COUNTING GEODESICS ON COMPACT SYMMETRIC SPACES

LUCAS SECO∗ AND MAURO PATRÃO∗

Abstract. We describe the inverse image of the Riemannian exponential map

at a basepoint of a compact symmetric space as the disjoint union of so called
focal orbits through a maximal torus. These are orbits of a subgroup of the
isotropy group acting in the tangent space at the basepoint. We show how
their dimensions (infinitesimal data) and connected components (topological
data) are encoded in the diagram, multiplicities, Weyl group and lattice of the
symmetric space. Obtaining this data is precisely what we mean by counting
geodesics. This extends previous results on compact Lie groups. We apply
our results to give short independent proofs of known results on the cut and
conjugate loci of compact symmetric spaces.

AMS 2010 subject classification: Primary: 53C22, 53C35; Secondary: 53C20.

Keywords: Geodesics, Symmetric spaces, Lie groups, Global Riemannian geometry.

1. Introduction
The set of smooth geodesics connecting two points p and q of a Riemannian
manifold M can be viewed as the tangent directions at p in the inverse image of
q by the Riemannian exponential map expp : Tp M → M at the basepoint p. To
describe this set is roughly what we mean by counting geodesics.
We can use the action of the isometry group of M to move around a geodesic
γ connecting p and q, producing more geodesics connecting them. Indeed, let K
be isotropy subgroup of isometries that fix p and let K q be the subgroup of K
of isometries that further fix q. Then the whole orbit K q γ consists of geodesics
between p and q. This orbit can also be seen in the tangent space Tp M to p if we
identify the geodesic γ with its initial velocity H = γ 0 (0). Then γ(t) = expp (tH),
q = γ(1) = expp (H) and, for k ∈ K, kγ(t) = expp (tkH), where we denote by
kH = dp kH the isotropy action of K in Tp M . The orbit K q γ then gets identified
with so called focal orbit of H
(1) F (H) = K q H ⊆ Tp M
It follows that K q acts in the inverse image exp−1
p (q) which is, thus, a union of focal
orbits. Describe this union and count the dimension and connected components of
each focal orbit is precisely what we mean by counting geodesics.
In this article we count the geodesics of compact Riemannian symmetric spaces
S = U/K: they are natural candidates for that, since their symmetry properties
allows us to study their isotropy orbits in terms of Lie algebraic data. We now state
our main results. Let t ⊆ Tp M be a maximal flat torus with corresponding lattice
(2) Γ = {γ ∈ t : expp (γ) = p}
1
2 L. SECO AND M. PATRÃO

restricted roots, Stiefel diagram S , root hyperplanes and Weyl group W (see Sec-
tion 2). Let q = expp (H) with H ∈ Tp M , using the isotropy action of K in Tp M
we can assume that H ∈ t. Consider in W the centralizer of H mod Γ given by
(3) W q = {w ∈ W : wH ∈ H + Γ}
and the normal subgroup W0q generated by reflections around root hyperplanes
paralell to the hyperplanes in S which cross H.
Theorem A. The geodesics between p and q of the compact symmetric space
S = U/K are given by the disjoint union of focal orbits through H + Γ
[
exp−1
p (q) = F (H + γ)
γ∈Γ

where:
(1) dim F (H + γ) = equals to the number of non-root hyperplanes of S that
cross H + γ, counted with multiplicity.
(2) F (H +γ) intersects t in the orbit W q (H +γ), its connected components are
diffeomorphic to one another, correspond to the left W0q -orbits in W q (H +γ)
and are in bijection with the quotient group W q /W0q .
(3) Each connected component of F (H +γ) corresponds to a unique homotopy
class of geodesics between p and q.
This result shows how infinitesimal (dimension) and topological (connected com-
ponents) data of the inverse image of the exponential of S can be read from data in
a maximal flat torus of S, showing how we can count geodesics just by looking at
t. Note that the lattice Γ is topological data of S = U/K (see Theorem B), while
the roots, their multiplicity, Stiefel diagram and Weyl group is infinitesimal data
and depends only on (u, k).
This extends previous results of [SS18], where the subgroup (3) was introduced
to count the geodesics of compact Lie groups (see Example 1.4) Note that, since
focal orbits of S can be disconnected (see Section 1.1), but are always connected
for simply connected S (see Theorem C), we cannot obtain our general result by
projecting it from the result on the simply connected covering. For the same reason,
since only simply connected S is guaranteed to embed in its isometry group U , we
cannot obtain the general result from the compact group case. Note, furthermore,
that the universal cover of S is not compact when S has an euclidean factor.
Remark 1.1. If one wishes to obtain all the geodesics of a fixed length |H| between
p and q, one first obtains the so called focal equivalents of H on t, i.e. those Γ-
equivalents H +γ with length |H +γ| = |H|, then considers the focal orbits through
them. The focal equivalents can be obtained from the lattice Γ by a geometric
construction involving mediator hyperplanes (see Section 4 of [SS18]).
We apply Theorem A to extends to compact symmetric spaces S the following
result, which is well known for symmetric spaces of compact type (see Proposition
VI.2.4 of [Loo69]).
Theorem B. Let Γ0 be the fundamental lattice. Then π1 (S) = Γ/Γ0 .
We use this to prove the following connectedness result.
Theorem C. Let S be simply connected. Then Γ = Γ0 and W q = W0q . In
particular, K q is connected and thus all focal orbits are connected.
 COUNTING GEODESICS ON COMPACT SYMMETRIC SPACES 3

We then apply our results to provide short independent proofs of classical results
by Crittenden [Cri62] and Sakai [Sak77] that characterize, respectively, the first
conjugate locus and the cut locus of a compact symmetric space.

The structure of the article is as follows. In Section 2 we introduce the notation

and standard results of our setup. In Section 3 we extend to a maximal flat torus
of a compact symmetric space a diagonalization result about the action of the Weyl
group. In Section 4 we prove Theorem A. In Section 5 we prove Theorem B. In
Sections 6 we apply our results to provide independent proofs of classical results
by Crittenden and Sakai.
In a forthcoming paper we will use the results of this article to refine the results
of Bott [Bot59] about the space of minimimal geodesics of S and provide a related
characterization of symmetric flag manifolds.

We would like to thank K-H. Neeb and C. Gorodski for helpful conversations
about symmetric spaces.
1.1. Examples. We finish this introduction by illustrating our main results in
examples of low rank.
A rank 0 Scompact symmetric space is a torus T , for which it is well known that
exp−1
p (q) = γ∈Γ (H + γ) and π1 (T ) = Γ. These become special cases of Theorems
A and B, since the Stiefel diagram, fundamental lattice and Weyl group of a torus
are trivial.
Example 1.2. (rank 1) The round 2-sphere S 2 ⊆ R3 , has exponential map expp (H) =
cos(|H|)p + sin(|H|)H/|H| for H 6= 0. The inverse image exp−1 p (q) is a concentric
union of circles of radii 2kπ if q = p, a concentric union of circles of radii (2k + 1)π
if q = −p, otherwise it is the discrete set H + 2πkH/|H| along the direction of H,
where expp (H) = q, k ∈ Z.
This is recovered by Theorem A as follows. S 2 = SO(3)/SO(2) is a symmetric
space, where the isotropy K = SO(2) of p = (0, 0, 1) is contained in the fixed point
set of the conjugation of SO(3) by the reflection on the z-axis. The tangent plane
Tp S 2 is the (x, y)-plane with the isotropy action of SO(2), where the x axis is a
maximal flat torus t with simple root α(x) = x with multiplicity 1, thus Stiefel
diagram S = πZ, whose nonzero points are conjugate tangent points along x, and
lattice Γ = 2πZ.
Let q = expp (H). Then q = p iff H ∈ Γ, in this case K q = K so that focal
orbits through non-zero tangent vectors are circles and exp−1 p (q) is the union of
concentric circles centered on the origin passing through Γ, thus circles of radii
2kπ. Furthermore q = −p iff H ∈ π + Γ, in this case once again K q = K so that
exp−1p (q) is the union of circles centered on the origin passing through π + Γ, thus
circles of radii (2k + 1)π. Note that S = Γ ∪ (π + Γ) so that in both previous cases
all focal orbits cross the diagram, thus W0q = {±1}. Furthermore, W q = {±1} so
that the focal orbit through H + γ meets t in ±(H + γ) and is connected. Finally,
if q 6= ±p, then K q = 1 so that the focal orbits are singletons and exp−1 p (q) is the
discrete set H + Γ in the x axis. Note that in this case no focal orbit crosses the
diagram and W q = W0q = 1. In all cases both K q and (thus) all focal orbits are
connected.
Let π : S 2 → RP 2 , π(v) = ±v, be the canonical projection onto the real pro-
jective 2-space. RP 2 = SO(3)/(SO(2) ∪ (−I)SO(2)) is a symmetric space, where
4 L. SECO AND M. PATRÃO

K = SO(2) ∪ (−I)SO(2) is the isotropy of [p] = (0, 0, ±1) and I is the 3 × 3 identity
matrix. The exponential at [p] is π ◦expp , so that the infinitesimal setup is the same
as the 2-sphere, now with lattice Γ = πZ. Let q = (1, 0, 0) then K [q] = {I, −I} is
disconnected. Let H = π/2 ∈ t, then [q] = π ◦ expp (H) and exp−1 p (q) is the union of
the disconnected focal orbits ±π(1/2 + 2k), k ∈ Z. Note that no focal orbit crosses
the diagram so that W0q = 1, while W q = {±1}.

