Representation theory and topology

Representation theory and homological stability

Thomas Church and Benson Farb

91 pages66 claims indexed80 referencesManuscript 2013-06-10

gpt-6.2-sol high: cf-braids

Reviewer record, not the publication verdict. The full response includes unaccepted proposals and source limitations. The curated report states what was adopted. Initial packets used a conservative 2010 cutoff; focused packets used the verified 2012 journal received date. The focused Fable answer needed a metadata-only derivative; its original failure and partial source access remain recorded.

Completed visible assessment, preserved as a readable check record. It is a model assessment, not a formal certificate. This document contains no private reasoning stream.

The pure-braid proof separates the structure-map conditions from the multiplicity argument correctly. Its 4i range follows from the active-support bound and the exact part of Hemmer's theorem invoked, but the supplied record does not establish that Hemmer's needed version was available by 2010-08-07. The separate Pieri reduction explains why this is a historical proof obligation, not evidence that the range is false. The type-B argument likewise gives uniform representation stability after correcting the ambient dimension and making the horizontal-strip comparison precise. Both twisted corollaries establish isomorphism of finite-dimensional rational vector spaces in legal ranks, not stability of specified maps. Pointwise multiplicity stability, simultaneous multiplicity stability, and the forward structure-map conditions are distinct here; strong stability adds IV, and reversed-map, simple, mixed tensor, and modular periodic stability are not the conclusions of these targets. The cutoff is the supplied first-arXiv fallback, not a verified journal received date; v3 and its retrospective notes are not evidence of what was available at that cutoff. No external texts or tools were accessed. Later mathematical knowledge cannot be excluded from model training, so historical admissibility has been assessed only from the recorded policy and the dated evidence in the supplied manuscript.

B1

Checked statement: Theorem 4.1: for each fixed i ≥ 0, the rational cohomology of the pure braid groups, with pullbacks along the maps forgetting the last strand, is uniformly representation stable for n ≥ 4i.

Returned assessment: unresolved

The mathematical deduction of the stated range checks out conditional on the imported inputs. Forgetting the last point is equivariant for the standard S_n inclusion. Its pullback fixes each old w_{jk}; the displayed basis maps into the next displayed basis, proving injectivity. Alternatively, appending a positive real point of modulus 1 + max_j |z_j| gives a continuous section. Degree-i monomials involve at most 2i indices, so for n ≥ 2i each is an S_{n+1}-translate of a monomial in the image. This is orbit-span surjectivity, not surjectivity of the linear map. For III′, the Orlik–Solomon splitting groups the degree-i summands by set partitions with n−i blocks. Write the nonsingleton block sizes as s_1,…,s_b. Then i = Σ(s_j−1), and their total size k = i+b is at most 2i. Once n ≥ 2i the list of nonsingleton block types is fixed. Each orbit gives induction from H × S_{n−k}, where H is the stabilizer on the k active coordinates. Appending singleton blocks leaves the active representation unchanged, and permutations of singleton blocks act trivially. The finite sum therefore has a common multiplicity threshold n ≥ 2 max k ≤ 4i. Degree zero is the constant trivial sequence and does not require Hemmer's k ≥ 1 theorem. There is no need to identify the active representations by characters, or to prove that any individual irreducible travels to the same type. Indeed the degree-one invariant sum printed in the source is not invariant after pullback, so IV fails in degree one. The unresolved point for the original proof is historical availability of the precise Hemmer result, not a counterexample to the range.

Premises used

  • Arnol’d, \bibitem[Ar]{Ar}: rational cohomology of the ordered configuration space is generated by the logarithmic degree-one classes with the displayed three-term relations; the stated basis is used in the printed injectivity proof. available_from: 1969 at publication; date_evidence: CF bibliography, Mathematical Notes 5, 138–140; date_basis: publication year; version: cited published article, not inspected.
  • Orlik–Solomon, \bibitem[OS]{OS}, Proposition 2.10 as cited: the rational arrangement cohomology splits by flats, equivariantly under the arrangement symmetries; for the braid arrangement the degree-i pieces are the stated top-cohomology summands for the blocks. available_from: 1980 at publication; date_evidence: CF bibliography, Invent. Math. 56, 167–189; date_basis: publication year; version: cited published article, not inspected.
  • Lehrer–Solomon, \bibitem[LS]{LS}: the braid-arrangement description of the flat summands and their permutation action, cited for details. The finer one-dimensional-centralizer-character description is not used. available_from: 1986 at publication; date_evidence: CF bibliography, J. Algebra 104, 410–424; date_basis: publication year; version: cited published article, not inspected.
  • Hemmer, \bibitem[He]{He}, as restated in Theorem 4.2: for fixed H ≤ S_k and a fixed rational H-module, induction with a trivial S_{n−k} factor has simultaneous padded multiplicities constant at n ≥ 2k. Only the multiplicity conclusion for the finite-dimensional active modules is used here. available_from: 2011 for the identified publication; pre-cutoff availability unknown; date_evidence: CF bibliography, JCTA 118, 1136–1139, with no dated preprint given; date_basis: publication year; version: cited journal article, not inspected.

