gpt-6.2-sol high: cef-moduli
Reviewer record, not the publication verdict. This preserves a completed model response, including unaccepted proposals and its stated source limitations. The curated report and repair note specify which conclusions were adopted. Seven initial source-label failures were reconciled in explicit derivatives; native statuses and original access declarations remain in the provenance record.
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.
All three assigned targets were checked at their proof interfaces. The principal defects are the ordinary Gysin map for a punctured fiber and the missing κ pullback convention; an all-genus surface Torelli FI identification that is incompatible with the manuscript's genus-2 discussion; and, for lower-central Lie algebras, a unital-algebra mismatch plus an unstated map interface for Malcev stability. The new arguments above repair the elementary steps without finding a counterexample to any of the intended eventual-polynomiality theorems. The cited abelianization and Malcev inputs remain conditional where their exact versions are unavailable. The review uses the recorded received-date cutoff of 2013-07-14. Bibliographic years establish chronology only to the precision indicated, not the external contents of the references. Later knowledge cannot be excluded from model training; no later result is used to close a proof obligation. This is a source-only assessment of CEF-v4-TeX, with rendered branches distinguished from comments. The journal metadata is recorded, but the journal full text was not collated.
M1
Checked statement: Theorem 7.1.1: for fixed g ≥ 2, the tautological rings form a graded FI-algebra of finite type, and their grading-j characters and dimensions are eventually polynomial of degree at most j. Also checked the forgetful maps, the κ-class convention, and the interface with the cohomological stability remark.
Returned assessment: repairable_gap
Forgetting and relabeling points compose as required on the moduli functor, so pullback gives the claimed FI structure on rational cohomology. The cotangent line at a retained label pulls back to the cotangent line at that label. The problem is with the stated definition and naturality of κ, not with the FI argument once the convention is fixed. New repair: use the compact universal curve over the moduli orbifold of smooth marked curves, allowing the moving point in the fiber to equal a retained mark. On that curve let K be the first Chern class of the relative cotangent line. Use the proper oriented fiber integral κ̃_j = π_!(K^{j+1}). The compact universal curve with its relative cotangent line is pulled back from the unmarked one: forgetting labels does not change the underlying smooth curve. Fiber integration commutes with that pullback, as can be checked in oriented local trivializations, including the equivariant trivializations on orbifold charts. Thus κ̃_j is compatible with every injection. If instead one uses compact-curve logarithmic κ-classes, with K replaced by K + Σ_s[D_s] for the disjoint marked sections D_s, the same subalgebra results. Indeed, K restricts on D_s to ψ_s, the normal of D_s has first Chern class −ψ_s, and the section self-intersection formula gives π_!(K^{j+1-u}[D_s]^u) = (−1)^{u-1}ψ_s^j for u ≥ 1. The binomial expansion gives κ_j^log = κ̃_j + Σ_s ψ_s^j. This also explains why the logarithmic classes oneself do not define the constant FI generators asserted in the proof. In either convention the degree-zero classes are scalars and the unit suffices. With the pullback-compatible generators, each grading-j piece is spanned by monomials in the κ̃'s and ψ's involving at most j distinct labels. There are finitely many monomials in that grading on sets of size at most j, and every monomial on a larger set is an injection image of one of those. Thus the piece is finitely generated in FI degree ≤ j, not merely of finite type; its weight is ≤ j. Character polynomiality and the specialization at the identity now give the stated eventual bounds. No numerical stable range follows just from this generation argument.
Premises used
- The compact-family repair uses the oriented compact surface family of the moduli orbifold, not the punctured forgetful map of coarse spaces. At the family level the cartesian diagram follows by retaining the same underlying curve and the same arbitrary moving point while forgetting some of the sections.
- \label{pr:fgweight}: over a characteristic-zero field, generation in FI degree ≤ j implies weight ≤ j. \label{thm:persinomial}: for a finitely generated FI-module over such a field the polynomial has weighted degree ≤ weight and is valid for n ≥ stability degree + weight. These general FI results are used at their stated interfaces, not given a whole-proof verdict here.
