Cited by 280Homotopy and Cohomology in Algebraic TopologyGeometric and Algebraic TopologyAdvanced Algebra and Geometry

The DOI matches its Crossref record; 5 of the 5 references it cites were found in a registry or an index; none of the 6 sources its claims cite is checked for support yet.

An application of the Morse theory to the topology of Lie-groups

Open access
Bott, Raoul·1956·Bulletin de la Société mathématique de France

Abstract

This paper is primarily devoted to a detailed account of the results announced and sketched in [5 |.It will be followed by a joint paper witli H. SAMELSON on the loop-space of symmetric spaces in general.We recall the principal results of [3].Let G be a compact Lie group; T'C G shall denote a torus in G and C( T) the centralizer of T in G.The space of loops on G is denoted by ^2(G).(For detailed definitions, see §9.) THEOREM A. -If G is a connected^ simply connected^ compact Lie groups then the spaces 12 (G) aud G/C( T') have the following properties : a.They are free of torsion; b.Their odd Betti numbers vanish; c.Their Betti numbers can be read off from the diagram of G. '25l> R. BOTT.THEOREM B. -1° The Poincare series of the loop-space, i2(G'), is given by P(^(G),t)=^t^\where A runs over the cells of the fundamental chamber ^.2° If^y is a lattice point ofF in t, and if C(^Y) is the centralize r of the direction JT, then the Poincare series of G/C(A^) is given bywhere now the summation is extended over the cells of 5 7 which contain X in their closure.As every C(7'' / ) is conjugate to some C(^T), ^T^r, part 2° of this theorem describes the homology of all the spaces G/C(T').It also relates their Poincare series to that of^(G).These formulae can of course be expressed entirely in terms of the rootforms of G on t : Let { 0; \rj (i' = i, . .., /n), denote the root-forms on t which take positive values in 5 7 '.Also let H^ denote the Weyl group ofG in t.Finally, ifJTet, define a function ^*, on W by : A* (w) == number of the root-forms j 6^-J£F whose values at X and w'A" differ in sign.Thus, A *(w) == number of planes of the infinitesimal diagram D'\ crossed by the line from X to w'X.Finally let p(^T) denote the order of the subgroup of W which leaves JT fixed.In terms of these notions, theorem B is equivalent to : THEOREM B'. -i° The Poincare series of^l(G) is given by " 2^[8,W] P^(G)^t)=---j t^ dv, I A 1 ŵhere the integral is taken overly \ A | denotes the volume of a fundamental cell o/Z), and [a] denotes the greatest integer less than a; 2° The Poincare series of G/C(^T), X arbitrary in t, is given by P{(G/CW);t}^--.^t^.w€Î n particular (for X in ^) the iq^ Betti number of G/T is equal to the

Overview

AI snapshot

Written from the paper’s full text.

TL;DR

Bott applies Morse theory to the Riemannian geometry of compact connected Lie groups to study path and loop spaces, showing these spaces are torsion-free with vanishing odd Betti numbers. The paper derives Poincaré-series formulas in terms of the group’s diagram and Weyl group, and uses them to compute Betti numbers for exceptional groups.

Key findings

  • The path space Ω(G,N,P) is torsion-free, has no odd Betti numbers, and its Poincaré series equals the Morse series of (G,N,P) (Theorem I).
  • The Morse series is given by summing t^{λ(s)} over straight-line segments in S(t,N,P), where λ(s) counts how many planes of the diagram D(G) the segment crosses (Theorem II).
  • For G/T, the q-th Betti number equals the number of Weyl-group elements that change the sign of exactly q positive root forms (Theorem B′).
  • Appendix II computes Betti numbers for the exceptional groups E6, E7, and E8; the text reports E6 dimension 78 and E7 dimension 133, and gives Poincaré polynomials for these groups.

Methods

  • The study uses a compact connected Lie group G with a bi-invariant Riemannian metric, and analyzes spaces of paths starting on an adjoint orbit N and ending at a regular point P.
  • The adjoint action is shown to be variationally complete, so geodesics reduce to straight lines in a maximal torus t; the Morse index of a segment is computed by counting intersections with the diagram D(G) or D′(G).
  • Morse inequalities are used; because the relevant indices are even, the inequalities become equalities, yielding Poincaré series and torsion-freeness results.

Limitations

  • The paper states that at the time there was no alternate proof that Ω(G) is free of torsion.
  • It notes that the Morse series may fail to exist if its coefficients are infinite; existence is likely when π1(M,N) is finite, but no proof is given and it is not needed for the applications.
  • The general variational-completeness case does not give equalities for dimensional reasons; that broader situation is deferred to a subsequent paper with Samelson.
  • Evident from the text: the OCR/scan is fragmentary and garbled in places, making some formulas and numerical details difficult to verify.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.24033/bsmf.1472

Cites

5 of 5 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

0 of 6 cited sources checked · 0 supported · 0 flagged

  • Findingreasoning
    § 4. THE MORSE SERIES OF M mod TV.

    The Morse series has only a countable number of terms.