Missing inputs

  • A dated pre-2010-08-07 version, or equivalent dated evidence identifying a version, of Hemmer's result that contains the uniform n ≥ 2k bound.
  • The cited arrangement-cohomology texts themselves. Their dates are supported by the bibliography, but the content of the exact cited propositions has not been independently verified.

Scope changes

  • The theorem proves forward uniform representation stability, not strong stability or classical homological stability of P_n.
  • A sufficient range is checked, not its optimality. The n ≥ 2i range in the generation argument alone does not establish multiplicity stability.

Located concern

CF-v3-TeX lines 2140–2151, use of \ref{thm:hemmer} in the final range calculation

The proof's decisive imported range is not documented as available by the recorded cutoff.

The supplied bibliography identifies a 2011 publication and gives no earlier dated version. The statements that the authors conjectured the result to Hemmer and that he proved it do not date the proof.

Supply a dated version containing the required range, or replace this invocation by the explicit Pieri argument in B2, labelled as a new argument relative to the supplied proof.

The original proof's historical dependency is unresolved. The formula n ≥ 4i follows from the stated induction input, and no mathematical failure of that range was found.

Located concern

CF-v3-TeX lines 2073–2081 and 2113–2127, orbit representatives and the rank-indexed sums

A few indices obscure the otherwise valid identification of summands.

The first direct sum at rank n+1 indexes ν but constrains ℓ(μ). The subsequent running-rank expression uses the old subscript on the cohomology piece. The representative written with a final block indexed using n−1 should run over the nonzero parts of μ.

Use ℓ(ν) = n+1−i in the first sum and index the running-rank piece by the enlarged set partition, or write it throughout as the fixed active module tensored with the trivial singleton action.

Notational corrections only. The append-singleton construction and the bound k ≤ 2i resolve the intended proof.

B2

Checked statement: Theorem 4.2, Remark 4.3, and Corollary 4.4: induced symmetric-group representations, which cohomological information is used, and what the twisted braid homology argument proves.

Returned assessment: repairable_gap

For Theorem 4.2 the natural map sends g⊗v to the same tensor at the next rank. Since (H × S_{n+1−k}) ∩ S_n = H × S_{n−k}, distinct old cosets give distinct new cosets, proving injectivity. The image contains the inducing copy, so its S_{n+1}-span is the entire induced module. The manuscript delegates III′ to Hemmer. New repair relative to the supplied proof: use transitivity to replace V by Ind_H^{S_k} V, then decompose into actual partitions α of k. The one-row induction rule gives one copy of each partition ν with ν/α a horizontal strip of size n−k. Set τ = (ν_2,ν_3,…). The strip condition is exactly α_r ≥ τ_r ≥ α_{r+1} for all r and ν_1 = n−|τ| ≥ α_1. The first inequalities give |τ| ≤ k, so the last inequality is automatic for n ≥ k+α_1. The allowed tails and hence the padded multiplicities are constant at n ≥ 2k. This argument uses only the one-row case of the induction rule, with the identical argument applying to arbitrary multiplicities since S_k has finitely many irreducible types. It does not date Hemmer's proof. For Corollary 4.4 let M = V(λ)n at a rank where it exists. It is trivial on P_n. The finite quotient S_n has exact coinvariants over Q by averaging, and averaging identifies coinvariants with invariants. For example, restrict a free B_n-resolution to P_n, form the P_n chain coinvariants with coefficients in M, and then take S_n coinvariants; the latter recover the B_n chains. Consequently H_i(B_n;M) identifies with (H_i(P_n;Q)⊗M)^{S_n}. This invariant space is Hom{S_n}(M*, H_i(P_n;Q)); the finite-dimensional rational Specht irreducibles are absolutely irreducible and self-dual. Thus its dimension is the multiplicity from Theorem 4.1, using rational duality between homology and cohomology. Only III′ of Theorem 4.1 is needed, not either structure-map condition. This proves equal finite dimensions, hence abstract Q-linear isomorphisms at the stated bound on the common domain of the padded labels; it does not show that stabilization maps in twisted homology induce isomorphisms.