Example 1.3. (rank 2) Let S = Gr2 (R4 ) be the Grassmanian of planes in R4 , de-
note by [u, v] ∈ S the plane spanned by two l.i. vectors u, v ∈ R4 and by e1 , e2 , e3 , e4
the canonical basis of R4 . S is a symmetric space of SO(4) with the canonical bi-
invariant metric and orthogonal involution given by conjugation by I2,2 = I0 −I 0


,
which has fixed point set K = S(O(2) × O(2)), precisely the isotropy of p = [e1 , e2 ]
in SO(4).
We have that k = {( A0 B0 ) : A, B ∈ so(2)}, thus its orthocomplement is s =
0 −C t : C ∈ gl(2) , with maximal torus t = 0 −D t : D = θ1




C 0 D 0 θ2 depicted
below as vectors (θ1 , θ2 ).
θ2

H θ1
H 00

The simple restricted roots are α = θ1 − θ2 , β = θ1 + θ2 , each with mutiplicity 1

and, since these roots are orthogonal, this is a symmetric space of type A1 × A1 .
The corresponding coroot vectors are Hα∨ = (1, −1), Hβ∨ = (1, 1) with fundamental
lattice Γ0 = {(πk, πl) ∈ Γ : k, l ∈ Z such that k + l is even}, depicted by the larger
dots above. The Stiefel diagram is then given by θ1 − θ2 , θ1 + θ2 ∈ πZ, depicted by
the slanted lines above, where the root hyperplanes are depicted as thicker lines.
Since expp (X) = eX [e1 , e2 ], for X ∈ s, where
!
−θ1 ! !
−θ2 cos(θ1 ) − sin(θ1 ) 1
θ1 1 cos(θ2 ) − sin(θ2 )
θ2
= sin(θ1 ) cos(θ1 ) 1
1 sin(θ2 ) cos(θ2 )
e

it follows that the lattice of S is Γ = π(Z × Z), depicted by the dots above. The
Weyl group W is given by either the symmetries of the dark grey square above,
with vertices (± π2 , ± π2 ), which is the closure of the Dirichilet domain D of Γ. By
Theorem A it follows that π1 (S) = Γ/Γ0 = Z2 .
 COUNTING GEODESICS ON COMPACT SYMMETRIC SPACES 5

Let q = [e3 , e4 ] and note that q = expp (H), for H = ( π2 , π2 ) ∈ t. By inspecting

the diagram, it follows that W0q = W , which implies that W q = W , so that all
focal orbits F (H + γ) are connected. Furthermore, note that each H + γ crosses
either one or two nonroot hyperplanes, thus F (H + γ) is connected with dimension
1 or 2. Note that K q = K, thus the focal orbits F (H + γ) are flag manifolds of S.
Since S is of type A1 × A1 , it follows that F (H + γ) is a product of projective lines
whenever both α(H + γ), β(H + γ) are not zero, and a projective line if one of them
is zero. It follows that exp−1 p (q) is a disjoint union of 1-spheres passing through
the non-regular elements of H + Γ and tori passing through the regular elements of
H + Γ. Since γ = (πk, πl), k, l ∈ Z, note that H + γ = ( π2 (2k + 1), π2 (2l + 1)).
Note that H lies in the boundary of D and that among the Γ-equivalents to H
with the same norm, both H and rβ (H) = −H lie in the focal orbit F (H), while
both H 0 = (− π2 , π2 ) and rα (H 0 ) = ( π2 , − π2 ) do not. Thus the minimal geodesics
from p to q in S are given by the disjoint union of F (H) and F (H 0 ), a disjoint
union of projective lines,
Now let q 00 = [e3 , e2 ] and note that q 00 = expp (H 00 ), for H 00 = (− π2 , 0) ∈ t. By
inspecting the diagram, it follows that W0q = 1 and W q = reflection around the θ2
axis, thus all focal orbits F (H 00 +γ) have two connected components. Furthermore,
note that none H 00 +γ = (πk, π2 (2l+1)) crosses root hyperplanes, thus F (H 00 +γ) =
(πk, ± π2 (2l+1)). If follows that exp−1 00
p (q ) is a disjoint union of 0-spheres in H +Γ.
00

The universal cover of S is the Grassmanian of oriented planes Se = Gr f 2 (R4 ),

which is a symmetric space with the same setup as above, except for the isotropy
K = SO(2) × SO(2) of the oriented plane p = [e^ 1 , e2 ] and the lattice, which is now
the fundamental lattice Γ0 . Now all focal orbits are connected (by Theorem C),
with circles (or products) taking place of projective lines (or products). Another
way to obtain this is to use that Se is isometric to product of 2-spheres, since it
is simply connected of type A1 × A1 , and use Example 1.2. Note that H also lies
in the boundary of the Dirichilet domain D0 of Γ0 , given by the light grey square
above, with vertices (±π, 0), (0, ±π). Now the Γ0 -equivalents to H with the same
norm are only H and rβ (H) = −H, which lie in the focal orbit F (H). Thus, the
minimal geodesics from p to q in Se are given by the connected focal orbit F (H), a
circle.
For n > 4, the Grassmannian Gr2 (Rn ) is a symmetric space of rank 2 with setup
analogous to the above, except that now θ1 , θ2 are also positive roots, so that it
is irreducible of type B2 , with additional vertical and horizontal lines θ1 , θ2 ∈ πZ
in the above Stiefel diagram. In particular, its universal cover Gr f 2 (Rn ) is not a
product.

Example 1.4. Taking care to divide by 2, the above symmetric space results for
H/2 generalize the compact Lie group results of [SS18] for H.
Indeed, a compact connected Lie group U with a biinvariant metric is a sym-
metric space of U × U acting on itself by left translations and acting on U by left
and right isometries (g, h) · u = guh−1 . The isotropy K of the basepoint p = 1 ∈ U
is given by gh−1 = 1 so that K = {(u, u) : u ∈ U } is the diagonal of U , which
is the fixed point set of the involution σ(g, h) = (h, g). Note that the action of
the isotropy K in U is given by conjugation. Thus, the isotropy K h of h ∈ U is
the diagonal of the centralizer Uh of h. The equivariant diffeomorphism that maps
6 L. SECO AND M. PATRÃO

S = U × U/K to U intertwining the respective actions is given by the slope map

π((g, h)K) = gh−1 .
The (−1)-eigenspace of the involution on u × u is s = {(X, −X) : X ∈ u}. Let t
be a maximal torus of u, then a = {(H, −H) : H ∈ t} is a maximal torus of s. The
differential of π at K maps s → u, and is given by
(X, −X) 7→ 2X
It maps the roots of S to the roots of U , the lattice Γ of S to the unit lattice Γ of
U , the diagram of S to the diagram of U . It also maps the isotropy action of K
in s to the adjoint action of U in u, thus the Weyl group action of S to the Weyl
group action of U and the isotropy action of K h in s to the adjoint action of the
centralizer Uh in u. Finally, the focal orbit K h (H, −H) ⊆ s gets mapped to the
centralizer orbit Uh (2H) ⊆ u.
Note that the unit quaternions S 3 as a Riemannian manifold with induced metric
from R4 has I(S 3 )0 = SO(4). As a compact group with the biinvariant metric, thus
as a symmetric space, it also has I(S 3 )0 = SO(4), since S 3 has an S 3 × S 3 action
by isometries, as above, thus a homomorphism S 3 × S 3 → I(S 3 )0 = SO(4) which
is surjective since it is, in fact, the universal covering of SO(4).

2. Preliminaries
Let us recall the definitions, notations and basic results on compact symmetric
spaces and their Lie algebras (see for example [Bor98], [Hel78] or [Sak77]). Since
many of the standard results in the literature are for symmetric spaces of com-
pact type, thus without euclidean factor, we take care to extend them to compact
symmetric spaces, whenever needed.
A compact Riemannian symmetric pair (U, K) consists of: a compact connected
Lie group U , a closed subgroup K, a U -invariant Riemannian metric on S = U/K
and an involutive automorphism σ of U such that K lies between U0σ and U σ ,
where U σ is the subgroup of fixed points of σ. For short, we will call this a compact
symmetric space. Denote by u, k the Lie algebras of U , K, respectively, and again
by σ the involutive automorphism of u given by the differential at the identity
of the automorphism of U . Then (u, k) is a compact orthogonal symmetric pair
corresponding to (U, K), where k = (+1)-eigenspace of σ. Let s = (−1)-eigenspace
of σ, then u = k ⊕ s is an orthogonal decomposition. For short, we will call this a
compact symmetric Lie algebra. We may identify the the tangent space Tp S with s
via the canonical projection π : U → S, where p = π(1) is the canonical basepoint.
In the following we shall fix in u an Ad(U )-invariant inner product. Such inner
product is essentially unique in the semisimple factor of u and is arbitrary on the
center of u. We then restrict it to s, obtaining an Ad(K)-invariant inner product
which extends uniquely to a U -invariant Riemannian metric on S = U/K. Then
the Riemannian exponential map of S at p is then given by
(4) expp : s → S, X 7→ exp(X)p = π(exp(X))
where exp : u → U denotes the exponential mapping of the Lie group U .
2.1. Restricted roots. Let t ⊆ s be a maximal (flat) torus. They are all conjugate
by the action of Intu (k) in s. To a real functional α ∈ t∗ corresponds the restricted
root space
(5) uα = {X ∈ uC : [H, X] = iα(H)X, H ∈ t}
 COUNTING GEODESICS ON COMPACT SYMMETRIC SPACES 7