- J1, Theorem concerning uniform representation stability of H^i of the pointed-curve moduli spaces for fixed g ≥ 2 and i with the forgetful pullbacks, is used only in the final remark, not in Theorem 7.1.1. available_from: 2011 (year precision); date_evidence: \bibitem{J1}, Algebr. Geom. Topol. 11 (2011), 3011–3041; date_basis: journal year; version: journal citation, also arXiv:1106.0947 without a revision specified. Exact map, hypotheses, and range remain unverified from the external text.
- J2 is a pointer to further bounds, not an input to this proof. available_from: exact dated version unresolved; date_evidence: \bibitem{J2}, arXiv:1207.6828 and 'to appear'; date_basis: unversioned preprint citation; version: unresolved. No bound is imported.
Missing inputs
- The original text does not specify a proper compactification or a support convention for (p_{n+1})_!. This verification cannot be supplied by reading the displayed κ formula literally.
- Dated J1 theorem text if the representation-stability remark is to be independently closed; a dated J2 version if its quantitative results are to be checked.
Scope changes
- The compact-family convention and the logarithmic change-of-generators calculation are new repairs, not definitions supplied by the printed proof.
- Theorem 7.1.1 has no active TeX label; the commented label is a locator only.
Located concern
CEF-v4-TeX, \label{subsection:mgn}, lines 2071–2076
The forgetful map is described as a fibration with a punctured genus-g fiber, without distinguishing orbifold and coarse spaces.
Over a coarse point (X;y) of M_{g,1}, the coarse fiber of M_{g,2} is (X−{y})/Aut(X;y), not necessarily X−{y}. For example, take the compact double cover of the sphere branched at six distinct points (genus 2 by Euler characteristic) and mark a ramification point. The stabilizer contains the deck involution, so the fiber is a further finite quotient of a punctured sphere. Its compactification has genus zero: pullback along a finite branched map injects H^1 of the quotient surface into H^1 of the sphere, by the nonzero degree pairing. In contrast, a mark off the finite union of fixed-point sets of the nonidentity automorphisms has trivial stabilizer, and the fiber is a punctured genus-2 surface. The two fibers even have different first Betti numbers.
Make the fibration assertion in the orbifold or family sense. The forgetful maps and FI pullbacks themselves also exist on coarse spaces; they do not need the coarse fibration assertion.
This is an incorrect intermediate description if read in the coarse category. It neither invalidates the co-FI maps nor disproves Theorem 7.1.1.
Located concern
CEF-v4-TeX, lines 2081–2085 and 2107–2112, Gysin display and proof of Theorem 7.1.1
Ordinary integration along the punctured fiber does not give the displayed Gysin map on unrestricted ordinary cohomology. Moreover, invariance of κ under permutations alone does not establish compatibility with forgetting points.
For a connected punctured orientable surface the orientation class with compact support integrates nontrivially but maps to zero in ordinary H^2. Thus an ordinary fiber integral extending the compactly supported one cannot factor through ordinary cohomology as written. In the logarithmic convention, removing a retained section changes the higher κ-class by the corresponding ψ power in the formula above; the constant summands in the proposed FI map should be represented by κ̃, not by logarithmic κ.
Use the proper compact universal curve and the pullback-compatible κ̃-classes as in the new argument above. State the change of generators if the intended convention is logarithmic.
The printed construction of the FI-generating map is incomplete. The degree and eventual-polynomiality conclusions survive for the usual smooth tautological ring with the convention made explicit.
Located concern
CEF-v4-TeX, lines 2110–2111, proof of Theorem 7.1.1
The proof ends with generation in degree ≤ i for a piece indexed by j, and the cited slope conclusion by itself does not imply a generation-degree bound.
Slope controls weight, not generation. Even a module supported only at a high degree m with the trivial S_m-action has weight zero and first requires generators in degree m.
For the character argument say weight ≤ j. If generation in degree ≤ j is also intended, use the direct monomial-support argument above after correcting the generators.
No larger degree bound is needed. The stronger generation statement holds by the new argument, not by the printed inference from slope.