    “In any case the series 4.J has only a countable number of terms.”
    Quote in the paperNot a substantial claim
  • Findingreasoning
    § 9. THE DIAGRAM OF G.

    If σ in T is on precisely l of the sets Ui, then dim C(σ)=l+2l.

    “Further ifo-e T is on precisely/of the sets Ui(i=z i, . .., ni) then dim C (o-) == /+ 2/, and hence (9.2) (W==2/”
    Quote in the paperSubstantial claim
  • Findingreasoning
    § A=[^^\^e^}, ^==JA[A€^J.

    The function f from A to £ is onto.

    “(2)/is onto. Let the closure of A e ^, contain the lattice point Y. Then the closure of y-1 A contains the origin of t, and therefore differs from Ay by an element of W.”
    Quote in the paperSubstantial claim
  • Findingcitation
    § 3. FOCAL POINTS.

    Let s be a geodesic segment of (M,N). Then the index of s relative to N is always a finite integer.

    “PROPOSITION 3.2.-Let s be a geodesic segment of (M^ N).Then the index of 5, relative to TV, defined by (3 9) ^N(s)==^dlmA N (s f ) s'Cs is always a finite integer.”
    Quote in the paperSubstantial claim
  • Findingreasoning
    § 9, (JT) == o mod i;. 1=1

    If G and G1 are locally isomorphic, such an isomorphism induces corresponding isomorphisms of W onto W1.

    “Thus if G and GI are locally isomorphic, such an isomorphism induces corresponding isomorphisms of W^ onto W^_.”
    Quote in the paperNot a substantial claim
  • It may state two separate claims as one. 4
  • Probably not a substantial claim. 3
  • The quote only partly states this claim. 5
  • It may state two separate claims as one.

    Findingreasoning
    § 8. THE GEODES1CS OF (G, TV, P).

    Every segment s in S(G,N,P) is perpendicular to o(P) at its end point and tangent to T at P; hence s lies in T.

    “Hence every segment s<E S (G, 7V, P) (which must be perpendicular to the orbit N at its start) is perpendicular to o (P) at its end point.The segment s is therefore tangent to T at P. Because T is a subgroup, s will have to lie in T.”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim.

    Findingnone given
    § 9, (JT) == o mod i;. 1=1

    The extended Weyl group W is generated by W and the covering transformations of t→T.

    “This^ is the transformation group W of t, generated by W and the covering transformations of i-> T.”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim.

    Findingreasoning
    § j--\

    D'(G) consists of the points which are in the kernel of at least one of the root forms θ_i.

    “In any case D' {G) consists of the points which are in the kernel of at least one of the 9^.”
    Quote in the paperSubstantial claim
  • The quote only partly states this claim.

    Findingreasoning
    § 8. THE GEODES1CS OF (G, TV, P).

    Conversely, a geodesic segment on T joining T' in N∩T to P is in S(M,N,P).

    “Conversely a geodesic segment on T joining T'eA'nT 7 lo P, is perpendicular to o(r) at its initial point and therefore is in S(M, A, P).”
    Quote in the paperSubstantial claim
  • The quote only partly states this claim.

    Findingnone given
    § 9, (JT) == o mod i;. 1=1

    The lattice Γ is left invariant by the action of W.

    “The lattice T is left invariant by the action of W.”
    Quote in the paperSubstantial claim
  • The quote only partly states this claim.

    Findingnone given
    § --i /'

    For a compact connected simply-connected Lie group G, a maximal torus T, an orbit N of the adjoint action, and a general point P of T not on N, the space of paths Ω(G,N,P) is free of torsion.

    “THEOREM I. -The space of paths, ^l (G, N, P), is free of torsion.”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: G is a compact connected and simply-connected Lie group.
    • The authors state this as an assumption: T is a maximal torus of G.
    • The authors state this as an assumption: N is an orbit of the adjoint action of G on G.
    • The authors state this as an assumption: P is a general point of T not on N.
  • It may state two separate claims as one.

    Findingnone given
    § 9. THE DIAGRAM OF G.

    The adjoint action of T on g decomposes g into orthogonal invariant subspaces g = t ⊕ e1 ⊕ e2 ⊕ ... ⊕ em, with t pointwise fixed under Ad|T, while each ei is a two-plane on which T is represented nontrivially by rotations.

    “This action of T on 9 decomposes q into orthogonal invariant subspaces, (^i) 9= : t®el®e,©...®e^ with t poinlwise fixed under Ad | 7 7 , while each Oi is a two-plane on which T is represented nonlrivially by rotations.”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Findingnone given
    § 9, (JT) == o mod i;. 1=1

    A fundamental chamber singles out precisely m root forms which take positive values in it.

    “A fundamental chamber 5^, singles out precisely m-root forms which take on positive values in ^F.”
    Quote in the paperNot a substantial claim
  • The quote only partly states this claim.

    Findingreasoning
    § Additional information

    The function space Ω(M,N,P) is of the same homotopy type as N.

    “PROPOSITION.11.5.-The function space i2(Af, TV, P) is of the same homotopy type as TV.”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: M is an open spherical disc about the origin in g, so small that exp|M maps M homeomorphically into G.
  • It may state two separate claims as one.

