Fundamental groups of split real Kac-Moody groups and generalized real flag manifolds. With appendices by Tobias Hartnick and Ralf K\"ohl and by Julius Gr\"uning and Ralf K\"ohl
Paula Harring, Ralf K\"ohl

TL;DR
This paper computes the fundamental groups of split real Kac-Moody groups with the Kac-Peterson topology by analyzing their flag varieties, extending known results from finite-dimensional cases to infinite-dimensional Kac-Moody groups.
Contribution
It determines the fundamental groups of symmetrizable split real Kac-Moody groups and their flag varieties, generalizing finite-dimensional results to infinite-dimensional Kac-Moody settings.
Findings
Fundamental groups of Kac-Moody groups are determined via flag varieties.
Results apply to all symmetrizable cases with CW decompositions.
The structure of fundamental groups extends to non-symmetrizable two-spherical cases.
Abstract
We determine the fundamental groups of symmetrizable algebraically simply connected split real Kac-Moody groups endowed with the Kac-Peterson topology. In analogy to the finite-dimensional situation, the Iwasawa decomposition provides a weak homotopy equivalence , implying . It thus suffices to determine which we achieve by investigating the fundamental groups of generalized flag varieties. Our results apply in all cases in which the Bruhat decomposition of the generalized flag variety is a CW decomposition in particular, we cover the complete symmetrizable situation; the result concerning the structure of more generally also holds in the non-symmetrizable two-spherical situation.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAlgebraic structures and combinatorial models · Homotopy and Cohomology in Algebraic Topology · Advanced Combinatorial Mathematics
Fundamental groups of split real Kac–Moody groups and generalized real flag manifolds
Paula Harring Ralf Köhl
With appendices
by Tobias Hartnick and Ralf Köhl and
by Julius Grüning and Ralf Köhl
Abstract
We determine the fundamental groups of symmetrizable algebraically simply connected split real Kac–Moody groups endowed with the Kac–Peterson topology. In analogy to the finite-dimensional situation, because of the Iwasawa decomposition the embedding is a weak homotopy equivalence, in particular . It thus suffices to determine which we achieve by investigating the fundamental groups of generalized flag varieties. Our results apply in all cases in which the Bruhat decomposition of the generalized flag variety is a CW decomposition – in particular, we cover the complete symmetrizable situation; furthermore, the results concerning only the structure of actually also hold in the non-symmetrizable two-spherical case.
1 Introduction
The structure of maximal compact subgroups in semisimple Lie groups was investigated by Cartan and, later, Mostow: In [Mos49], Mostow gives a new proof of a Cartan’s theorem stating that a connected semisimple Lie group is a topological product of a maximal compact subgroup and a Euclidean space, implying in particular that and have isomorphic fundamental groups. Subsequent case-by-case analysis provided the isomorphism types of these maximal compact subgroups – which in the split real situation turn out to be all classical – and their fundamental groups; tables of the maximal compact subgroups can be found in [Hel78, p 518], their fundamental groups in [SBG*+*95, 94.33].
Starting in the 1940’s, Dynkin diagrams, introduced in [Dyn46], have been used to describe the structure of simple Lie groups. In this article, we present a uniform result which makes it possible to determine the fundamental group of any algebraically simply connected split real simple Lie group – and, more generally, any algebraically simply-connected semisimple split real topological Kac–Moody group – directly from its Dynkin diagram.
In [Tit87, Theorem 1], Tits for every generalized Cartan matrix provides a functor from commutative rings into groups. Let be the Dynkin diagram of .
Definition 1.1.
We set and refer to this group as the algebraically simply-connected semisimple split real Kac–Moody group of type .
Kac–Moody groups endowed with the Kac–Peterson topology have been studied extensively by the second author together with Glöckner and Hartnick in [GGH10] and with Hartnick and Mars in [HKM13]. Our result is applicable to those Kac–Moody groups whose Bruhat decompositions are CW decompositions and for which the embedding is a weak homotopy equivalence.