√
where i = −1 and we do not introduce the 2π factor used in [Sak77] and [SS18].
The (restricted) roots of (u, k) are given by Π = {α ∈ t∗ : uα 6= 0} with correspond-
ing restricted root space decomposition of uC given by
X
(6) uC = u0 ⊕ uα
α∈Π

The Stiefel diagram of the symmetric space is the union of the affine hyperplanes
(7) S = {H ∈ t : α(H) ∈ πZ, for some α ∈ Π}
where the root hyperplanes are the hyperplanes through the origin.
The set of regular elements treg is the complement of the Stiefel diagram in t.
A connected component of the complement of treg is an alcove, and a connected
component of the complement of the root hyperplanes is a Weyl chamber. The
Weyl group acts simply transitively on the Weyl chambers.
Fix a Weyl chamber t+ and corresponding positive roots Π+ and simple roots
Σ ⊆ Π+ . For α ∈ Π+ let kα = k ∩ (uα ⊕ u−α ), sα = s ∩ (uα ⊕ u−α ), and let m =
centralizer of t in k, then
X X
(8) k=m⊕ kα s=t⊕ sα
α∈Π α∈Π

where dim kα = dim sα = mα , the multiplicity of α. Note that u0 = m ⊕ t.

2.2. Centralizers and the center. Restricted roots of the compact symmetric
Lie algebra (u, k) are related to the roots of the compact Lie algebra u as follows. Let
et be maximal torus of u which contains t. To a real functional α e ∈ et∗ corresponds
the root space
euαe = {X ∈ uC : [H, X] = ie α(H)X, H ∈ et}
The roots of u are given by Π e = {e α ∈ et∗ : euαe 6= 0} and we have the root space
decomposition X
uC = eu0 ⊕ uαe
e
e ∈Π
α e
Since the adjoint representation of t in uC is the restriction of the adjoint represen-
tation of et, it follows that the restricted root spaces of (u, k) split as a sum of roots
spaces of u so that, for α ∈ Π, we have
X
uα = uαe
e
e|t =α
α

e 0 = {e
Let Π α∈Π e: αe|t = 0}, it follows that Π = {e
α|t : α
e∈Π e −Π
e 0 }. Furthermore,
a Weyl chamber of t can be chosen with corresponding positive roots Π e + so that
+ +
Π = {e α|t : α
e∈Πe −Π e 0 }.
For H ∈ t, we have that ad(H) is diagonal in uC with 0-eigenspace eu0 and iα(H)-
eigenspace uα , α ∈ Π. For h = exp(H) we have that Ad(h) = ead(H) is diagonal
in uC with 1-eigenspace e u0 and eiα(H) -eigenspace uα , α ∈ Π. Thus, by (8) the
centralizers
kH = {X ∈ k : ad(H)X = 0} ⊆ kh = {X ∈ k : Ad(h)X = X}
decompose as
X X
(9) kH = m ⊕ kα kh = m ⊕ kα
α(H)=0 α(H)∈2πZ
8 L. SECO AND M. PATRÃO

where α ranges through the positive restricted roots.

Since u compact is reductive, it splits as the direct sum
(10) u = g ⊕ z(u)
where g = [u, u] is compact semi-simple and z(u) the center of u. The compact
symmetric Lie algebra (u, k) is said to be of compact type when z(u) = 0. Note that
both z(u) and g are σ-invariant, since they are characteristic ideals. Thus
(11) g = (k ∩ g) ⊕ (s ∩ g) z(u) = (k ∩ z(u)) ⊕ (s ∩ z(u))
(12) k = (k ∩ g) ⊕ (k ∩ z(u)) s = (s ∩ g) ⊕ (s ∩ z(u))
(13) t = (t ∩ g) ⊕ (t ∩ z(u))

Let bt = t ∩ g. The rank of the compact symmetric Lie algebra (u, k), hence of the
symmetric space U/K, is dim bt. Restricting σ to g we get a symmetric Lie algebra
of compacty type with maximal torus bt. Since the roots Π of u annihilate t ∩ z(u),
it follows that the roots of g are the roots of u restricted to bt. It also follows that
S = (S ∩ bt ) ⊕ (t ∩ z(u)) so that the diagram hyperplanes are, in fact, the sum of
hyperplanes in bt with t ∩ z(u).
To each restricted root α corresponds the dual root vector Hα ∈ bt such that
hHα , Hi = α(H), for all H ∈ bt. The corresponding coroot vector is Hα∨ =
2Hα /hα, αi, note that α(Hα∨ ) = 2.

Lemma 2.1. (Lemma 1.1 of [Sak77]) For α e+ − Π

e∈Π e 0 , there exist Xαe ∈ k, Yαe ∈ s
such that:
(i) {Xαe : α e|t = α} is a basis of kα , {Yαe : α
e|t = α} is a basis of sα , for α ∈ Π.
Let α e|t = α, then
(ii) [H, Xαe ] = α(H)Yαe , [H, Xαe ] = −α(H)Yαe , for H ∈ t.
(iii) [Xαe , Yαe ] = Hα∨ /2
In particular, since α(Hα∨ /2) = 1, to each positive restricted root α corresponds
a subalgebra u(α) of u spanned by Hα∨ /2, Yα , Xα with bracket relations
(14) [Yα , Xα ] = Hα∨ /2 [Hα∨ /2, Yα ] = Xα [Hα∨ /2, Xα ] = Yα

2.3. Weyl group. Consider the normalizer M ∗ and the centralizer M of t in K.

The Weyl group of the symmetric space U/K is then W = M ∗ /M . We can see an
element w ∈ W either as an element of M ∗ acting by the adjoint action on t or as
a linear isometry of t. We have that W also acts on the roots Π by the coadjoint
action w∗ α(H) = α(w−1 H). By seeing an element of W as an element of M ∗ it
follows at once that the lattice Γ is invariant by the Weyl group. We have that W
is generated by simple reflections rα around the root hyperplanes α = 0, where α
varies in the simple roots and the reflection rα of t around the hyperplane α = 0
is given by rα (H) = H − α(H)Hα∨ . Furthermore, the centralizer WH of H in W
is generated by reflections rα around the root hyperplanes which cross H, where α
varies in the simple roots. In particular, if H is regular then WH is trivial.
Regarding the maximal torus t as diagonal matrices, we have the following si-
multaneous diagonalization result: if t0 ⊆ s is a maximal torus, then there exists
k ∈ K0 such that Ad(k)t0 = t (see Theorem II.2.3 of [Bor98]). In particular,
(15) Given X ∈ s there exists k ∈ K0 such that Ad(k)X ∈ t
 COUNTING GEODESICS ON COMPACT SYMMETRIC SPACES 9

Furthermore, for H ∈ t and k ∈ K, we have that

(16) If Ad(k)H ∈ t then Ad(k)H = wH for some w ∈ W
which can be regarded as: a change of orthogonal base of a diagonal matrix which
diagonalizes it again, can be attained by a permutation in the Weyl group (see
Proposition VII.2.2 p.285 of [Hel78], the proof applies equally to compact symetric
spaces).
2.4. Lattices. Consider the fundamental and the central lattices given, respec-
tively, by
(17) Γ0 = Z-span {πHα∨ : α ∈ Π} ⊆ Γ1 = {H ∈ t : α(H) ∈ πZ, α ∈ Π}
Actually, Γ0 is a lattice of bt. Furthermore, z(u) ⊆ Γ1 so that Γ1 may not be a lattice
of t, nevertheless, projecting onto bt parallel to t ∩ z(u), we get
(18) b 1 = {H ∈ bt : α(H) ∈ πZ, α ∈ Π}
Γ
which is a lattice of bt, in fact is the lattice of the adjoint symmetric space Sb of S
obtained as follows.
Note that the kernel of Ad : U → U b = Ad(U ) is the center Z(U ), a characteristic
subgroup, thus invariant by the involution σ, which then descends to the involution
σ
b of Ub given by σ b(Ad(u)) = Ad(σ(u)). Note that, since Ad(exp(X)) = ead(X) it
follows that σ b(ead(X) ) = ead(σ(X)) . Let K b =U b σb be the fixed point set of σ
b, then
(U , K) is a compact Riemannian symmetric pair with correspoding symmetric space
b b
Sb = U b and compact symmetric Lie algebra g. Since Ad(K) ⊆ K,
b /K b Ad descents
to the projection
(19) S → Sb gK 7→ Ad(g)K b

