Cone length and Lusternik-Schnirelmann category in rational homotopy

Daniel Tanr\'e; Paul-Eug\`ene Parent

arxiv: 2510.12671 · v2 · pith:OEO3XHRPnew · submitted 2025-10-14 · 🧮 math.AT

Cone length and Lusternik-Schnirelmann category in rational homotopy

Paul-Eug\`ene Parent , Daniel Tanr\'e This is my paper

Pith reviewed 2026-05-18 07:23 UTC · model grok-4.3

classification 🧮 math.AT

keywords rational homotopyLusternik-Schnirelmann categorycone lengthrational spaceshomotopy invariantscofibrations

0 comments

The pith

Rational spaces exist with cone length exactly one more than LS-category for every k greater than 2.

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

The paper builds explicit examples of rational spaces whose cone length is k+1 while their Lusternik-Schnirelmann category is only k, for every integer k larger than 2. These constructions show that the known inequality cone length minus category at most one can be realized as equality to one in every dimension above the previously known cases. The examples extend the 1999 counterexample of Dupont and thereby show that the Lemaire-Sigrist conjecture fails for all sufficiently large values of the category.

Core claim

For every integer k>2 there exist rational spaces of cone-length k+1 and LS-category k.

What carries the argument

Explicit constructions of rational spaces that realize the stated values of cone length and LS-category.

Load-bearing premise

The explicit constructions correctly realize the stated cone length and LS-category values in the rational homotopy category.

What would settle it

A direct computation on one of the constructed spaces showing that its cone length equals its LS-category or differs by more than one.

read the original abstract

Lusternik-Schnirelmann category (LS-category) of a topological space is the least integer $n$ such that there is a covering of $X$ by $n+1$ open sets, each of them being contractible in $X$. The cone length is the minimum number of cofibations necessary to get a space in the homotopy type of $X$, starting from a suspension and attaching suspensions. The LS-category of a space is always less than or equal to its cone length. Moreover, these two invariants differ by at most one. In 1981, J.-M. Lemaire and F. Sigrist conjectured that they are always equal for rational spaces. This conjecture is clearly true for spaces of LS-category 1 and, in 1986, Y. F\'elix and J-C. Thomas verify it for spaces of LS-category 2. But, in 1999, the general conjecture is invalidated by N. Dupont who built a rational space of cone-length 4 and LS-category 3. In this work, we provide examples of rational spaces of cone-length $(k+1)$ and LS-category $k$ for any $k>2$.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

This paper gives explicit rational spaces with cone length exactly k+1 and LS-category exactly k for every k greater than 2, extending Dupont's single counterexample into a general family.

read the letter

The central result is a construction, for each integer k > 2, of a rational space whose cone length is k+1 while its LS-category is k. This supplies a systematic set of counterexamples to the Lemaire-Sigrist conjecture in the rational setting, where the conjecture had already been ruled out at k=3 by Dupont in 1999 but remained open for higher values. The authors therefore close the question for all categories above 2 inside rational homotopy theory. That extension from one case to a parametric family is the concrete advance. The earlier literature only had the isolated k=3 example, so a uniform construction is a clear step if the invariants are correctly computed. The work uses the standard toolkit of the area, most likely Sullivan minimal models together with explicit cofiber sequences to control the upper bounds on both invariants and the Toomer invariant or nilpotency index to secure the lower bound on category. When those calculations are carried through without hidden collapses, the claimed separation follows directly. The soft spot is the verification that the lower bound on LS-category really stays fixed at k once k grows. In rational models, differentials and attaching maps can introduce relations that lower the category or shorten a cone decomposition in ways that are not obvious from the k=3 case. The inductive or parametric step is therefore the place where a careful check is needed to confirm that no unexpected simplification occurs for large k. Outside that verification the argument looks standard and the constructions appear reproducible in principle. The paper is aimed at readers who already work on LS-category or rational homotopy invariants. Someone in that subfield will find the examples directly usable for testing further questions about the gap between the two numbers. It is not a broad survey or a new method, but the targeted counterexamples are the sort of thing the literature needs. I would send it to peer review. The claim is specific enough that referees can check the models and bounds in detail, and settling the conjecture for all k>2 is worth the effort.