In order to fix notations, let be the algebraically simply-connected split real semisimple Kac–Moody group associated to an irreducible diagram endowed with the Kac–Peterson topology (for definitions, see Section 2). Let be the so-called maximal compact subgroup of the topological group , i.e., the subgroup fixed by the Cartan–Chevalley involution of . We stress that in the infinite-dimensional non-Lie case this topological group is not a compact group, only a -group, in fact a -group.
Given the Dynkin diagram with a fixed labelling , we define a modified diagram with vertex set and edge if and only if , where denotes the parity of the corresponding Cartan matrix entry. To each connected component of we then assign a colour as follows: Let be coloured red (denoted by ) if it contains a vertex such that there exists a vertex satisfying and ; let be coloured green () if it is not red and consists only of an isolated vertex; and blue () else.
One can then read off the isomorphism type of from the coloured diagram as specified in the following theorem.
Theorem**.**
Let be an irreducible Dynkin diagram such that the Bruhat decomposition of provides a CW decomposition (i.e., such that the conclusion of Proposition 3.7 holds) and such that the embeding is a weak homotopy equivalence (i.e., such that the conclusion of Theorem A.15 holds). Let and be the number of connected components of of colour and , respectively. Then
[TABLE]
In particular, this statement holds in the symmetrizable case.
Example: 0 Isomorphism types of for the spherical Dynkin diagrams111Dynkin diagram LaTeX styles kindly provided by Max Horn at [Hor].
[TABLE]
Example: 0 Isomorphism types of for selected indefinite Dynkin diagrams222Dynkin diagram LaTeX styles kindly provided by Max Horn at [Hor].
[TABLE]
While in the classical finite-dimensional Lie case, one has a topological Iwasawa decomposition with and contractible, implying , it is currently unknown whether the corresponding Iwasawa decomposition in the general Kac–Moody case is also topological. However, using a fibration result by Palais (see Proposition A.13), in the appendix Hartnick and the second author prove that the isomorphism between the fundamental groups still exists in the general symmetrizable case, therefore reducing the problem to the computation of .
In [GHKW17, Section 16], the group – where denotes a so-called admissible colouring of the vertices of – is defined as the canonical universal enveloping group of a -amalgam where the isomorphism type of depends on the - and -entries of the Cartan matrix of as well as the values of on the corresponding vertices.
It is shown in [GHKW17, Section 17] that there exists a finite central extension which implies that the subspace topology on inherited from the Kac–Peterson topology on defines a unique topology on that turns the central extension into a covering map. The resulting group topology on is called the Kac–Peterson topology on .
In the simply-laced case, there is a unique non-trivial admissible colouring and the corresponding group double-covers as shown in [GHKW17]. We prove here that in the simply-laced case is simply connected which then implies that .
The strategy of proof in the simply-laced case is to study fibre bundles of the form
[TABLE]
arising from embeddings of along subdiagrams of type , which yield exact sequences of the form
[TABLE]
and establishes the equivalence of simple-connectedness of with the simple-connectedness of .
A key to the proof both in the simply-laced and in the general case is the computation of the fundamental groups of generalized flag varieties – that is, spaces of the form for a parabolic subgroup of corresponding to an index subset . It turns out that the aforementioned space is a universal covering space of an appropriately chosen generalized flag variety. In general, we prove the following theorem:
Theorem**.**
Let be an irreducible Dynkin diagram such that the Bruhat decomposition of provides a CW decomposition (i.e., such that the conclusion of Proposition 3.7 holds), let be the index set of the Dynkin diagram, let , and let be a parabolic of type . Then a presentation of is given by
[TABLE]
In particular, this statement holds in the -spherical and in the symmetrizable case.
We refer to [Wig98] for the analog result in the finite-dimensional situation.
In order to determine in the general case, we compute subgroups of corresponding to the index sets of connected components of using the above theorem and covering maps of the type where is a maximal split torus of and is the subgroup of fixed points of a Levi factor of with both and invariant under the Cartan–Chevalley involution. We then show that is a direct product of appropriately chosen such subgroups.