We have that H ∈ bt belongs to the lattice of Sb iff ead(H) ∈ K,

b thus iff
b(ead(H) ) = ead(−H) = ead(H)
σ
thus iff ead(2H) = 1. Since the eigenvalues of ead(2H) in gC are 1 and e2α(H)i , for α
varying in the roots of g, it follows that the lattice of Sb is given by (18).
Recalling the lattice Γ = {γ ∈ t : expp (γ) = p} of S, the projection (19) then
implies that, projecting onto bt parallel to t ∩ z(u), Γ projects inside Γb 1 . Since the
roots Π of u annihilate t ∩ z(u), it follows that Γ ⊆ Γ1 . Furthermore, we have the
following.
Proposition 2.2. We have that Γ0 ⊆ Γ ⊆ Γ1 . In particular, the Stiefel diagram
S of S is invariant by translations in the lattice Γ of S, so that Γ ⊆ S .
Proof: Let U 0 = S 3 be the unit sphere of the quaternions H, whose Lie algebra is
u0 = Ri⊕Rj ⊕Rk with bracket coming from the quaternionic product. Consider the
orthogonal involution σ 0 (q) = kqk −1 , with fixed point subgroup K 0 = unit circle of
the R ⊕ kR-plane, whose Lie algebra is k0 = Rk, complemented by s0 = Ri ⊕ Rj.
Take H 0 = i/2 in the maximal torus t0 = Ri of s0 , whose exponential is the unit
circle of the R ⊕ iR-plane, thus its intersection with K 0 is {±1} so that U 0 has
lattice Γ0 = πZi.
Let Y 0 = j/2, X 0 = k/2, which have bracket relations [Y 0 , X 0 ] = H 0 , [H 0 , Y 0 ] =
X , [H 0 , X 0 ] = Y 0 . By (14), to each positive root α corresponds a Lie algebra
0

isomorphism u0 → u(α) ⊆ u that maps H 0 7→ Hα∨ /2, Y 0 7→ Yα , X 0 7→ Xα , thus

10 L. SECO AND M. PATRÃO

i 7→ Hα∨ , t0 into t, s0 into s and k0 into k. Since U 0 is simply connected, this

isomorphism can be integrated to a Lie group homomorphism U 0 → U that maps
eθi 7→ exp(θHα∨ ) and K 0 into K. Since eπi = −1 ∈ K 0 , it follows that exp(πHα∨ ) ∈
K so that πHα∨ ∈ Γ.
For H ∈ t, recall the Weyl subgroup W q , centralizer of H modulo the lattice Γ,
and its subgroup W0q , generated by reflections around root hyperplanes paralell to
the hyperplanes in S which cross H. Since W permutes the roots and Γ ⊆ S , we
have the following.
Proposition 2.3. Wq0 is a normal subgroup of W q .
We will need some definitions and results from [SS18]. Two vectors X, Y ∈ t are
Γ-equivalent if X −Y ∈ Γ and are called focal equivalents if, furthermore, |X| = |Y |.
Lemma 2.4 (Lemma 4.2 of [SS18].). If X 6= Y ∈ t are focal equivalents then, for
 > 0, (1 + )X 6= Y + X are Γ-equivalent and such that
|(1 + )X| > |Y + X|
The Dirichlet domain of Γ is given by
(20) D = {H ∈ t : |H| < |H + γ|, γ ∈ Γ − 0}
so that
D = {H ∈ t : |H| ≤ |H + γ|, γ ∈ Γ}
The next result shows that D is a fundamental domain of Γ and characterizes its
geometry.
Proposition 2.5 (Proposition 4.3 of [SS18]). We have that
i) Given H ∈ t, the minimal norm among H + Γ is attained at D.
ii) D is nonempty, convex and compact.
iii) D is the interior of D so that H ∈ D iff H is the unique element of minimal
norm in H + Γ.
iv) Its boundary is ∂D = {H ∈ D : H has nontrivial focal equivalents in t}.
Now denote the alcove at the origin of the Stiefel diagram by
(21) D0 = {H ∈ t : α(H) < π, α ∈ Π}
so that
(22) D0 = {H ∈ t : α(H) ≤ π, α ∈ Π}
Theorem 2.6. If Γ = Γ0 then
i) Its Dirichilet domain is D0 ,
ii) The focal equivalents of H ∈ D0 in t are given by the orbit W0q H,
iii) W q = W0q .
Proof: Items (i) and (ii) are given by Theorem 4.5 of [SS18], where here Γ0 is
πZ-spanned by Hα∨ , instead of Z-spanned. For item (iii), let w ∈ W q and take a
regular H ∈ D◦ . Since wH is clearly focal equivalent to H, by item (ii) we have
that wH = vH, for v ∈ W0q . It follows that v −1 wH = H, and since H is regular,
this implies v −1 w = 1, thus w = v ∈ W0q , as claimed.
Note that this provides an alternative definition of W0q as the centralizer of H
modulo the fundamental lattice Γ0 .
 COUNTING GEODESICS ON COMPACT SYMMETRIC SPACES 11

2.5. Cartan covering. Consider the following involution of U

(23) u∗ = σ(u−1 )
We clearly have that
(24) u∗∗ = u, (uv)∗ = v ∗ u∗ , 1∗ = 1,
(25) uu∗ = uσ(u−1 ) = 1 ⇐⇒ u ∈ fix(σ).
By (25) we have in particular that
(26) kk ∗ = 1, for k ∈ K,
since K ⊆ fix(σ). By (24) we have an action of U on itself given by
(27) u · h = uhu∗
which restricts to conjugation when u ∈ K.
A Riemannian geometric symmetric space S with geodesic reflection sp at p can
be (almost) embeeded in its isometry group by the Cartan map
(28) η : S → I(S), q 7→ sq s−1
p ,

more precisely, this map is a covering onto its image. Indeed, consider U = I(S)0
with involution σ(u) = sp us−1
p and isotropy K = Up . Since sup = usp u
−1
, it follows
that
η(up) = usp u−1 s−1
p = uσ(u
−1
) = uu∗
so that η is U -equivariant, where on the right hand side U acts on itself by the
action (27). It follows that the image of η is the orbit of 1 under this action, which
by (25) has isotropy fix(σ). Thus (28) is a covering onto its image, with discrete
fiber fix(σ)/K.

3. Weyl group action on U/K

In this section we extend to a maximal (flat) torus of a compact symmetric space
U/K the diagonalization result (16) about the action of the Weyl group on t.
Recall that, since the Weyl group W of the symmetric space U/K is given by
infinitesimal data, it is given analytically as either one of the normalizers M ∗ =
NK (t) or NK0 (t). Also, an element w ∈ W can be viewed both as a class kM or
as an orthogonal map Ad(k)|t , where k ∈ NK0 (t). Recall the centralizer M of t in
K, the next result shows that this subgroup intersects each connected component
of K.
Proposition 3.1. K = K0 M .
Proof: Let k ∈ K, then t0 = Ad(k)t is maximal abelian in s. Thus, there exists
l ∈ K0 such that Ad(l)t0 = Ad(lk)t = t. It follows that lk ∈ M ∗ , so that there
exists u ∈ NK0 (t) such that Ad(lk)H = Ad(u)H, for all H ∈ t. It follows that
u−1 lk ∈ M , so that
k = (l−1 u)(u−1 lk)
where l−1 u ∈ K0 , as claimed.
We first generalize (16) to T , the connected subgroup of U with Lie algebra t.
Proposition 3.2. Let k ∈ K and h ∈ T be such that khk −1 ∈ T , then
khk −1 = whw−1
for some w ∈ W .
12 L. SECO AND M. PATRÃO

Proof: Let khk −1 = h0 ∈ T . We have that

T ⊆ ZU (h0 )0 = kZU (h)0 k −1 ⊇ kT k −1 .
Furthermore, ZU (h0 )0 is σ-invariant. In fact, if u ∈ ZU (h0 ) then, since ZU (h0 ) =
ZU (h0−1 ), we have that
σ(u)h0 = σ(u)σ(h0−1 ) = σ(uh0−1 ) = σ(h0−1 u) = σ(h0−1 )σ(u) = h0 σ(u).
The claim now follows since ZU (h0 ) ⊇ σ(ZU (h0 )0 ) is connected and contains the
indentity. Thus σ restricts to an involution σ 0 of ZU (h0 )0 which, intersected with
fix(σ)0 ⊆ K ⊆ fix(σ), furnishes
fix(σ 0 )0 ⊆ K ∩ ZU (h0 )0 ⊆ fix(σ 0 )
where ZK (h0 )0 is contained in the middle subgroup. It follows that (ZU (h0 )0 , ZK (h0 )0 )
is a symmetric pair, where T and uT u−1 are maximal abelian symmetric in ZU (h0 )0 .
Thus, there exists l ∈ ZK (h0 )0 such that
lkT k −1 l−1 = T
so that lk ∈ NK (T ) and thus w = lkM ∈ W satisfies
whw−1 = lkhk −1 l−1 = lh0 l−1 = h0
which proves the item.