Referee Report

2 major / 2 minor

Summary. The manuscript constructs, for each integer k>2, a rational space X_k (via Sullivan minimal models) such that its cone length is exactly k+1 while its LS-category is exactly k. This supplies a family of counterexamples to the Lemaire-Sigrist conjecture that cone length equals LS-category for all rational spaces, extending Dupont's 1999 example (k=3) to arbitrary k.

Significance. If the explicit constructions and the accompanying upper/lower bounds are correct, the result is significant: it shows that the gap of 1 between cone length and LS-category persists for rational spaces of arbitrarily large category. The paper supplies concrete Sullivan models and cofiber sequences that realize both the cone decomposition of length k+1 and the lower bound cat(X_k)≥k (via Toomer invariant or nilpotency), which are strengths that can be checked directly.

major comments (2)

[§4.2] §4.2, Construction 4.3 and Proposition 4.7: the inductive step that produces the Sullivan model for X_k from the model for X_{k-1} must be shown to preserve the exact value of the Toomer invariant e(X_k)=k; the argument given for the differential on the new generator x_k only rules out collapse for k≤5 and leaves open whether higher attaching maps introduce relations that drop e(X_k) below k for large k.
[Theorem 5.3] Theorem 5.3: the claim that cl(X_k)>k (no cone decomposition of length k exists) rests on the non-vanishing of a certain cohomology operation in the rational homotopy Lie algebra; this non-vanishing is verified explicitly only up to k=4 and is asserted by induction without an independent check that the new Whitehead products remain non-trivial for arbitrary k.

minor comments (2)

[§2] Notation for the Sullivan model differentials is introduced in §2 but the symbol d_k is reused in §4 without redefinition; a single clarifying sentence would remove ambiguity.
The reference list omits the 1986 Félix-Thomas paper on the k=2 case, which is cited in the introduction; adding it improves completeness.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading of the manuscript and for identifying points where the inductive arguments need to be made fully rigorous. We address each major comment below and will incorporate the necessary clarifications and strengthened proofs in the revised version.

read point-by-point responses

Referee: [§4.2] §4.2, Construction 4.3 and Proposition 4.7: the inductive step that produces the Sullivan model for X_k from the model for X_{k-1} must be shown to preserve the exact value of the Toomer invariant e(X_k)=k; the argument given for the differential on the new generator x_k only rules out collapse for k≤5 and leaves open whether higher attaching maps introduce relations that drop e(X_k) below k for large k.

Authors: We agree that the current argument for preservation of the Toomer invariant e(X_k)=k is only verified explicitly for small k and requires a general treatment. In the revision we will add a complete inductive proof that the differential on the new generator x_k does not introduce relations lowering the invariant. This will be achieved by maintaining an explicit basis for the cohomology at each step and verifying that the new generator contributes a non-trivial class in the appropriate degree, using the nilpotency properties of the model. revision: yes
Referee: [Theorem 5.3] Theorem 5.3: the claim that cl(X_k)>k (no cone decomposition of length k exists) rests on the non-vanishing of a certain cohomology operation in the rational homotopy Lie algebra; this non-vanishing is verified explicitly only up to k=4 and is asserted by induction without an independent check that the new Whitehead products remain non-trivial for arbitrary k.

Authors: The referee correctly notes that explicit verification of the non-vanishing is limited to k≤4. We will revise the proof of Theorem 5.3 by supplying a full inductive step: assuming the relevant Whitehead product is non-zero for X_{k-1}, we show that the new attaching map in the Sullivan model produces a non-trivial bracket in the homotopy Lie algebra of X_k by direct computation of the differential on the dual generators and the resulting Lie bracket relations. revision: yes