M2
Checked statement: Theorems \label{theorem:albanese1} and \label{theorem:albanese2}: the surface and free-group Torelli constructions, naturality of rational first homology, and eventual degree-≤3i dimension polynomials for Albanese cohomology. Also checked the IA character-polynomial conclusion.
Returned assessment: repairable_gap
The surface proof uses the genus-≥3 rational abelianization formula as if it applied over all of FI, despite explicitly distinguishing genus 2. This does not disprove an eventual statement. New repair: construct the surface FI-group up to conjugacy on the full subcategory in cardinalities ≥3 and extend by the trivial group below 3. For m<n in the tail, use an orientation-preserving interior genus-m subsurface embedding with connected complement and with the prescribed map on symplectic bases. One can start with the standard inclusion of m handles and use the stated surjectivity of the ambient mapping class group onto the symplectic group to realize the prescribed map. The conjugacy interface can be verified directly given that same surjectivity: for two embeddings with the same homology map, their complements are oriented surfaces of the same genus with two boundary circles. Cutting each complement into a handle part and a collar of its boundary gives an orientation-preserving identification with the prescribed boundary maps. Gluing to the identification of the embedded subsurfaces gives an ambient mapping class carrying one embedding to the other and acting as the identity on the common homology image. Its action preserves the symplectic orthogonal complement. Realize the inverse of its action there by a mapping class on a one-boundary genus-(n−m) handle part of the complement, extended by the identity. This uses the stated symplectic surjectivity for genus ≥2; in genus 1 the two twists on curves intersecting once give the two elementary shears, which generate SL_2(Z) by the Euclidean algorithm. The corrected ambient map is in Torelli and still carries one embedding to the other. Equal-size injections are realized by whole-surface mapping classes; equal homology actions differ by Torelli. The composite embedding realizes the composite homology map, so these facts give functoriality up to target conjugacy. Extending by the identity on the complement is a group homomorphism; injectivity of this homomorphism is not needed for any of the FI deductions. On the tail, the first Johnson map can be checked for naturality on based loops: use a path through the complement from the outer boundary to the inner one, and compare φ(x)x^{-1} modulo the third lower central term. For loops in the embedded surface it is the inclusion of the old value; the complementary handle loops have zero value. Contraction by the symplectic form identifies that inclusion and extension by zero with the map on Λ³H, where H is surface first homology. This checks the needed naturality once the stated Johnson identification with the rational abelianization is granted. The repaired H_1 is the truncation of Λ³(M(1)⊕M(1)) below degree 3. Its monomials have at most three labels; in cardinality ≥3 every monomial is an injection image from a 3-element set. It is finitely generated in degree ≤3. For IA, the specified map replaces the basis letters by their images, extends the automorphism by the identity on the unused letters, and has inverse obtained by extending the inverse automorphism. Composition is strict. The formula φ(x)x^{-1} modulo the third term gives the corresponding first Johnson map; in Hom(H,Λ²H) the old values are included and the unused basis vectors go to zero. Thus the basis identification of the dual with M(1) in the tensor formula is covariant extension by zero, not a reversal of FI variance. In Λ²M(1)⊗M(1), terms with two labels have the sign twist of the two-point permutation module, hence trivial plus sign; terms with three labels have the sign twist of the three-point permutation module, hence standard plus sign. This verifies the displayed decomposition and weight 3 over Q. For both groups, each objectwise first cohomology is finite-dimensional in the construction actually used. Its exterior powers surject naturally onto the algebra generated in degree one. Dualizing gives submodules of the corresponding tensor powers of H_1, so finite generation follows from characteristic-zero FI Noetherianity, and the weight is ≤3i by the tensor-weight bound. The resulting polynomial dimension for the surface agrees with the original groups on the tail; for IA the same argument gives the character polynomial. The proof does not specify a numerical range in g or n as a function of i.
Premises used
- FM / the exact sequence stated in the surface paragraph: the mapping class group of a compact oriented one-boundary surface maps onto Sp_{2g}(Z) by its homology action, with kernel Torelli. This is used for g ≥ 2 in the conjugacy verification. available_from: 2012 (year precision); date_evidence: \bibitem{FM}, Princeton Mathematical Series 49; date_basis: cited book publication year; version: cited 2012 book. The exact external passage was not inspected.