3.1. Lifting to the universal cover of U . We now lift some results from the com-
pact connected Lie group U to its universal cover π
e:Ue → U , which is noncompact
whenever U is not semisimple.
Recall from (10) that the Lie algebra u of U e splits as a direct sum of ideals
u = g ⊕ z(u), where g is compact semi-simple and z(u) the center of u. It follows
e = G × Z where G is the connected subgroup with Lie algebra
the direct product U
g, which is compact by a Theorem of Weyl (see Theorem V.1.3 of [Bor98]), and Z
is a connected and central vector subgroup (see Corollary V.1.4 of [Bor98]).
Proposition 3.3. If σe is an automorphism of U
e , then the set fix(e
σ ) of its fixed
points is connected.
Proof: We have that σ e (G) = G and that σ e (Z) = Z, since these subgroups
are characteristic. It follows that σ e = σ e|G × σe|Z , which implies that fix(e σ) =
σ |G ) × fix (e
fix (e σ |Z ), which is connected, since fix (e
σ |Z ) is obviously connected and
since fix (e σ |G ) is also connected, by a Theorem of Borel (see Theorem V.3.3 of
[Bor98]).
Recall from (12) that
(29) u = (k ∩ g) ⊕ (s ∩ g) ⊕ (k ∩ z(u)) ⊕ (s ∩ z(u))
Proposition 3.4. We can always assume that k ⊆ g and that z(u) ⊆ s.
Proof: We have that
u = (k ∩ g) ⊕ (s ∩ g) ⊕ (s ∩ z(u))
is isomorphic to u/(k ∩ z(u)) and denote by σ the restriction of the involution σ to
u, and denote by an overline the image of the projection u → u parallel to k ∩ z(u).
Thus
g = g, z(u) = s ∩ z(u), k = k ∩ g, s=s
 COUNTING GEODESICS ON COMPACT SYMMETRIC SPACES 13

and hence
u = g ⊕ z(u) = k ⊕ s
so that
z(u) = s ∩ z(u) = z(u)
and
(30) fix(σ) = k ∩ g = k = fix(σ)
Now denote by Z(U ) the center of U , let
U = U/(K ∩ Z(U ))
σ the induced involution by σ on U , and denote by an overline the image of the
projection U → U . We thus get the subgroups
K ⊆ fix(σ) ⊆ fix(σ)
which by (30) have the same Lie algebra. It follows that U /K is a compact Rie-
mannian symmetric space such that
U U
'
K K
and such that k ⊆ g and z(u) ⊆ s, as claimed.
Let Te be the connected subgroup of U e with Lie algebra t. Denote by σ
e the lift
e◦σ
of the involution σ of U to U so that π
e e =σ◦π e.

 3.5. Assume that k ⊆ g, and that z(u) ⊆ s. Then we have that

Proposition

T = T ∩ G × Z, that K
e e σ ) ⊆ G, and we have the polar decomposition
e = fix(e

U
e =K
e TeK
e

Proof: Itis immediate

 that z(u) ⊆ t, which implies that t = (t ∩ g) ⊕ z(u) and thus
that Te = Te ∩ G × Z. Since U e → U is a covering, the Lie algebra of K e = fix(e
σ)
e ⊆ G, since K
is given by k, which implies that K e is connected by Proposition 3.3.
Finally, since G is compact and semi-simple and since t ∩ g ⊆ s ∩ g is maximal
abelian, it follows that  
G=K e Te ∩ G K e

e → G be the natural projection homomorphism. It follows that π(K)

Let π : U e = K,
e
that π(T ) = T ∩ G, and thus that
e e
 
π(Ue) = π K e TeK
e

This implies that, given u ∈ U

e , there exist k, l ∈ K,
e h ∈ Te, and z ∈ Z such that
u = khlz. Hence u = khzl ∈ K T K, since Z ⊆ T .
e e e e
We finish this section generalizing (16) to Te. Let W
f = N e (Te).
K

k∈K
Proposition 3.6. Let e h ∈ Te be such that e
e and e ke
hek −1 ∈ Te, then
ke
ehek −1 = w
ee e−1
hw
e∈W
for some w f . Furthermore, π
e(Wf) ⊆ W .
14 L. SECO AND M. PATRÃO

Proof: Denote by Te0 = Te∩G. By the previous proposition, we have that Te = Te0 ×Z
and Ke ⊆ G. Restricting σ e to G, we get the compact Riemannian symmetric pair
e with maximal torus Te0 .
(G, K)
Let e
ke
hek −1 = e
t ∈ Te. Decompose eh=e h0 z, e t0 w, where e
t=e h0 , e
t0 ∈ Te0 and z, w ∈ Z.
Since Z is contained in the center of U we get that
e
h0 e
ke
e k −1 z = e t0 w
e = G × Z, it follows that e
Since U h0 e
ke k −1 = et0 . Applying Proposition 3.2 to (G, K)
e
we get we∈W
f such that e h0 e
ke k −1 = w h0 w
ee e−1 . It follows that
ke
e k −1 = e
he keh0 e
k −1 z = weeh0 w
e−1 z = w h0 z w
ee e−1 = w
ee e−1
hw
as claimed. Furthermore, since π e ⊆ K and π
e(K) e(NKe (Te)) ⊆
e(Te) = T , it follows that π
NK (T ).
3.2. Lifting to the Klein universal cover of U/K. We now realize the Klein
universal cover from Ue to S: a good understanding of this covering allows us to
transfer some results from U/K to U and vice versa. For U compact semisimple
this is standard material (see Chapter V of [Bor98]) and here we extend this for U
compact.
Note that the Cartan covering (28) factors through
(31) η : U/K → U, uK 7→ uu∗ (u ∈ U ),
and the natural homomorphism from to U to the isometry group I(U/K), where
now u∗ is defined in terms of the involution σ of U . This map is well defined by
(26) and is automatically U -equivariant, where on the right hand side U acts on
itself by the action (27). By the same reasoning as in Section 2, (31) is a covering
onto its image, with fiber fix(σ)/K, which we will also call the Cartan covering.
Denote by π e:U e → U the universal cover. Let K
e = fix(e
σ ) which is connected by
Proposition 3.3, so that
(32) π e = fix(σ)0 ⊆ K.
e(K)
and Ue /K
e is simply connected, by the homotopy exact sequence. It follows that
e : U → U descends to the universal cover of U/K, which we denote by the same
π e
symbol
(33) π
e:U e → U/K,
e /K uKe 7→ πe(u)K (u ∈ Ue ).

Since K
e = fix(eσ ), the Cartan covering (31) becomes the embedding
ηe : U
e /Ke →U e , uK 7 uu∗ (u ∈ U
e → e)

Denoting by Se the image of ηe in U e , we have the following commutative diagram of

universal covers
η
Ue /K Se ⊆ U
e
e e
∼

π
e λ
S
which defines the Klein universal cover λ.
We now furnish a group theoretic description of the fibers of these coverings. Let
Te be the connected subgroup of Ue with Lie algebra t and let
Fe = Te ∩ Z(U
e)
 COUNTING GEODESICS ON COMPACT SYMMETRIC SPACES 15

Below we adapt ideas from Sections 6.10 of [HI12] and Theorem V.4.5 of [Bor98]
to prove the following.
Theorem 3.7. We have that ηe restricts to a group isomorphism
(34) ηe : NUe (K)/
e K e → Fe
where the corresponding groups act effectively by isometries on the right of the
respective universal cover.
Furthermore, ηe restricts to an isomorphism between the corresponding fibers
e−1 (p), λ−1 (p), which are subgroups isomorphic to π1 (U/K).
π
Proof: Let πe:U e−1 (K) ⊆ NUe (K).
e → U , we claim that π e(u) ∈ K.
e Indeed, let π
By (32) we have that
π e −1 ) ⊆ π
e(uKu π (u)−1 = K,
e(u)Ke
e −1 is connected with the same Lie algebra as K,
so that uKu e −1 = K,
e thus uKu e as
−1
e : U /K → U/K, it follows that π
claimed. Going back to π e e e (p) 6 NUe (K)/K. By
e e
the commutativity of the diagram, we have that
(35) λ−1 (p) = ηe(e
π −1 (p)) ⊆ ηe(NUe (K)/
e K).
e
Now we prove that
(36) e = (Te ∩ N e (K))
NUe (K) e K,e
U

where
(37) e = {h ∈ Te : h2 ∈ Fe}.
Te ∩ NUe (K)
For (36), let u ∈ NUe (K)
e and consider the polar decomposition u = khl, where
k, l ∈ K
e and h ∈ Te, given by Proposition 3.5. Since u, k, l normalize K,
e it follows
e Furthermore, k −1 u = hl, where k −1 u = um, for m ∈ K,
that h ∈ Te ∩ NUe (K). e so
−1 −1
that u = h(lm ), with lm ∈ K, as claimed. For (37), first note that
e

(38) Te ∩ ZUe (K)

e = Te ∩ Z(U
e ),

which follows from the polar decomposition of u = khl ∈ U h ∈ Te ∩ ZUe (K)

e , since e e
−1
commutes with h, k and l. Now let h ∈ T and k ∈ K, then h kh ∈ K iff
e e e
h−1 kh = σ
e(h−1 kh) = hkh−1
iff h2 kh−2 = k. It follows that
e = {h ∈ Te : h2 ∈ Te ∩ Z e (K)}
Te ∩ NUe (K) e
U

and the claim follows from (38).