Premises used

  • Hemmer, \bibitem[He]{He}: III′ for fixed H ≤ S_k and rational V with trivial action on the remaining coordinates, at n ≥ 2k. available_from: 2011 for the identified publication; earlier availability unknown; date_evidence: CF bibliography, JCTA 118, 1136–1139; date_basis: publication year; version: cited journal article, not inspected. The manuscript itself checks I and II.
  • For the new reduction: rational one-row Littlewood–Richardson induction for S_k × S_{n−k} ≤ S_n, giving a multiplicity-one sum over horizontal strips. This is the symmetric-group case of the rule displayed in the type-B proof when both minus partitions are empty. available_from: 2000 for that cited rule; date_evidence: CF citation to Geck–Pfeiffer, \bibitem[GP]{GP}, Lemma 6.1.3 and bibliographic year; date_basis: publication year; version: cited monograph, not inspected.
  • Rational Specht classification, including split irreducibility and finite-dimensional self-duality, as reviewed in \ref{section:reps1}, with Fulton–Harris \bibitem[FH]{FH} as reference. available_from: 1991 for the cited edition; date_evidence: CF bibliography; date_basis: publication year; version: cited GTM 129 edition, not inspected. The averaging and chain-coinvariant steps above are checked directly rather than treated as an unverified transfer citation.
  • Lehrer–Solomon, \bibitem[LS]{LS}: the finer centralizer-character description mentioned in Remark 4.3, not a premise of Corollary 4.4. available_from: 1986 at publication; date_evidence: CF bibliography, J. Algebra 104, 410–424; date_basis: publication year; version: cited published article, not inspected.

Missing inputs

  • A dated pre-cutoff text for Hemmer's multiplicity theorem, and the cited versions themselves for a direct comparison of hypotheses.
  • If classical stability is intended in the Section 1 sense, specify compatible coefficient maps over B_n → B_{n+1} and prove that their induced homology maps are isomorphisms. The manuscript expressly does not make this step.

Scope changes

  • Under the shared counter, 4.3 is a remark and 4.4 is the twisted-homology corollary. The following ordinary-configuration-space remark is 4.5.
  • The induction input is fixed active data with trivial action on the unused letters. Neither the explicit Lehrer–Solomon character formula nor type-preservation is needed for the corollary.

Located concern

CF-v3-TeX lines 2153–2162, unlabelled Remark 4.3

The remark is not a corollary, and the scalar field for the finer centralizer-character description should be made explicit.

The proof uses rational flat summands, not one-dimensional rational inducing characters. For example, in degree two at n = 3 the exterior square of the three-dimensional edge permutation representation is standard plus sign, and the Arnold relation is the sign line. The quotient is standard. The 3-cycle centralizer C_3 has only the trivial rational one-dimensional character, whose induction gives trivial plus sign, not standard.

Distinguish the complex one-dimensional-character description from the rational flat decomposition. The latter suffices in Theorem 4.1.

No effect on either stability deduction. It prevents reading an unnecessary complex-character input as a rational one.

Located concern

CF-v3-TeX lines 2174–2186, \ref{corollary:twistedstability}

The term 'classical homological stability' is stronger than what the proof establishes under the paper's own definition. The range also presupposes existence of V(λ)_n.

The first sentence of the proof explicitly replaces a map statement by eventual equality of dimensions. If n < |λ|+λ_1, the padded coefficient module is not defined even if n ≥ 4i. The zero multiplicity convention for an undefined label in III′ does not define a coefficient module for twisted homology.

State eventual equality of dimensions and abstract isomorphism of rational vector spaces for n ≥ max(4i, |λ|+λ_1), with the empty partition interpreted in the usual way. Treat an isomorphism induced by specified maps as a separate assertion.

The displayed abstract isomorphism survives on its proper domain. No map-based twisted homological stability conclusion has been proved here.

B3

Checked statement: Theorem 4.6, Corollary 4.7, and Conjecture 4.8: type-B pure and twisted braid results over Q, their structure maps and ranges, and the historical status of the string-motion statement.