In a very similar way, the fundamental group of is determined, establishing the following theorem:
Theorem**.**
Let be an irreducible Dynkin diagram such that the Bruhat decomposition of provides a CW decomposition (i.e., such that the conclusion of Proposition 3.5 holds). Let be the number of connected components of of colour . Let be the number of connected components of on which takes the value 1 and which have colour . Then
[TABLE]
In particular, this statement holds in the -spherical and in the symmetrizable case.
Acknowledgements. The research leading to this article has been partially funded by DFG via the project KO 4323/11. The authors thank Julius Grüning and two anonymous referees for various helpful remarks on earlier versions of this article.
2 Split-real Kac–Moody groups
In [Kac90, §1.3], Kac associates with every generalized Cartan matrix a quadruple of a complex Lie algebra , an abelian subalgebra and linearly independent finite subsets and called simple roots and simple coroots, respectively, such that . Associated with such a quadruple is a Lie algebra generating set . The complex Lie algebra is called the complex Kac–Moody algebra associated with , and its standard Cartan subalgebra.
Since , one can analogously define a quadruple where is a real Lie algebra that embeds naturally into as the real form given by the involution induced by complex conjugation. One refers to as the split real Kac–Moody algebra associated with and to as its standard split Cartan subalgebra.
Let be the group generated by and the subsemigroups generated by , respectively. For and define the root space
[TABLE]
The set of roots in is defined as . One has the root space decomposition
[TABLE]
The set decomposes as a disjoint union into the subsets called positive, respectively negative roots. The restriction of the Lie bracket on to
[TABLE]
turns and into Lie subalgebras of .
For define the fundamental root reflection by
[TABLE]
Then the Weyl group of is defined as and forms a Coxeter system together with the set of fundamental root reflections. Finally, define the set of real roots and , the positive, respectively negative real roots.
The construction in [Tit87] of (see Definition 1.1) provides a representation of on by Lie algebra automorphisms, which is denoted by
[TABLE]
and referred to as the adjoint representation of . Since the subgroup of under this representation preserves the commutator subalgebra , one obtains an adjoint representation
[TABLE]
for . The kernels of the adjoint representations of and are given by the respective centres.
An element is -locally-finite if for every element there exists an -invariant finite-dimensional subspace with . As pointed out in [Mar18, p. 64], this implies that is a (finite) matrix in some basis of , so the exponential can be defined in the ususal way. By [Tit87, (KMG5), p. 545] and the uniqueness properties of established in [Tit87, Theorem 1], . Let and be the subsets of -locally-finite elements of the respective algebras. The maps and given by can be lifted to exponential functions and .
For , one has
[TABLE]
cf. [Kum02, Section 6.1.6], [Tit87, (KMG5), p. 545].
The same constructions apply also to instead of . Since , one can define . Note that . There is a unique Lie group topology on in which and . The centre of is contained in .
The intersection is called the standard split maximal torus of ; again, is of finite index in and contains the centre of .
The Lie algebra admits a unique involution which maps to for all and acts as on . There exists a unique involutive automorphism such that for all , and this involutive automorphism is called the Cartan–Chevalley involution of . We denote by the fixed point subgroup of this involution and define .
Let be a real root. Then is one-dimensional and consists of ad-locally-finite elements. One can therefore define the root group . Each root group carries a unique Lie group topology such that as topological groups. Root groups corresponding to positive real roots are called positive root groups, root groups corresponding to negative real roots are called negative root groups.
Define the positive, respectively negative maximal unipotent subgroup of as the group generated respectively by the positive and negative root groups. One has . The groups are normalized by and intersect trivially. In particular, they intersect the centres of and trivially and hence embed into both and .
If , then and the group is isomorphic to . The groups with are called the rank- subgroups and the groups are called the fundamental rank- subgroups of .
One can show that the pair defines an RGD system for . For details concerning RGD systems, we refer the reader to [AB08, Chapter 8].