For h ∈ Te we have that ηe(hK)e = h2 , so that (36) and (37) imply ηe(N e (K)/
e K)e ⊆
U
Fe. Since elements in Te have square roots in Te, equality follows from (36) and (37).
It follows that (34) is a bijection. To prove that it is a group homomorphism, let
u∈U e and n ∈ N e (K).
U
e Since ηe(nK) e ∈ Fe ⊆ Z(U e ), it follows by U
e -equivariance
that
ηe(unK)
e = ue e ∗ = uu∗ ηe(nK)
η (nK)u e = ηe(uK)ee η (nK),
e
which implies the claim. From (35) it then follows that λ−1 (p) 6 Fe so that ηe
restricts to a group isomorphism between π e−1 (p) and λ−1 (p), as claimed.
16 L. SECO AND M. PATRÃO

By the homotopy exact sequence π1 (U e /K)

e is trivial, since K
e is connected. It
follows that both πe and λ are coverings, so that π1 (U/K) is isomorphic to both
e−1 (p) and λ−1 (p).
π
Remark 3.8. Note that h ∈ Te ∩ K e iff h∗ = h−1 = h, iff h2 = 1. Thus, a square
root a ∈ T of h ∈ T , that is, such that h = a2 , is defined mod K.
e e e Indeed, if
√
2 2 −1 2
h = a = b , for a, b ∈ Te, then (ab ) = 1 so that a = b mod K. e Denote h = a
mod K,
e then by (36) and (37) we have the isomorphism
p
Fe → Te ∩ NUe (K)/
e K, e f 7→ f ,
which is the inverse to ηe, through which Fe acts in the universal cover U
e /K.
e

3.3. Projecting back to U/K. We finish this section by extending (16) to the
Weyl group W of a compact symmetric space U/K.
Theorem 3.9. Let M = U/K be a symmetric space with U compact and basepoint
p = K. If k ∈ K and h ∈ T are such that khp ∈ T p, then
khp = whp
for some w ∈ W in the Weyl group of U/K.
Proof: Let khp = h0 p, for some h0 ∈ T . By Proposition 3.1, since M ⊆ K
centralizes T , we can assume w.l.o.g that k ∈ K0 . We now use the notation of and
results of Subsection 3.1. Since K
e is connected, we have that π e(K)
e = K0 . Let then
0
k ∈ K, h ∈ T , h ∈ T such that
e e e e e e
π
e(e
k) = k, π e(e
h) = h, π e(eh0 ) = h0 .
Denote the basepoint of U
e /K
e by pe = K,
e since

π
e(e
ke
hep) = khp = h0 p = π h0 pe)
e(e
it follows that e
kehe
p and eh0 pe in Ue /K
e are in the same fiber of π e. By Theorem 3.7,
their image by ηe are in the same fiber of the Klein universal cover λ, so that there
exists f ∈ Fe ⊆ Te such that
h2 e
ke
e k −1 = ηe(e
ke
hep) = ηe(h0 pe)f = h02 f
where h02 f ∈ Te. By Proposition 3.6, there exists w
e∈W
f , such that
ke
e k −1 = w
h2 e e−1
h2 w
ee
Thus
ηe(e
ke
hep) = ηe(w
eehe
p) =⇒ eke
hep=w
eehe
p,
since ηe is an embedding. Projecting back to U/K we get that khp = whp, for
w=π e ∈ W , as claimed.
e(w)

4. Inverse image of the exponential map

In this section we will describe the inverse image of the exponential map expp :
s → S given by (4), as a disjoint union of focal orbits passing through t, and will
count their dimensions and connected components by inspecting data in t.
Given H ∈ s, recall that by (15) we can assume w.l.o.g. that H ∈ t. Let
h = exp(H) and q = hp ∈ S. Recall the focal orbit F (H) = K q H given by (1).
Recall the lattice Γ given by (2) and the centralizer W q of H mod Γ given by (3).
 COUNTING GEODESICS ON COMPACT SYMMETRIC SPACES 17

Theorem 4.1. Given q = exp(H)p, we have that

[
exp−1
p (q) = F (H + γ)
γ∈Γ

where two such focal orbits intersect, and then coincide, if and only of their respec-
tive H + γ lie in the same W q -orbit.
Proof: First recall that h = exp(H) and note that the whole focal orbit
F (H + γ)
exponentiates to hp since, for k ∈ K q , we have
exp(Ad(k)(H + γ))p = k exp(H + γ)k −1 p = k exp(H)p = kq = q
Reciprocally, let exp(X)p = hp. There exists k ∈ K such that Ad(k)X ∈ t so that
exp(Ad(k)X)p = khp ∈ T p. By Theorem 3.9, we have that khp can be realized by
khp = whp
for some w ∈ W , which we can consider belonging to K. It follows that k −1 w ∈ K q .
In the above equation, replacing hp by exp(X)p in the left hand side and by exp(H)p
in the right hand side, we get that
exp(Ad(k)X)p = exp(wH)p
Since both Ad(k)X and wH belong to t, it follows that
Ad(k)X = wH + γ 0 = w(H + γ)
for γ = w−1 γ 0 ∈ Γ, since Γ is W -invariant. Then
X = Ad(k −1 )w(H + γ) = Ad(k −1 w)(H + γ)
where we have seen that k −1 w ∈ K q . This proves thet X is in the focal orbit
F (H + γ), as claimed.
For the second part, if F (H + γ) and F (H + γ 0 ) intersect, then H + γ 0 =
Ad(u)(H + γ) for some u ∈ K q . By (16), it follows that Ad(u)(H + γ) = w(H + γ)
for some w ∈ W , so that
H + γ 0 = w(H + γ) = wH + γ 00
where γ 00 = wγ ∈ Γ, since Γ is W -invariant. Thus we get
wH = H + γ 0 − γ 00 ,
from which it follows that w ∈ W q , as claimed. Reciprocally, let H + γ 0 = w(H + γ)
for some w ∈ W q . Since w can be realized by u ∈ K, we have that H + γ 0 =
Ad(u)(H + γ). We get that q = hp = uhp = uq and then u ∈ K q . It follows that
H + γ 0 ∈ Ad(K q )(H + γ) = F (H + γ), as claimed.
Consider the centralizers KH of H and Kh of h, so that
(39) KH ⊆ Kh ⊆ K q
Note that F (H) is diffeomorphic to K q /KH , where K q fixes the endpoints p and
q while KH fixes pointwise the geodesic expp (tH), t ∈ [0, 1]. Recall from Equation
(9) the Lie algebras of the centralizers
X X
(40) kH = m ⊕ kα kh = m ⊕ kα
α(H)=0 α(H)∈2πZ
18 L. SECO AND M. PATRÃO

where α ranges through the positive restricted roots. Since the Cartan covering
(31) takes kq = q to kh2 k −1 = h2 , it follows that the isotropy K q is contained in
Kh2 with the same Lie algebra
X
(41) kq = kh2 = m ⊕ kα
α(H)∈πZ

Note that the sum runs through the diagram hyperplanes that cross H.
Recall that elements of the Weyl group W are represented by the centralizer M∗
of t in K, that is, W = Ad(M∗ )|t . Let
M∗q = M∗ ∩ K q
The next result shows that this subgroup intersects each connected component of
K q . Recall the subgroup W0q generated by reflections around root hyperplanes
paralell to the diagrams hyperplanes which cross H.
Proposition 4.2. K q = K0q M∗q where
W q = Ad(M∗q )|t and W0q = Ad(K0q ∩ M∗q )|t
0
Proof: For the fisrt statement, let q 0 = exp(−H)p, we claim that K q ⊆ K q .
Indeed, let u ∈ K be such that uq = q = exp(H)p. Applying the reflection of U/K
at p we get
s(u exp(H)p) = σ(u) exp(−H)p = uq 0 = s(exp(H)p) = exp(−H)p = q 0
0 0
Analogously we have K q ⊇ K q , and thus K q = K q . Thus, by the covering
η : U/K → U (31), it follows that K q is contained in
U 0 = centralizer of η(q) = h2 and η(q 0 ) = h−2 in U
which is clearly closed and σ-invariant. Its Lie algebra then splits as u0 = k0 ⊕ s0 ,
where k0 = u0 ∩ k and s0 = u0 ∩ s. It follows that k0 = kh±2 = kh2 = kq , the Lie algebra
kq of K q . Clearly, s0 contains t. Let u ∈ K q , then Ad(u)t ⊆ s0 is a maximal torus.
By Proposition 3.1 applied to the symmetric pair (U 0 , K q ), there exists v ∈ K0q such
that Ad(vu)t = t. Thus vu = w ∈ K q ∩ M∗ = M∗q and u = v −1 w, with v ∈ K0q ,
which proves the first statement.
For the second statement, let w ∈ W q so that wH = H + γ, γ ∈ Γ. Let v ∈ M∗
represent w. Thus, vq = v exp(H)v −1 p = exp(wH)p = exp(H + γ)p = exp(H)p =
q, so that v ∈ K q . Reciprocally, for v ∈ M∗q we have exp(Ad(v)H)p = v exp(H)p =
vq = q = exp(H)p. Thus, Ad(v)H = H + γ, γ ∈ Γ, so that Ad(v) ∈ W q , as
claimed.
For the third statement, since u0 = uh±2 it follows from (6) that
X
(u0 )C = u00 ⊕ {u0α : α(2H) ∈ 2πZ}

Since t ⊆ s0 , it follows that the restricted roots of (u0 , k0 ) are Π0 = {α ∈ Π : α(H) ∈