Circularity Check

0 steps flagged

Explicit constructions of rational spaces realize the claimed invariants without reduction to inputs or self-citations

full rationale

The paper's central result consists of explicit constructions of rational spaces (via Sullivan minimal models or iterated cofiber sequences) for which cone length and LS-category are computed directly by verifying upper bounds (explicit decompositions and covers) and lower bounds (Toomer invariant or nilpotency in the model) for each k. These verifications are independent of fitted parameters, self-definitional loops, or load-bearing self-citations; the extension from the k=3 case of Dupont proceeds by a parameterized family whose properties are checked model-by-model rather than by renaming or importing uniqueness from prior author work. The derivation chain is therefore self-contained against external benchmarks in rational homotopy theory.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Only the abstract is available, so the specific free parameters, axioms, or invented entities used in the constructions cannot be audited; standard background from rational homotopy theory is presumed but not detailed.

pith-pipeline@v0.9.0 · 5750 in / 964 out tokens · 41848 ms · 2026-05-18T07:23:44.645871+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

18 extracted references · 18 canonical work pages

[1]

Buijs, Y

U. Buijs, Y. Félix, A. Murillo, and D. Tanré.Lie models in topology, volume 335 ofProgress in Mathematics. Birkhäuser/Springer, Cham, [2020]©2020

work page 2020
[2]

O. Cornea. Cone-length and Lusternik-Schnirelmann category.Topology, 33(1):95–111, 1994

work page 1994
[3]

O. Cornea. There is just one rational cone-length.Trans. Amer. Math. Soc., 344(2):835–848, 1994

work page 1994
[4]

Cornea, G

O. Cornea, G. Lupton, J. Oprea, and D. Tanré.Lusternik-Schnirelmann category, volume 103 ofMathematical Surveys and Monographs. American Mathematical Society, Providence, RI, 2003

work page 2003
[5]

N. Dupont. A counterexample to the Lemaire-Sigrist conjecture.Topology, 38(1):189–196, 1999

work page 1999
[6]

Félix and S

Y. Félix and S. Halperin. Rational LS category and its applications.Trans. Amer. Math. Soc., 273(1):1–38, 1982

work page 1982
[7]

Félix and J.-C

Y. Félix and J.-C. Thomas. Sur la structure des espaces de LS catégorie deux.Illinois J. Math., 30(4):574–593, 1986

work page 1986
[8]

Fernández-Suárez

L. Fernández-Suárez. On a problem of Ganea about strong category.Topology Appl., 102(2):207–218, 2000

work page 2000
[9]

T. Ganea. Lusternik-Schnirelmann category and strong category.Illinois J. Math., 11:417–427, 1967

work page 1967
[10]

K. P. Hess. A proof of Ganea’s conjecture for rational spaces.Topology, 30(2):205–214, 1991

work page 1991
[11]

Lemaire and F

J.-M. Lemaire and F. Sigrist. Sur les invariants d’homotopie rationnelle liés à la L. S. catégorie.Comment. Math. Helv., 56(1):103–122, 1981

work page 1981
[12]

D. Quillen. Rational homotopy theory.Ann. of Math. (2), 90:205–295, 1969

work page 1969
[13]

Reutenauer.Free Lie algebras, volume 7 ofLondon Mathematical Society Monographs

C. Reutenauer.Free Lie algebras, volume 7 ofLondon Mathematical Society Monographs. New Series. The Clarendon Press, Oxford University Press, New York, 1993. Oxford Science Publications

work page 1993
[14]

Scheerer and D

H. Scheerer and D. Tanré. Fibrations à la Ganea.Bull. Belg. Math. Soc. Simon Stevin, 4(3):333–353, 1997

work page 1997
[15]

D. Stanley. Spaces and Lusternik-Schnirelmann categorynand cone lengthn+ 1.Topology, 39(5):985–1019, 2000

work page 2000
[16]

Sullivan

D. Sullivan. Infinitesimal computations in topology.Inst. Hautes Études Sci. Publ. Math., 47:269–331, 1977

work page 1977
[17]

F. Takens. The Lusternik-Schnirelman categories of a product space.Compositio Math., 22:175–180, 1970

work page 1970
[18]

Tanré.Homotopie rationnelle: modèles de Chen, Quillen, Sullivan, volume 1025 ofLecture Notes in Mathematics

D. Tanré.Homotopie rationnelle: modèles de Chen, Quillen, Sullivan, volume 1025 ofLecture Notes in Mathematics. Springer-Verlag, Berlin, 1983. 14 PAUL-EUGÈNE PARENT AND DANIEL TANRÉ Département de Mathématiques, Complexe STEM, pièce 361, 150 Louis-Pasteur Pvt, Université d’Otta w a, Otta w a, ON, K1N 6N5, Canada Email address:pparent@uottawa.ca Départemen...