Recall that the generalized Cartan matrix is called -spherical, if for all ; in other words, if the orders of the products are always finite. The generalized Cartan matrix is symmetrizable, if it is the product of a symmetric and a diagonal matrix. These notions are also applied to any and all objects that are derived from such as the (extended) Weyl group, the Kac–Moody group, their buildings, etc.
Definition and Remark 2.1.
The Kac–Peterson topology on equals the finest group topology on such that the natural embeddings and are continuous when and the root groups are endowed with their Lie group topologies.
The Kac–Peterson topology is by [HKM13, Proposition 7.10] and, in particular, Hausdorff. Moreover, for every , it induces the unique connected Lie group topology on and on by [HKM13, Corollary 7.16]
For more details on the Kac–Peterson topology, see [HKM13, Chapter 7].
Notation 2.2.
Throughout this paper, let be the algebraically simply connected semisimple split real Kac–Moody group associated to an irreducible generalized Dynkin diagram with (bijective) labelling . Let be the maximal compact subgroup of , i.e., the subgroup fixed by the Cartan–Chevalley involution .
Denote by the positive Borel subgroup of the twin -pair of , by the standard split maximal torus and by the Weyl group of with generating set . For each , take to be a fixed representative of order for . The group is called the extended Weyl group. By [DMGH09, Corollary 1.7], one has an Iwasawa decomposition .
The groups and are always endowed with the subspace topologies induced by the Kac–Peterson topology on and with the quotient topology.
Unless specified more explicitly, the symbol will always denote an arbitrary subset of the index set , the symbol the subdiagram of corresponding to , the symbol the subgroup of , and the symbols and the intersections and , respectively. This is consistent with the notation for the fundamental rank one subgroups: One has .
Remark 2.3.
Due to the structure theory of RGD systems (cf. [AB08, Chapter 8], most notably the fact that restricting an RGD system to a subdiagram again yields an RGD system), for each fundamental rank one subgroup there exists an (abstract) isomorphism with the following properties: Let be the group of upper triangular matrices in and let denote the canonical root subgroups of . Then
- •
.
- •
.
- •
For each , , and hence
- •
.
By [HKM13, Corollary 7.16], the restriction of the Kac–Peterson topology to any spherical subgroup of coincides with its Lie topology. That is, the groups inherit their Lie group topology from the topological Kac–Moody group . By the classical theory of Lie groups this yields the existence of a diffeomorphism with the desired properties; in particular, is an open map.
Definition 2.4.
Using the Bruhat decomposition ([AB08, Theorem 6.56, Remark (1)]), let
[TABLE]
be the Weyl distance function on , and let be the length function that associates to each element the (unique) length of a corresponding reduced expression in . Let be the strong Bruhat order on : Recall that for one has if there exist reduced expressions of and such that the former is a (not necessarily consecutive) substring of the latter.
For and a chamber define
[TABLE]
[TABLE]
and
[TABLE]
In particular, one has and for with representative in the extended Weyl group . A set is called a -panel.
Moreover, for a subset with representatives define to be the standard parabolic subgroup corresponding to the index set , that is, .
Throughout this paper, and will always be endowed with the subspace topologies induced by .
Lemma 2.5.
Let . Then the following hold:
- (a)
* . In particular, .* 2. (b)
. In particular,
Proof.
Assertions (a) and (b) follow from [AB08, Remark 8.51] and [AB08, Remark (2) after Theorem 6.56], respectively, and the Iwasawa decomposition . ∎
3 The fundamental group of the generalized flag variety
For a moment, let be an irreducible simply-laced diagram distinct from , and let and be as in the preceding section. Moreover, let be the double cover of constructed in [GHKW17, Lemma 16.18] (see Definition 4.6 below). By construction, any -subdiagram of yields an embedding and, since inherits the Lie topology from the Kac–Peterson topology on by [HKM13, Corollary 7.16], one obtains a locally trivial fibre bundle
[TABLE]
by [Pal61] (see Proposition A.13). It will turn out in Section 4 below that is a universal covering space of the generalized flag variety where equals the set consisting of the two types involved in the chosen -subdiagram. The fundamental group of then follows from the homotopy exact sequence
[TABLE]
This motivates our interest in the fundamental group and covering theory of generalized flag varieties .