Returned assessment: repairable_gap

The type-B proof uses the rank-n complement with n coordinates and the maps forgetting the last coordinate. This interpretation is fixed by the later hyperplanes, codimension calculation, signed-permutation action and square-coordinate quotient, notwithstanding the opening ambient-dimension typo. For the rank-n complement the projection has a section appending a positive real coordinate of modulus 1 + max_j |z_j|. This proves injectivity directly; naturality alone need not imply injectivity for an arbitrary presentation. Local triviality of forgetting follows by moving the finitely many punctures 0, ±z_1,…,±z_n in disjoint small disks; the fibers are punctured planes, giving asphericity inductively. Signed permutations fixing the last coordinate give consistency. The rational logarithmic generators involve at most two indices, so orbit-span surjectivity holds for n ≥ 2i. A flat of codimension i has a zero block of size ℓ and nonsingleton signed blocks of sizes s_j; sign changes put the latter in all-positive form. Thus i = ℓ+Σ(s_j−1) and its active size m = ℓ+Σs_j ≤ 2i. The flat can be represented on the first m coordinates. A hyperplane containing it cannot involve an unrestricted coordinate, and the section argument ensures that its flat summand does not acquire new relations by pullback. Therefore H^J comes from the fixed rank-m summand, the W_{n−m} factor acts trivially, and the orbit is induced from G_J × W_{n−m}. With induction factored through W_m × W_{n−m}, the displayed wreath-product rule adds a horizontal strip only to the plus partition α, leaving the minus partition β fixed. New repair of the last combinatorial step: writing the new plus partition as (n−|β|−|τ|,τ), the allowed tails satisfy α_r ≥ τ_r ≥ α_{r+1} and n−|β|−|τ| ≥ α_1. Since |τ| ≤ |α| and |α|+|β| = m, the final inequality is automatic at n ≥ m+α_1. This gives a simultaneous sufficient bound n ≥ 4i (for positive ranks; degree zero is trivial) for Theorem 4.6, without claiming sharpness. For Corollary 4.7 the free W_n cover gives quotient group W_n; the square-coordinate quotient is the unordered configuration space of n points in C minus 0. A configuration with one distinguished point is identified with C times this space by translating that point to zero, explaining B_{1,n}. The same rational averaging argument as in B2 applies. The rational signed-wreath irreducibles have split rational models from the displayed induction of Specht modules; they are self-dual, also seen from an invariant positive definite rational form obtained by averaging. Thus the corollary gives stable dimensions in legal ranks, with no proof about induced twisted stabilization maps. Conjecture 4.8 is a conjecture in the historical argument, not a theorem proved here. In the McCool description, forgetting the last generator is given by quotienting by its normal closure and is split by extension by the identity. With the stated indexed cohomology generators and their usual forgetful pullbacks, this checks consistency and injectivity, and the same support count checks orbit-span surjectivity for n ≥ 2i. These observations alone do not supply III′ in the manuscript's conjecture paragraph.

Premises used

  • Brieskorn, \bibitem[Bri]{Bri}, and Orlik–Solomon, \bibitem[OS]{OS}, Theorem 5.2 as cited: logarithmic generators and the arrangement cohomology presentation for the complex central type-B arrangement over Q. available_from: 1973 and 1980 respectively for the cited publications; date_evidence: CF bibliography, Bourbaki Exp. 401 in LNM 317 and Invent. Math. 56; date_basis: publication years; version: cited published texts, not inspected.
  • Orlik–Solomon Proposition 2.10 as cited: splitting by intersection flats, compatible with permutations of the arrangement. Lehrer–Solomon Equation 2.2 as cited: dependent monomials vanish, so each flat contributes only in its codimension. available_from: 1980 and 1986 respectively; date_evidence: CF bibliography \bibitem[OS]{OS} and \bibitem[LS]{LS}; date_basis: publication years; version: cited published articles, not inspected. No complete Douglass character calculation is required.
  • Geck–Pfeiffer, \bibitem[GP]{GP}, Sections 1.4 and 5.5 and Lemma 6.1.3 as cited: signed-wreath irreducible labelling and product induction for W_m × W_{n−m}, only the case of a trivial second factor being used. available_from: 2000; date_evidence: CF bibliography, cited monograph; date_basis: publication year; version: cited Oxford edition, not inspected.
  • Jensen–McCammond–Meier, \bibitem[JMM]{JMM}: the cohomology presentation for the pure symmetric automorphism group, used over Q here. CF gives the displayed relations but no theorem number. available_from: 2006 at publication; date_evidence: CF bibliography, Geom. Topol. 10, 759–784; date_basis: publication year; version: cited published article, not inspected. Compatibility of the indexed generator description with forgetting is needed for the map check. No imported proof of the conjecture is used.