πZ}, so that the Weyl group of (u0 , k0 ) is W0q . On the other hand, since k0 = kq , it
follows that the Weyl group of (u0 , k0 ) is given by Ad of K0q ∩M∗ = (K0q ∩K q )∩M∗ =
K0q ∩ M∗q , as claimed.
The dimension and connected components of a focal orbit F (H) = K q H can be
read in t as follows.
 COUNTING GEODESICS ON COMPACT SYMMETRIC SPACES 19

Theorem 4.3. F (H) intersects t in the orbit W q H and its connected components
are K0q wH, w ∈ W q . They are in bijection with the quotient group W q /W0q where
each connected component corresponds to a left W0q -orbit in W q H and is diffeomor-
phic to K0q H, which has dimension
X
(42) dim F (H) = mα
where mα is the multiplicity of the root α and the sum runs through the non-root
hyperplanes of the Stiefel diagram that cross H.
Proof: For the first statement, let Ad(k)H ∈ t, for k ∈ K q . By (16) there exists
w ∈ W such that Ad(k)H = wH: since the left hand side exponentiates to q =
expp (H), it follows that expp (wH) = expp (H). Thus, there exists γ ∈ Γ such that
wH = H + γ, so that w ∈ W q , as claimed.
For the second statement, note that the centralizer WH is generated by reflections
around root hyperplanes that cross H, so that WH ⊆ W0q . It follows from the
previous Theorem that the connected components of F (H) are K0q wH, w ∈ W q .
Thus they are in bijection with the double coset W0q \W q /WH . Since W0q is normal,
it follows that
W0q wWH = wW0q WH = wW0q
so that the double coset is in bijection with W q /W0q . For the last statement,
let w ∈ W q be represente by v ∈ M∗q . Since conjugation preserves the identity
component, it follows that each connected component of F (H) satisfies
K0q wH = v(v −1 K0q vH) = v(K0q H)
since conjugation by v leaves invariant the connected component of the identity,
thus is diffeomorphic to K0q H, as claimed. Their dimension is then dim kq /kH ,
which by (40) and (41) is given by
X
dim kq /kH = mα
α(H)∈πZ−0

as claimed.
Remark 4.4. In the compact group case [SS18], the isotropy KH of F (H) is
connected, so that the connected components of F (H) are in bijection with those
of K q . This is not the case in general since K itself can be disconnected. What
holds in general is the bijection between the connected components of F (H) and
the quotient group W q /W0q .

5. Fundamental group
In this section we generalize to compact symmetric spaces a well known descrip-
tion of the fundamental group of compact type symmetric spaces as a quotient of
lattices: this is not immediate from the compact type case since the lattices in t do
not need to be adapted to the splitting into center and its complement. We then
use this result to investigate the connected components of K q .
Let treg be the complement in t of the Stiefel diagram and denote by Sreg =
K expp (treg ) be the regular elements of S. Note that the left action by W and
translation by Γ leaves treg invariant (see Proposition 2.2), thus so does the affine
action of the semidirect product Γ o W . Also, W acts on the right of K/M : it is
the right action of W = M∗ /M on the principal bundle K/M → K/M∗ . It follows
20 L. SECO AND M. PATRÃO

that Γ o W acts in the right of treg × K/M by the right W -action on K/M and by
the affine action on treg :
(H, kM ) · (γ, w) = (γ + w−1 H, kwM )
This action is free since (γ + w−1 H, kwM ) = (H, kM ) implies wM = M , so that
w ∈ M , thus w−1 H = H, which implies γ + H = H so that γ = 0.
Proposition 5.1. The map
Ψ : treg × K/M → Sreg , (H, kM ) 7→ k expp (H)
is the quotient map of the right Γ o W -action. In particular, Ψ is a covering of
Sreg with fiber Γ o W .
Proof: Clearly, Ψ(γ + w−1 H, kwM ) = Ψ(H, kM ). Reciprocally, let Ψ(H, kM ) =
Ψ(H 0 , k 0 M ), with H, H 0 ∈ treg . Then k expp (H) = k 0 expp (H 0 ) so that

q = expp (H) = expp (Ad(k −1 k 0 )H 0 )

where q ∈ Sreg . By Theorem 4.1 it follows that
Ad(k −1 k 0 )H 0 = Ad(u)(H + γ)
for some γ ∈ Γ and u ∈ K q . Since q ∈ Sreg , it follows from equation (41) that
kq = m, so that (K q )0 = M0 . By Proposition 4.2 it then follows that
K q = M0 M∗q
where M∗q acts in t as Wq while M centralizes t, so that
Ad(u)(H + γ) = H + γ 0
for some γ 0 ∈ Γ, by the definition W q . By two previous equations it follows that
Ad(k −1 k 0 )H 0 = H + γ 0
so that by (16) we have Ad(k −1 k 0 )H 0 = wH 0 , for some w ∈ W . Then
wH 0 = H + γ 0 =⇒ H 0 = γ 00 + w−1 H
where γ 00 = w−1 γ 0 ∈ Γ. Now Ad(w−1 k −1 k 0 )H 0 = H 0 with H 0 regular implies that
w−1 k −1 k 0 ∈ M , so that k 0 M = kwM , and then
(H 0 , k 0 M ) = (H, kM ) · (γ 00 , w)
as claimed.
The determination of the fundamental group of compact symmetric spaces fol-
lows using the same strategy which is used in the proof of the analogous result for
compact Lie groups in Chapter 13 of [Hal15] apart from the first and, specially, the
last item, which we addressed above.
Theorem 5.2. Is S is a compact symmetric space, then
π1 (S) = Γ/Γ0
Proof: The proof is given by the following items.
 COUNTING GEODESICS ON COMPACT SYMMETRIC SPACES 21

1) By a standard curve-shortening process, one shows that any nontrivial class in

π1 (S) can be represented by a closed geodesic through p (see Lemma 8.3.10 of
[Wo11]), say γ(t) = expp (tX), X ∈ s, with γ(1) = expp (X) = p. By (15) there
exists k ∈ K0 such that kX = H ∈ t. Note that expp (H) = k expp (X) = kp = p,
thus H ∈ Γ. Since K0 is connected, it follows that γ(t) is homotopic the the
closed geodesic expp (tH) on T p. Thus, the inclusion
π1 (T p) → π1 (S)
is surjective. Note that [Hal15] uses the simply connectivity of the full flag
manifold of K to prove this (Proposition 13.37 of [Hal15]), but here the corre-
sponding object is the full flag manifold K/M of S, which is not always simply
connected.
2) Since T p is a compact abelian Lie group, expp : t → T p is a homomorphism with
kernel Γ, which induces an isomorphism t/Γ → T p. It follows that the inclusion
of the closed geodesics Γ → π1 (T p) induces an isomorphism.
3) By the standard construction given in the proof of Proposition 2.2 we can realize
each closed geodesic of S corresponding to a coroot vector as the equator of a
corresponding 3-sphere immersed in U and acting in S. By a standard argument
(entirely analogous to Proof of Theorem 13.17, one direction, p.382 of [Hal15]),
it follows that every closed geodesic that comes from the fundamental lattice Γ0
is nullhomotopic in S.
4) By a standard dimension argument (entirely analogous to the Proof of Theorem
13.27, p.386 of [Hal15]) it follows that the fundamental group of S equals the
fundamental group of Sreg .
5) The proof that every loop that does not come from Γ0 is not nullhomotopic
(Proof of Theorem 13.17, the other direction, p.398 of [Hal15]) depends on the
construction of a covering of Sreg . [Hal15] obtains the universal covering, but
a covering is enough for the proof. This construction is the only step whose
proof fully departs from the compact group case of [Hal15] and it is provided by
Proposition 5.1.

Corollary 5.3. If S is simply connected then

Γ = Γ0 D = D0 W q = W0q
In particular, all focal orbits are connected.
Proof: Since S is simply connected, Theorem 5.2 gives us that Γ = Γ0 . Theorem
2.6 then implies that D = D0 and W q = W0q . The connectedness of the focal orbits
then follows from Theorem 4.3.
The connectedness of focal orbits for a simply connected S also follow from the
next result.
Proposition 5.4. Each connected component of a focal orbit of corresponds to a
distinct homotopy class of curves between p and q.
Proof: Fixing the geodesic γH (t) = expp (tH) from p to q = expp (H), H ∈ t, gives
us the map
−1
πp,q (S) → π1 (S), [γ] 7→ [γ ∗ γH ]
between the homotopy classes of curves between p and q and the fundamental group
of loops at p. This map is a bijection with inverse π1 (S) → πp,q (S), [γ] 7→ [γ ∗ γH ].
22 L. SECO AND M. PATRÃO

Restricting this bijection to the geodesics corresponding to a connected component

of a focal orbit F (H) we get the injection
π0 (F (H)) = W q /W0q ,→ π1 (S) w mod W0q 7→ (wH − H) mod Γ0
which proves the result.