- Johnson surface input needed at \label{eq:Johnsonhom}: for a compact oriented genus-g surface with one boundary fixed pointwise and g ≥ 3, the first Johnson homomorphism induces H_1(I_g^1;Q) ≅ Λ³H_1(S_g^1;Q), with the stated symplectic action. The use of the first Johnson homomorphism, rather than just an abstract representation isomorphism, matters. available_from: unresolved on the supplied record; date_evidence: lines 2137–2143 and 2169 attribute the result to Johnson but give no dated abelianization source; date_basis: undated in-text attribution; version: unresolved. The genus-≥3 finite-generation assertion is not needed when Albanese cohomology is taken to mean the algebra generated by H^1.
- JohnsonCR, Theorem 1A as invoked at line 2164: the needed assertion is target-Torelli conjugacy of the extension maps for orientation-preserving embeddings inducing the same homology map. available_from: 1980 (year precision); date_evidence: \bibitem{JohnsonCR}, Math. Ann. 249 (1980), 243–263; date_basis: journal year; version: cited journal article. Its exact scope was not verified; the new check above supplies the needed conjugacy on the tail from the stated symplectic surjectivity.
- Mess is used in the low-genus discussion, not to prove the repaired tail theorem. The needed low-genus input is infinite-dimensional rational first cohomology in genus 2. An infinite-rank abelianization for the closed genus-2 Torelli group would suffice: capping gives a surjection from the one-boundary Torelli group, since a closed mapping class can be represented fixing a disk and capping identifies first homology. available_from: 1992 (year precision); date_evidence: \bibitem{Mess}, Topology 31 (1992), 775–790; date_basis: journal year; version: cited journal article. Exact statement not independently verified.
- Ka / the attributions at \label{eq:H1IAn}: the first Johnson homomorphism induces the integral abelianization H_1(IA_n;Z) ≅ Λ²Zⁿ⊗(Zⁿ)*. Only the rational isomorphism is required here. Ranks 0 and 1 give zero on both sides directly. available_from: 2006 according to the supplied bibliography for the cited revision; date_evidence: \bibitem{Ka}, arXiv:math/0505497v3, listed as a 2006 preprint; date_basis: stated preprint year; version: v3. The text of v3 and the theorem's range were not inspected.
- \label{th:noetherian}, \label{pr:tensor}, \label{pr:tensorwt}, \label{pr:cofisubalgebra}, and \label{thm:persinomial} are used over Q. All dual/tensor comparisons in the argument are objectwise finite-dimensional. No general assertion about commuting tensor products with infinite-dimensional duals is needed.
Missing inputs
- A dated source and exact theorem for the one-boundary surface Johnson abelianization; JohnsonCR is cited for a different step.
- The cited Kawazumi v3 theorem or another dated pre-cutoff account identifying the IA abelianization with the first Johnson homomorphism in the required ranks.
- If retaining the low-genus discussion as an independently verified claim, the exact Mess statement and its passage from the closed to the one-boundary group.
Scope changes
- The surface finite-type assertion is for a modified FI-group agreeing with the original groups from genus 3 onward, not for the original sequence in every genus.
- The new conjugacy and naturality checks do not purport to verify the exact text of JohnsonCR, Mess, or Kawazumi.
- The literature survey about computations of higher Albanese cohomology is not used to establish eventual polynomiality.
Located concern
CEF-v4-TeX, \label{eq:Johnsonhom}, lines 2137–2143, and proof of \label{theorem:albanese1}, lines 2161–2170
The global FI isomorphism and ensuing finite-type assertion in the surface proof do not follow from the genus-≥3 identification and in fact contradict the paper's own genus-2 description if used for the original groups.
At genus 2 the displayed exterior cube of a four-dimensional vector space has dimension 4. The preceding surface paragraph states that H^1 of the genus-2 one-boundary Torelli group is infinite-dimensional. A finitely generated FI-module necessarily has finite-dimensional evaluations over Q, by \label{prop:charfg}. In particular the printed all-genus duality and finite-generation assertions cannot be used as written; the surface groups below the stated genus range were not supplied either.