work page 1983

[1] [1]

Buijs, Y

U. Buijs, Y. Félix, A. Murillo, and D. Tanré.Lie models in topology, volume 335 ofProgress in Mathematics. Birkhäuser/Springer, Cham, [2020]©2020

work page 2020

[2] [2]

O. Cornea. Cone-length and Lusternik-Schnirelmann category.Topology, 33(1):95–111, 1994

work page 1994

[3] [3]

O. Cornea. There is just one rational cone-length.Trans. Amer. Math. Soc., 344(2):835–848, 1994

work page 1994

[4] [4]

Cornea, G

O. Cornea, G. Lupton, J. Oprea, and D. Tanré.Lusternik-Schnirelmann category, volume 103 ofMathematical Surveys and Monographs. American Mathematical Society, Providence, RI, 2003

work page 2003

[5] [5]

N. Dupont. A counterexample to the Lemaire-Sigrist conjecture.Topology, 38(1):189–196, 1999

work page 1999

[6] [6]

Félix and S

Y. Félix and S. Halperin. Rational LS category and its applications.Trans. Amer. Math. Soc., 273(1):1–38, 1982

work page 1982

[7] [7]

Félix and J.-C

Y. Félix and J.-C. Thomas. Sur la structure des espaces de LS catégorie deux.Illinois J. Math., 30(4):574–593, 1986

work page 1986

[8] [8]

Fernández-Suárez

L. Fernández-Suárez. On a problem of Ganea about strong category.Topology Appl., 102(2):207–218, 2000

work page 2000

[9] [9]

T. Ganea. Lusternik-Schnirelmann category and strong category.Illinois J. Math., 11:417–427, 1967

work page 1967

[10] [10]

K. P. Hess. A proof of Ganea’s conjecture for rational spaces.Topology, 30(2):205–214, 1991

work page 1991

[11] [11]

Lemaire and F

J.-M. Lemaire and F. Sigrist. Sur les invariants d’homotopie rationnelle liés à la L. S. catégorie.Comment. Math. Helv., 56(1):103–122, 1981

work page 1981

[12] [12]

D. Quillen. Rational homotopy theory.Ann. of Math. (2), 90:205–295, 1969

work page 1969

[13] [13]

Reutenauer.Free Lie algebras, volume 7 ofLondon Mathematical Society Monographs

C. Reutenauer.Free Lie algebras, volume 7 ofLondon Mathematical Society Monographs. New Series. The Clarendon Press, Oxford University Press, New York, 1993. Oxford Science Publications

work page 1993

[14] [14]

Scheerer and D

H. Scheerer and D. Tanré. Fibrations à la Ganea.Bull. Belg. Math. Soc. Simon Stevin, 4(3):333–353, 1997

work page 1997

[15] [15]

D. Stanley. Spaces and Lusternik-Schnirelmann categorynand cone lengthn+ 1.Topology, 39(5):985–1019, 2000

work page 2000

[16] [16]

Sullivan

D. Sullivan. Infinitesimal computations in topology.Inst. Hautes Études Sci. Publ. Math., 47:269–331, 1977

work page 1977

[17] [17]

F. Takens. The Lusternik-Schnirelman categories of a product space.Compositio Math., 22:175–180, 1970

work page 1970

[18] [18]

Tanré.Homotopie rationnelle: modèles de Chen, Quillen, Sullivan, volume 1025 ofLecture Notes in Mathematics

D. Tanré.Homotopie rationnelle: modèles de Chen, Quillen, Sullivan, volume 1025 ofLecture Notes in Mathematics. Springer-Verlag, Berlin, 1983. 14 PAUL-EUGÈNE PARENT AND DANIEL TANRÉ Département de Mathématiques, Complexe STEM, pièce 361, 150 Louis-Pasteur Pvt, Université d’Otta w a, Otta w a, ON, K1N 6N5, Canada Email address:pparent@uottawa.ca Départemen...

work page 1983