Missing inputs

  • The cited published texts for exact verification of the arrangement and wreath-product input and of the string-motion presentation and generator conventions.
  • For a map-based interpretation of Corollary 4.7, compatible stabilization maps with the varying coefficients and a proof that they induce isomorphisms.
  • A precise space model and a source for its identification with the McCool group if the string-motion statement is read literally as a claim about the stated space of embeddings, rather than the algebraic group.

Scope changes

  • Theorem 4.6 concerns forward uniform representation stability over Q, not modular stability or type-preservation. Its printed statement gives no numerical range; n ≥ 4i is an extraction from the repaired proof.
  • Conjecture 4.8 is not an original theorem. Retrospective only: its footnote credits Wilson, \bibitem[Wi1]{Wi1}. available_from: 2012 for the cited publication; date_evidence: CF bibliography, Algebr. Geom. Topol. 12, 909–931, also listing arXiv:1108.1255; date_basis: publication year, with the listed preprint identifier also post-cutoff; version: cited published article, with preprint version unspecified and neither inspected.

Located concern

CF-v3-TeX lines 2215–2229 and 2241–2244, definition of X′_n

The printed ambient dimension is 2n, but the arrangement, action, and quotient used in the proof are those for n coordinates.

The standard signed permutation action is on the n coordinates z_1,…,z_n, and the free quotient in the source keeps exactly their n squares. The proof later explicitly takes codimension in C^n. The diagonal equations also require distinct indices when defining the pair hyperplanes.

Write X′_n ⊂ C^n with z_i ≠ 0 and z_i ≠ ±z_j for i ≠ j.

The proof and theorem concern the intended rank-n complement. The rank-2n display does not literally specify the claimed arrangement quotient.

Located concern

CF-v3-TeX lines 2291–2297 and 2351–2363, stabilizer example and stabilization of the plus partition

The block stabilizer notation is ambiguous, and the assertion about adding multiple boxes to the first row is not justified by the printed bound.

On the two nonzero 3-blocks the stabilizer is ((S_3 × C_2)^2) ⋊ S_2: each block permits an internal permutation and a common sign, and the two blocks may be exchanged. For the strip statement take α=(s), s≥1, and add s+1 boxes to obtain (s+1,s); only one box is added in the first row. Also ν[j+1] literally pads ν itself, whereas the stability operation keeps its tail fixed.

Use the explicit block stabilizer or just the needed product G_J × W_{n−m}, and replace the first-row argument by the interlacing inequalities above.

The required stabilizer factorization is correct. The corrected strip argument proves uniform multiplicity stability and yields the sufficient range given above.

Located concern

CF-v3-TeX lines 2366–2373, \ref{cor:gen}

As in Corollary 4.4, the transfer argument proves abstract isomorphism, not the map-based assertion called 'classical homological stability'.

The proof is imported from Corollary 4.4 and uses only stable multiplicities. The double-partition coefficient module is defined only for n ≥ |λ^+|+|λ^-|+λ^+_1, with λ^+_1=0 for the empty plus partition.

State stable dimensions and abstract Q-linear isomorphism in legal ranks. For example, the sufficient bound from the repaired argument is the maximum of 4i and the label's starting rank.

The vector-space conclusion is supported; the stronger terminology should be removed unless an additional map argument is supplied.

Outside the audit

  • Journal typesetting and received/oblatum metadata, original v1 wording, external article contents, and any earlier Hemmer manuscript or correspondence.
  • The exact numerical decompositions in the degree-two examples, the broader assertion of failure of strong stability in every positive degree beyond the degree-one check, and optimal stability thresholds.
  • The proof and range of Wilson's retrospective theorem, and topological model-identification details for the space of string motions.
  • Results and conjectures outside the assigned braid and string-motion targets.
Source record

Additional supporting archive items are identified in the original report. View the manuscript hashes and repository snapshot.

Cite this review

If this review or its supporting reports inform your work, please cite it.

Igor Rivin. Review of Representation theory and homological stability. Math Referee, .

Download .bib
View BibTeX
@misc{rivin2026review-human-church-farb-homological-stability,
  author = {Rivin, Igor},
  title = {{Review of Representation theory and homological stability}},
  year = {2026},
  month = {November},
  date = {2026-11-21},
  howpublished = {Math Referee},
  url = {https://mathreferee.dimensionreducers.ai/papers/human-church-farb-homological-stability/panel/},
  note = {Review report}
}