Use the tail-and-trivial-group extension above, or explicitly work with the degree-≥3 truncation and extend the relevant modules. Define the Albanese algebra as the subalgebra generated by H^1, as the manuscript also describes it, to avoid invoking a definition restricted to finitely generated groups at genus 2.
This is an actual failure of the intermediate all-genus argument, not a counterexample to the eventual polynomial. The repair leaves every genus ≥3 evaluation unchanged.
Located concern
CEF-v4-TeX, proof of \label{theorem:albanese1}, lines 2162–2169; proof of \label{theorem:albanese2}, lines 2188–2197
The imported objectwise identifications must be identifications induced by the Johnson maps, compatible with extension by identity. Equivariance for automorphisms of each individual group would not suffice by itself.
For the IA maps the Hom expression requires the dual input to be extended by zero on unused letters, which comes from the extended automorphism fixing those letters, not from covariantly dualizing an inclusion. For the surface formula the same check requires complementary loops and the symplectic contraction. These are map-level conditions, not consequences of the dimensions or the isomorphism classes.
State the exact imported abelianization results with valid ranges and include the map checks given above. For the surface argument state the embedding and complement hypotheses explicitly.
The FI deductions and degree bounds check out with those inputs. The original external abelianization premises remain conditional in this source-only review.
M3
Checked statement: Theorem \label{theorem:assoc:graded}, the action and functoriality assertions for lower-central-series Lie algebras, and the free-group, pure-braid/Malcev, and two Torelli applications, including the weight bounds supplied by the argument.
Returned assessment: repairable_gap
The bracket on the lower-central quotients is natural, and those quotients are generated under the bracket by degree one. Indeed, modulo Γ_{j+2}, conjugation of elements of Γ_{j+1} is trivial, and the generators [x,y] with x∈Γ and y∈Γ_j give the products of degree one and degree j; induction gives iterated brackets of degree-one elements. This also shows that an automorphism acting trivially on the abelianization acts trivially on all the graded pieces, including for the purposes of conjugacy in the target group of a homomorphism. New repair of the algebra interface: for each i≥1, take the direct sum of one copy of H_1(Γ;k)^{⊗i} for each bracketing of i letters, and send each tensor to its bracket. This is a natural surjection onto gr_i. It proves Theorem \label{theorem:assoc:graded} without treating a Lie bracket as a unital product. It also proves generation in degree ≤ bi if H_1 is generated in degree ≤ b, and, over a characteristic-zero field, weight ≤ ci if H_1 has weight ≤ c. For the free groups, H_1=M(1); the support bound is i, and representation stability is a characteristic-zero deduction. Stability for the degree-i free Lie pieces themselves also follows directly by using the same bracket-word surjection onto the free Lie algebra on M(1), without identifying that algebra with gr(F_n). For pure braids the arXiv expansion supplies the H^1 calculation: the preamble retains the \remove branch, which points to \label{example:arnoldalgebra}; the \antiremove branch is not a rendered assertion. An independent check of the exact FI# input needed here is possible for the plane. Compress the plane radially by h(x)=x/(1+|x|) into the open unit disk when adding labels, and place new points outside the closed disk; forget and relabel as required. In a composite the surviving old points are transformed by h^r for r=0,1,2. If any added labels survive, their locations are constant outside the closed image disk of h^r. For r=2 rescale the entire output by a factor increasing from 1 to 2, then interpolate the two strictly increasing radial functions for 2h² and h, both with image the open unit disk. The constants remain outside that disk, where their configuration can be moved to the chosen direct-addition configuration without collisions. If no constants survive, interpolate the radial embedding to the identity; the case of no surviving old points is just a path of constant configurations. This proves the required plane composition up to homotopy. The pairwise winding homomorphisms form a basis of the dual of H_1: a loop bounds a piecewise linear disk in the ambient vector space; a general-position choice of interior vertices misses intersections of diagonals (real codimension at least four) and meets individual diagonals transversely in finitely many interior points. Deleting small disks gives a homology expression in meridians, and the winding homomorphisms pair with the individual meridians as the identity matrix. Relabeling permutes these classes, with no sign on swapping the two labels. A pair involving a new fixed point contributes zero winding because the moving point lies in the disjoint disk; a forgotten label has no image. Hence H_1 is the FI# module induced from the trivial S_2-representation, as used in the paper. The new bracket-word proof gives generation in degree ≤2i and weight ≤2i for gr_i(P_•;Q); the printed FI# stability criterion gives range n≥4i for these pieces. For the Torelli pair use M2, with the genus-≥3 tail repair for surfaces. Both lower-central pieces have weight ≤3i and are finitely generated, hence have eventual dimension polynomials of degree ≤3i, a bound not explicitly stated in the final two theorems. In degree zero the lower-central grading has zero piece. For Malcev, applying stability to gr_i(P_•) gives the asserted result for the associated graded of the rational Malcev Lie algebra if the comparison is compatible with the stabilization maps; alternatively the maps may be defined by transport along the comparison. It does not give stability of arbitrary filtered subspaces or of independently specified maps on abstractly isomorphic S_n-representations.