Throughout this section, let , let be the subgroup of generated by , and let be a set of representatives of the cosets in that have minimal length in the coset they define.
Lemma 3.1 (Bruhat decomposition).
One has .
Proof.
This follows immediately from [AB08, Theorem 6.56, Remark (1)]. ∎
Lemma 3.2.
Let be a topological group and subgroups of and endow with the quotient topology. Then the following hold:
- (a)
The projection map is continuous and open. 2. (b)
The canonical map is continuous and open.
Proof.
This is a standard exercise for topological groups. ∎
Definition and Remark 3.3.
For , define the following restrictions of the canonical map :
- •
,
- •
.
Since is continuous, the same holds for the two restrictions. The space is compact by [HKM13, Corollary 3.10] and so is a quotient map.
Lemma 3.4.
Let be -spherical or symmetrizable and let . Then the canonical map is a homeomorphism.
Proof.
By Remark 3.3, is a quotient map. One has : Let such that . Then where the equality holds since by definition of one has for all which implies . The Bruhat decomposition of yields and hence, . This implies .
Now, since is open in its closure in (see [HKM13, Proposition 5.9] plus Corollary B.8), the preceding observations yield that is an injective quotient map and therefore a homeomorphism. ∎
Lemma 3.5.
Let . Then each panel is homeomorphic to the -sphere .
Proof.
The panel is a subbuilding of corresponding to the RGD system . By Remark 2.3 one has , and where denotes the subgroup of diagonal matrices and denote the canonical root subgroups of . This implies that is homeomorphic to the building . ∎
Definition 3.6.
Following [Rot88, Chapter 8], a CW complex is an ordered triple , where is a Hausdorff space, is a family of cells in , and is a family of maps, such that
- (a)
. 2. (b)
For , let be the union of all cells of dimension . Then for each -cell , the map , is a relative homeomorphism, i.e., it is a continuous map and its restriction is a homeomorphism. 3. (c)
If , then its closure is contained in a finite union of cells in . 4. (d)
has the weak topology determined by , i.e., a subset of is closed if and only if is closed in for each .
For , let be an index set for the -dimensional cells, so that and set . This map is called the characteristic map of .
Proposition 3.7.
Let be -spherical or symmetrizable. Then for each , the set is a cell of dimension that is open in its compact closure in . For each subset , the Bruhat decomposition is a CW decomposition.
Proof.
The first statement is immediate by [HKM13, Corollary 3.10 and Proposition 5.9] plus Corollary B.8, see also [Kra01, p. 170, 171]. Furthermore, [HKM13, Proposition 5.9] combined with Corollary B.8 states that the Bruhat decomposition of is a CW decomposition. By Lemma 3.4, is composed of cells that are homeomorphic to cells in , so composing the characteristic maps of the latter cells with the canonical map yields characteristic maps for the cells in .
For the closure-finiteness, let be a cell in . Since is continuous and restricts to a homeomorphism , it maps surjectively onto . Now, , which implies that
[TABLE]
where the last equality holds since . This proves that is contained in a finite union of cells.
It remains to show that has the weak topology determined by the cell closures.
For and a minimal-length representative of , one has . Let and . Let and .
Let be a closed subset of and let , , be an arbitrary cell. Then is closed in since is continuous, so is closed in since is a CW complex. Now,
[TABLE]
Since is a quotient map by Remark 3.3, this implies that is closed in .
Now, let be a subset of such that is closed in for all . Since for each one has for any minimal-length representative of , in fact is closed in for all . Therefore is closed in for all . Since , the fact that is a CW complex implies that is closed in . Since is open by Lemma 3.2, it follows that is closed in . This proves that is a CW complex. ∎
The preceding result combined with the following lemma (which is a consequence of [Mas77, Ch. 7, Thm 2.1]) will allow us to efficiently compute the fundamental group of a generalized flag variety in Theorem 3.15 below.