6. Conjugate and cut loci

Crittenden [Cri62] and Sakai [Sak77] characterized, respectively, the first conju-
gate locus and the cut locus of compact symmetric spaces showing that they are
determined by a maximal flat torus. In this section we use our results to reobtain
their results, with short and independent proofs.
Recall that a tangent vector X ∈ Tp S is a (first tangent) conjugate point at p
when tX has null index for 0 ≤ t < 1 and positive index for t = 1, where the index
of Y ∈ Tp S along the geodesic cY (t) = expp (tY ) is the dimension of the kernel of
the differential dY expp : Tp S → Texpp (Y ) S. The (first tangent) conjugate locus at
p is the set of first tangent conjugate points in Tp S. Conjugate points are related
to variations by geodesic as follows: X ∈ Tp S is a tangent conjugate point at p
if and only if there exists a nontrivial variation of geodesics around cX with fixed
start point p at t = 0 and, up to first order, fixed end point q = expp (X) at t = 1.
Now, recall that a tangent vector X ∈ Tp S is a (tangent) cut point at p when the
geodesic cX (t) = expp (tX) is minimizing for 0 ≤ t ≤ 1 and ceases to be minimizing
for t > 1. The (tangent) cut locus at p is the set of tangent cut points in Tp S.
The following result is well known and shows that tangent conjugate points of a
compact symmetric space S = U/K are intimately related to the focal orbits that
appear in Theorem 4.1.
Lemma 6.1. The index of H ∈ t along the geodesic cH (t) = expp (tH) equals
dim F (H). Furthermore, H is a tangent conjugate point of the geodesic cH if, and
only if, there exists a nontrivial variation of geodesics around cH with fixed start
point p and fixed end point q = expp (H). Note that such a variation corresponds
to a curve in F (H) that passes through H.
Proof: By Theorems IV.4.1 and Lemma VII.2.9 of [Hel78], the eigenvalues of
dH expp (H) are given by sin(α(H))/α(H) with multiplicity dim sα = mα , for
α ∈ Π+ , thus is zero iff α(H) ∈ πZ − 0. This, together with equation (42) proves
the first part.
For the the second part, if H is tangent conjugate then, by the previous para-
graph, dim F (H) > 0, so that there exists a nontrivial variation of geodesics around
cH with fixed endpoints p and q, which corresponds to a nontrivial curve in F (H)
passing through H. The reciprocal is immediate from the second equivalent defini-
tion of conjugate point.
For the cut points of a compact symmetric space S = U/K, we first investigate
the submanifold geometry of a maximal flat torus T p. It is immediate that T p is
totally geodesic in S, moreover, we have the following.
Proposition 6.2. T p is geodesically convex in S, that is, the Riemannian distance
of S induced in T p coincides with its intrinsic Riemannian distance.
Proof: Denote by dS , dT the Riemmanian distances of S and T p, respectively.
For x, y ∈ T p it is immediate that dS (x, y) ≤ dT (x, y). For the converse inequality,
 COUNTING GEODESICS ON COMPACT SYMMETRIC SPACES 23

since T acts transitively by isometries on T and S, we can assume that x = hp,

y = p, where h ∈ T . Let X of minimal norm in s such that hp = expp (X), then
dS (hp, p) = |X|, since expp (tX), 0 ≤ t ≤ 1, is a geodesic of length |X| between p
and expp (X) in S. Now let H of minimal norm in t such that hp = expp (H), then
dT (p, hp) = |H|, arguing as before, now in T p. Since expp (X) = expp (H) = hp,
it follows from Theorem 4.1 that X ∈ F (H + γ), for some γ ∈ Γ. Hence |X| =
|H + γ| ≥ |H|, by the minimality of |H|. Thus dS (hp, p) ≥ dT (hp, p), as claimed.

Consider the Dirichilet domain D of the lattice Γ and the Dirichilet domain D0
of the fundamental lattice Γ0 (see Section 2.3).
Theorem 6.3. Let S = U/K be a compact symmetric space, then
i) (Sakai) The cut locus of S is the K orbit of ∂D.
ii) (Crittenden) The conjugate locus of S is the K orbit of ∂D0 .
Furthermore, S is simply connected iff its cut locus equals its first conjugate
locus.
Proof: We will use freely the results of Proposition 2.5. We first claim that the
cut locus of T p is the boundary of D. Indeed, recall that D is the set of vectors of
t with minimal norm among its Γ equivalents. Let H be in the tangent cut locus of
T , then tH ∈ D for 0 ≤ t ≤ 1, and tH 6∈ D for t > 1 so that H ∈ ∂D. Reciprocally,
let H be in the boundary of D. Since D is a convex neighbourhood of the origin,
it follows that tH ∈ D is minimal for 0 ≤ t ≤ 1. For t = 1 we have H ∈ ∂D so that
there exists γ 6= 0 in Γ such that |H + γ| = |H|. Thus Y = H + γ is a nontrivial
focal equivalent to H and by Lemma 2.4 it follows that (1 + )H has norm greater
than Y +H, for  > 0, while exp((1 +)H) = exp(Y +H). It follows that exp(tH)
is not a minimizing geodesic for t > 1, which proves our claim.
For item (i), from Proposition 6.2 it follows that the cut locus of T p is contained
in the cut locus of S, and so its adjoint orbit, since Ad(K) acts by isometries in
s. By using that Ad(K) conjugates every X ∈ s to some H ∈ t, it follows from
Proposition 6.2 that the cut locus of S is contained in the adjoint orbit of the cut
locus of T p, which proves item (i).
We now claim that the conjugate locus of T p is the boundary of D0 . Indeed,
recall that D0 = {H ∈ t : α(H) ≤ π, α ∈ Π}. Thus its boundary ∂D0 is contained
in the union of hyperplanes α(H) = π, α ∈ Π. The claim now follows since D0
consists of regular elements and D0 is a convex neighbourhood of the origin so
that, by Lemma 6.1 the first conjugate point of any ray tH occurs when it hits the
boundary of D0 .
For item (ii), it is obvious that the conjugate locus of T p is contained in the
tangent conjugate locus of S, and so its adjoint orbit, since Ad(K) acts by isometries
in s. By using that Ad(K) conjugates every X ∈ s to some H ∈ t and that Ad(K)
acts by isometries, it follows that the tangent conjugate locus of S is contained in
the Ad(K) orbit of the tangent conjugate locus of T , which proves the first claim.
For the second claim, by the previous items, the cut locus of S equals its first
conjugate locus iff the cut locus of T p equals its first conjugate locus iff D = D0
since both are open convex sets whose closure is the convex closure of its boundary
points. If Γ = Γ0 then it is immediate that D = D0 . Reciprocally, if D = D0 we
claim that
min |H + γ0 | = min |H + γ|
γ0 ∈Γ0 γ∈Γ
24 L. SECO AND M. PATRÃO

Indeed, we have that minγ0 ∈Γ0 |H + γ0 | is attained at a H + γ0 ∈ D 0 , for some

γ0 ∈ Γ0 . Since D0 = D, we have that |H + γ0 | equals minγ∈Γ |(H + γ0 ) + γ|. Since
Γ0 ⊆ Γ, this equals minγ∈Γ |H + γ|, which proves the claim. Taking H = γ ∈ Γ it
follows that minγ0 ∈Γ0 |γ + γ0 | = 0 so that γ ∈ Γ0 and thus Γ ⊆ Γ0 so that Γ = Γ0 .
By Theorem 5.2, S is simply connected iff Γ = Γ0 , which proves our claim.

References
[Bor98] A. Borel: Semisimple Groups and Riemannian Symmetric Spaces, Text and Readings in
Math. 16, Hindustan Book Agency, India, 1998.
[Bou02] N. Bourbaki: Elements of Mathematics. Lie Groups and Lie Algebras, Springer-Verlag,
New York, 2002.
[Bot59] R. Bott: The Stable Homotopy of the Classical Groups, Ann. Math., Second Series, 70(2)
(1959), 313-337.
[Cri62] R. Crittenden: Minimum and conjugate points in symmetric spaces Canad. J. Math. 14
(1962), 320-328.
[Esch] J.-H. Eschenburg: Lecture Notes on Symmetric Spaces, https://myweb.rz.uni-
augsburg.de/ eschenbu/symspace.pdf
[Fer80] D. Ferus: Symmetric Submanifolds of Euclidean Space, Math. Ann. 247 (1980), 81-93.
[Hal15] B. Hall: Lie Groups, Lie Algebras, and Representations: An Elementary Introduction,
Springer, 2nd Edition, 2015.
[Hel78] S. Helgason: Differential Geometry, Lie Groups and Symmetric Spaces, Academic Press,
1978.
[HN12] J. Hilgert, K. Neeb: Structure and Geometry of Lie Groups, Springer, 2012.
[HI12] J. Holland, B. Ion: Notes on symmetric spaces, arXiv:1211.4159v1, 2012.
[Loo69] O. Loos: Symmetric Spaces II: Compact Spaces and Classification, W.A. Benjamin Inc.,
New York / Amsterdam, 1969.
[Sak77] T. Sakai: On cut loci of compact symmetric spaces, Hokkaido Math. J. 6 (1977), 136-161.
[SS18] L. Seco, L.A.B. San Martin: Counting geodesics on compact Lie groups, Diff. Geom.
Appl. 56 (2018) 1-19.
[Tak78] M. Takeuchi: On conjugate loci and cut loci of compact symmetric spaces I, Tsukuba J.
Math. 2 (1978), 35-68.
[Wo11] J.A. Wolf, Spaces of constant curvature, 6th ed., AMS Chelsea Publishing, Providence,
RI, 2011.
∗ Universidade de Brası́lia – Departamento de Matemática, Campus Darcy Ribeiro,

Asa Norte 70910-900 - Brası́lia, DF - Brazil

Email address: lseco@unb.br, mpatrao@mat.unb.br