Premises used
- \label{pr:tensor} and \label{pr:tensorwt} applied to finite tensor powers and finite sums; the weight assertion needs a characteristic-zero field. \label{th:FIequiv} also requires characteristic zero, and \label{co:fisharprange} needs the actual FI# structure.
- M2's IA construction and the repaired surface tail construction, together with the conditional rational abelianization inputs recorded there. The genus-≥3 surface Lie pieces are unchanged by replacing the smaller groups.
- Ha: the needed Malcev comparison is that the homomorphism from a finitely generated pure braid group to its rational Malcev completion induces, for each i, an isomorphism (Γ_iP_n/Γ_{i+1}P_n)⊗Q → gr_i(𝔭_n) for the lower-central filtration, natural under the braid-group maps, or an equivalent comparison with a stated transport of maps. available_from: 1997 (year precision); date_evidence: \bibitem{Ha}, J. Amer. Math. Soc. 10 (1997), 597–651; date_basis: journal year; version: cited journal article. The precise comparison and filtration are not independently verified.
- Ar: the application invokes the published calculation of pure-braid cohomology, in particular the degree-one pair classes; the full presentation is not required for M3. available_from: 1969 (year precision); date_evidence: active \bibitem{Ar}, Mathematical Notes 5 (1969), 138–140; date_basis: journal year; version: cited journal article. The FI# degree-one input is instead checked directly above.
- Ko is an alternative route, not a premise of the bracket argument. The relevant presentation would identify a graded rational model of 𝔭_n with generators t_{ab}=t_{ba} in degree one and relations [t_{ab},t_{cd}]=0 for four distinct labels and [t_{ab},t_{ac}+t_{bc}]=0 for three distinct labels, with compatible label maps. available_from: 1985 (year precision); date_evidence: \bibitem{Ko}, Invent. Math. 82 (1985), 57–87; date_basis: journal year; version: cited journal article. Neither the precise comparison nor its field and map hypotheses have been checked from that source.
Missing inputs
- The precise Ha statement, including its filtration, rational coefficients and functoriality, or an explicit convention in \label{theorem:malcevpn} that the maps on the reported pieces are transported.
- If retaining the identification of gr(F_n) with the free Lie algebra as an independently checked external result, a dated source for the natural lower-central comparison for a free group. It is uncited in the example and not needed for the direct free-Lie stability argument.
- For the surface and IA applications, the dated first-homology inputs listed under M2.
Scope changes
- The plane homotopy argument checks only what the braid application needs; it is not a referee verdict on the general manifold statement of \label{pr:CMhTop}.
- The additional ≤3i bound for the two Torelli Lie-algebra dimension polynomials concerns the repaired tail in the surface case. The n≥4i range concerns lower-central pure-braid pieces and transfers to Malcev pieces only with the indicated maps.
- No stability assertion is made here for arbitrary positive-characteristic representations or for the entire filtered Malcev completion.