    Findingreasoning
    § --i /'

    Let s be a transversal segment on G. Then: (a) every element V∈J^ is of the form V=π_*(Y)+ω_*(Z), with Y∈g, Z∈c(s); (b) if for V∈J^, Z∈c(s), π_*(Y)+ω_*(Z)=0.

    “LEMMA 7 . 4-. -Let s be a transversal segment on G. Then : (a) Every element V ^.J^ is of the form r==7r.,(r)+c*),s.(Z),[reg, Zec(^)]; (b) If for Fe^ Z^c(s), 7T,(r)+CO,(Z)=0.”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: G is a compact connected and simply-connected Lie group.
  • The quote only partly states this claim.

    Findingreasoning
    § --i /'

    The Poincare series of Ω(G,N,P) coincides with the Morse series of (G,N,P).

    “Its odd Betti numbers vanish, and the Poincare series of Sl(G, N, P) coincides with the Morse series of (G, TV, P).”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: G is a compact connected and simply-connected Lie group.
    • The authors state this as an assumption: T is a maximal torus of G.
    • The authors state this as an assumption: N is an orbit of the adjoint action of G on G.
    • The authors state this as an assumption: P is a general point of T not on N.
  • It may state two separate claims as one.

    Premisereasoning
    § 9, (JT) == o mod i;. 1=1

    The closure of every fundamental cell is compact and contains precisely one point of the lattice Γ.

    “(9.io)The closure of every fundamental cell is compact and contains precisely one point of the lattice r.”
    Quote in the paperSubstantial claim
12 of 12
  • The authors state this as an assumption.

    G is a compact connected and simply-connected Lie group.

    “Let G be a compact connected and simply-connected Lie group, TC.G a maximal torus of G, N any orbit of the adjoint action of G on G.”
  • The authors state this as an assumption.

    T is a maximal torus of G.

    “Let G be a compact connected and simply-connected Lie group, TC.G a maximal torus of G, N any orbit of the adjoint action of G on G.”
  • The authors state this as an assumption.

    N is an orbit of the adjoint action of G on G.

    “Let G be a compact connected and simply-connected Lie group, TC.G a maximal torus of G, N any orbit of the adjoint action of G on G.”
  • The authors state this as an assumption.

    P is a general point of T not on N.

    “Let P be a general point of T not on N.”
  • The authors state this as an assumption.

    M is an open spherical disc about the origin in g, so small that exp|M maps M homeomorphically into G.

    “Let Mbe an open, spherical disc about the origin in $, which is so small that exp | Sl maps Sl homeomorphically into 6r.”
  • The authors state this as an assumption.

    G is simple, connected, simply connected and compact.

    “Let G be simple^ connected^ simply connected and compact.”
  • The authors state this as an assumption.

    The Morse series of (M,N,P) exists, with P a regular point of M-N.

    “Suppose then that the Morse series of (M, N, P) exists (P a regular point of M-N).”
  • The authors state this as an assumption.

    E∞ is of finite type.

    “Suppose now that E^ is of finite type so that the above proposition applies.”
8 of 8

Citations

Cites

Flagged first, then by citations

Research works printed but not linked: 2

  1. [2]2. C. CHEVALLEY, Theory of Lie groups I (Princeton, New Jersey, Princeton University press, 1946).
  2. [4]4. A. BOBEL, a. Sur la cohomologie des espaces fibres... {Ann. Math.^ t. 57, i953, n 0 1, p. n5). A. BOREL, b. Kdhlerian^ coset spaces of semisimple Lie groups {Proc. Nat. Acad. Sc., t. 40, 1954, p. u47-n5i).

Cited by

Sorted by citations

Metadata

Authors
Bott, R.
Corresponding author
Bott, Raoul
Venue
Bulletin de la Société mathématique de France
Published
1956
Volume
79
Pages
251-281
ISSN
  • 0037-9484 (ISSN-L)
  • 2102-622X
OpenAlex
W35421105
Cited by
280
References
5
Citations by year
2026: 4 · 2025: 5 · 2024: 5 · 2023: 3 · 2022: 7 · 2021: 3 · 2020: 5 · 2019: 4 · 2018: 4 · 2017: 17 · 2016: 9 · 2015: 6 · 2014: 5 · 2013: 6 · 2012: 6
FWCI
0.00
Type
journal article
Language
zh
Open access
Locations
  • Bulletin de la Société mathématique de France (publishedVersion)
  • French digital mathematics library (Numdam) (submittedVersion)
Topics
  • Physical Sciences › Mathematics › Mathematical Physics › Homotopy and Cohomology in Algebraic Topology
  • Physical Sciences › Mathematics › Geometry and Topology › Geometric and Algebraic Topology
  • Physical Sciences › Mathematics › Algebra and Number Theory › Advanced Algebra and Geometry
Keywords
Morse theory, compact Lie groups, loop spaces, Poincaré series, Betti numbers, root form, Weyl group, simply connected Lie groups, free abelian groups, symmetric spaces

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-10-11, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-11 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review