Lemma 3.8.
Let be a complex with only one [math]-cell . For each , let be a loop whose homotopy class generates and whose image under is a loop in starting at . Then
[TABLE]
is a presentation of , where the brackets denote the respective homotopy classes in .
Next, we study the characteristic maps of the CW decomposition of a generalized flag variety explicitly.
Notation 3.9.
Define
Lemma 3.10.
* induces a continuous, surjective map which maps the interior homeomorphically onto its image and maps the boundary surjectively onto its image.*
Proof.
Let where denotes the real projective line, modelled as the subset of one-dimensional subspaces of . Since each one-dimensional subspace in contains exactly one element in the upper half circle while contains the two boundary points corresponding to and , one has a surjection from onto given by which maps bijectively onto . Since acts transitively on the real projective line with being the stabilizer of , one has a bijective correspondence between and . This yields the desired surjectivity and bijectivity properties of . Continuity is clear, as well as the fact that the restriction to the interior is a homeomorphism. ∎
Definition 3.11.
Let be the -dimensional unit disc and note that . For let be as in Remark 2.3. Let be the canonical projection. Define and by
- •
,
- •
.
The following lemma was inspired by [Pro07, Ch. 10, second Proposition of 6.8], see also [Kac85, §2.6, p. 198].
Lemma 3.12.
Let be -spherical or symmetrizable. Then the maps defined above are characteristic maps for the following cells:
- (a)
* for ,* 2. (b)
* for .*
Proof.
(a): One has to show that and that is a continuous map which maps homeomorphically to . The first assertion is clear, since by Lemma 2.5 one has .
By Lemma 2.5, one has . Let . Then . By Lemma 3.10, there exists a unique satisfying . Hence, is the unique preimage of under . This yields the desired bijectivity property. The continuity properties are clear.
(b): Since by Lemma 2.5 (b) one has , it is clear that . For the injectivity of the restriction, let such that . Then
[TABLE]
This implies , since otherwise the left expression is in , contradicting . Since , one obtains . It follows that , hence by (a).
For the surjectivity, note that by Lemma 2.5 (b), one has . Let be an arbitrary element of with and . By (a), there exists an with . Hence, there exists a with . Again by (a), there exists a with . This yields
[TABLE]
This proves that maps bijectively to The continuity properties are clear. ∎
Notation 3.13.
For , let where is the -entry of the Cartan matrix of .
Lemma 3.14 [GHKW17, Remark 15.4(1)].
Let with as in Remark 2.3 and . Then .
Theorem 3.15**.**
If the Bruhat decomposition satisfies the conclusion of Proposition 3.7, then a presentation of is given by
[TABLE]
In particular, this statement holds in the -spherical and the symmetrizable case.
Proof.
By Lemma 3.4 and Proposition 3.7, the Bruhat decomposition is a CW decomposition where each cell has dimension . The characteristic maps of the 1-cells and 2-cells are given by the compositions , respectively ( and denoting the canonical homeomorphisms from Lemma 3.4).
Lemma 3.8 gives a presentation of . The generating elements are given by the homotopy classes of the characteristic maps of the 1-cells – namely, the cells where . For the homotopy classes with , note that , and so which implies . This yields the desired generating set as well as the trivial relation for .
To obtain the set of relators, for let where
[TABLE]
Then the concatenation is a loop in the relative boundary which generates its fundamental group. Moreover, for each characteristic map of a 2-cell, one has where is the unique 0-cell of the CW complex. Therefore, Lemma 3.8 implies that the set of relators is given by . Now,
[TABLE]
where
with which implies
[TABLE]
Moreover,
[TABLE]
Since , this yields . One therefore obtains . This proves the assertion.
∎
Lemma 3.16.
Let be irreducible simply-laced distinct from and . Then .
Proof.
For each generator in the presentation of Theorem 3.15, one has : Recall that denotes the labelling map of the vertex set of . Since is connected, one has a minimal path in such that . If , one has by the presentation above. Let have order . Since is simply-laced, which implies and . Multiplying these expressions yields .