Located concern
CEF-v4-TeX, lines 2204–2221, lower-central series, automorphism, and fundamental-group assertions
The series is not necessarily strictly decreasing. Also, the action descends to the image in Aut(Γ^ab), not in general to an action of the entire abelianization automorphism group by graded Lie automorphisms. The unbased fundamental-group construction needs path-connected objects or additional component data.
An abelian group has Γ_2=Γ_3=1. For the action claim take Γ to be the integral upper-unitriangular 3×3 group times an infinite cyclic factor. Over Q, gr_1 has basis a,b,c, gr_2 has basis [a,b], and c is the only direction in gr_1 that brackets with no degree-one vector. The automorphism of Γ^ab≅Z³ swapping a and c cannot act compatibly on gr: it would not preserve that radical. For the component claim, a disconnected space consisting of a point and a circle gives different fundamental groups depending on the component; conjugacy does not resolve that choice.
Write weak inclusions; say that the action factors through the image of Aut(Γ) in Aut(Γ^ab) and hence descends to Out(Γ). Use the path-connected subcategory for the unbased construction, or explicitly retain component data. Interpret the absence of strict functoriality as the absence of a canonical strict lift in general, not an impossibility for every example.
These corrections leave the group-up-to-conjugacy theorem and the connected plane application intact. The literal whole-automorphism-group claim and the unqualified claim on Top are not valid.
Located concern
CEF-v4-TeX, \label{theorem:assoc:graded}, lines 2224–2232, invoking \label{thm:fisubalgebra} under \label{remark:FIalgebras}
The cited FI-algebra convention requires a unital multiplication, while the lower-central Lie bracket is not such a multiplication.
If an element e were a unit for the bracket, then [e,e]=e, but alternation gives [e,e]=0; a nonzero Lie algebra therefore cannot be unital in this sense. Listing the free Lie algebra among the later examples does not change the earlier definition.
Use the new bracket-word surjection above, or explicitly apply the finite-positive-grading argument to nonunital algebras. Specify characteristic zero where the example goes on to claim representation stability.
The interface in the printed proof is defective; the finite-type theorem is true by the elementary replacement, with the stated generation and weight implications.
Located concern
CEF-v4-TeX, \label{example:pbgraded} through \label{theorem:malcevpn}, lines 2240–2255, with \label{pr:CMhTop}
The cited homotopy composition argument suppresses cases where one of the maps adds no points, and the Malcev application invokes an objectwise equivariant comparison while uniform representation stability requires a sequence with maps.
In \label{pr:CMhTop}, Ψ_B^T is defined to be the identity when T=B, so the composition need not apply Φ twice to surviving points. The constant labels of the direct composite belong to its target outside the composite image, not in general to T−B. Separately, even a constant sequence of trivial one-dimensional representations can be stable or not, depending on whether its transition maps are identities or zero.
Use the explicit plane-only homotopy check above for M3. Specify that the Malcev pieces are associated-graded pieces, and provide naturality of the comparison or specify the maps by transport.
Stability for gr_i(P_•;Q) follows with the plane repair. The printed Malcev theorem is conditional on the identified comparison and map convention; no counterexample to the intended associated-graded statement is shown.
Outside the audit
- The subscription journal text and differences between it and the arXiv author version.
- External reference texts, including exact historical versions for the surface Johnson abelianization and the Malcev comparison; bibliography inspection is not reference verification.
- The proofs of the general FI theorems beyond the uses identified in the dependency list, and the general-manifold version of the homotopy FI# construction.
- The peripheral literature claims about higher Albanese computations, Torelli Lie presentations, pseudo-nilpotent cohomology and comparisons with earlier stability ranges.
- Preserved source-access limitation: J1
- Preserved source-access limitation: J2
- Preserved source-access limitation: FM
- Preserved source-access limitation: JohnsonCR
- Preserved source-access limitation: Mess
- Preserved source-access limitation: Ka
- Preserved source-access limitation: Ha
- Preserved source-access limitation: Ar
- Preserved source-access limitation: Ko
- Preserved source-access limitation: FH
- Preserved source-access limitation: Mac
Additional supporting archive items are identified in the original report. View the manuscript hashes and repository snapshot.