M

G. H ¨ohn, M. Seysen. The Order of the Monster Finite Simple Group. Preprint,https://arxiv.org/ abs/2508.01037

arXiv
[8]

P. E. Holmes, R. A. Wilson. PSL 2(59) is a subgroup of the Monster. J. London Math. Soc. 69:141–152, 2004

2004

Show all 15 references

[9]

A. A. Ivanov. The Monster Group and Majorana Involutions. Cambridge Tracts in Mathematics, V ol. 176. Cambridge University Press, Cambridge, 2009

2009
[10]

King’s Computational Research, Engineering and Technology Environment (CRE- ATE)

King’s College London. King’s Computational Research, Engineering and Technology Environment (CRE- ATE). 2022.https://doi.org/10.18742/rnvf-m076

2022 doi
[11]

M. Seysen. A fast implementation of the Monster group. J. Comput. Alg., 9:100012, 2024

2024
[12]

M. Seysen. Themmgrouppackage, version 1.0.8.https://github.com/Martin-Seysen/mmgroup
[13]

M. Seysen. Online manual ofmmgroup.https://mmgroup.readthedocs.io/en/latest/api. html#module-mmgroup.axes
[14]

R. A. Wilson. The maximal subgroups of the Baby Monster, I. J. Algebra 211:1–14, 1999

1999
[15]

R. A. Wilson. The odd-local subgroups of the Monster. J. Austral. Math. Soc., 44:1–16, 1988 (Dietrich, Lee, Pisani) SCHOOL OFMATHEMATICS, MONASHUNIVERSITY, CLAYTON, AUSTRALIA (Rizzoli) DEPARTMENT OFMATHEMATICS, KING’SCOLLEGELONDON, UK, HEILBRONNINSTITUTE FOR MATHEMATICALRESEAR...

1988

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.

[1] [1]

J. N. Bray, R. A. Parker, R. A. Wilson. Finding47:23in the Baby Monster. London Math. Soc. J. Comput. Math., 19:229–234, 2016

2016

[2] [2]

Breuer, CTBLLIB— aGAPpackage, Version 1.3.11, 2025

T. Breuer, CTBLLIB— aGAPpackage, Version 1.3.11, 2025. https://www.math.rwth-aachen.de/homes/Thomas.Breuer/ctbllib/

2025

[3] [3]

Dietrich, M

H. Dietrich, M. Lee, T. Popiel. The maximal subgroups of the Monster. Adv. Math. 469:110214, 2025

2025

[4] [4]

Dietrich, M

H. Dietrich, M. Lee, A. Pisani, T. Popiel. Explicit construction of the maximal subgroups of the Monster. J. Algebra 698:862–895, 2026. Accompanying code:https://github.com/melissa-maths/MonsterSubgroups

2026

[5] [5]

GAP — Groups, algorithms and programming.https://gap-system.org

The GAP Group. GAP — Groups, algorithms and programming.https://gap-system.org

[6] [6]

R. L. Griess, Jr. The Friendly Giant. Invent. Math., 69:1–102, 1982

1982

[7] [7]

H ¨ohn, M

G. H ¨ohn, M. Seysen. The Order of the Monster Finite Simple Group. Preprint,https://arxiv.org/ abs/2508.01037

arXiv

[8] [8]

P. E. Holmes, R. A. Wilson. PSL 2(59) is a subgroup of the Monster. J. London Math. Soc. 69:141–152, 2004

2004