Since each generator has order , the relations show that the group is abelian. One concludes that . ∎
4 The fundamental groups of and
The Iwasawa decomposition implies that acts transitively on the generalized flag varieties . In this section, we describe the generalized flag varieties and suitable covering spaces as coset spaces of and its various spin covers defined in [GHKW17]. This will then allow us to compute the fundamental group of and its various spin covers via locally trivial fibre bundles and homotopy exact sequences.
Lemma 4.1.
The canonical map is a homeomorphism. In particular, there exists a homeomorphism .
Proof.
Bijectivity follows from the product formula for subgroups since . By Lemma 3.2, the map is continuous, so the same holds for its bijective restriction .
In order to show that is closed, let and let . Consider the commutative diagram
[TABLE]
where denotes the canonical embedding and denotes the canonical map from to . Since is closed in by [FHHK20, Section 3F], the map is closed. By Lemma 3.2, is open.
Let be a closed subset of and suppose that is not closed in . Then the complement is not open in , hence the complement is not open in . Therefore, is not closed in . This yields that is not closed in , a contradiction.
For the second claim, since and , one has . Furthermore, is normal in which implies . The claim follows. ∎
The key advantage of the description of a generalized flag variety as a -coset space lies in the fact that is a finite group. It is therefore straightforward to write down covering spaces of generalized flag varieties via the following well-known basic observation from covering theory.
Lemma 4.2.
Let be a continuous, open, surjective map between Hausdorff topological spaces. If all fibers are finite and of constant cardinality, then is a covering map.
This readily applies in our setting:
Lemma 4.3.
The canonical map is a covering map of degree .
Proof.
By Lemma 3.2, is continuous, open and surjective.
By [FHHK20, Lemma 3.20 and the discussion after Prop 3.8], the group has order . Note that one has , since the Kac–Moody group being algebraically simply connected implies . Now, for one has , and since , one has for . This yields . Lemma 4.2 now shows that is a covering map.∎
Definition 4.4 [GHKW17, Definition 16.2].
Let be the graph on the vertex set with edge set
[TABLE]
where denotes the parity of the corresponding Cartan matrix entry, as defined in Notation 3.13.
An admissible colouring of is a map such that
- (a)
whenever there exists with and . 2. (b)
the restriction of to any connected component of the graph is a constant map.
Define to be the number of connected components of on which takes the value 2. For a subgraph of that is a union of connected components of let be the corresponding restriction of .
Definition 4.5.
Let be the colouring of that to each connected component of assigns a colour as follows: Let be coloured red (denoted by ) if it contains a vertex such that there exists a vertex satisfying and ; let be coloured green () if it is not red and consists only of an isolated vertex; and blue () else.
We refer to the introduction for a discussion of various examples.
Definition and Remark 4.6.
Recall from the introduction that in [GHKW17, Definition 16.16], the spin group with respect to and is defined as the universal enveloping group of a particular -amalgam where the isomorphism type of depends on the - and -entries of the Cartan matrix of as well as the values of on the corresponding vertices. The group can be regarded as (being uniquely isomorphic to) the universal enveloping group of an -amalgam where each covers via an epimorphism . By [GHKW17, Lemma 16.18] there exists a canonical central extension that makes the following diagram commute for all :
[TABLE]
Here, and denote the respective canonical maps into the universal enveloping groups.
By [GHKW17, Proposition 3.9], one has
[TABLE]
Each connected component of that admits a vertex with contributes a factor to the order of so that is a -fold central extension of .
In particular, this implies that the subspace topology on defines a unique topology on that turns the extension into a covering map. The resulting group topology on is called the Kac–Peterson topology on .
In the case of an irreducible simply-laced diagram , the only admissible colourings are the (trivial) constant colouring (that every diagram admits) and the constant colouring ; we define the spin group with respect to as .
Before turning to the general case, we will first consider the simply-laced case and formulate and prove the corresponding simplified versions of the main theorems.
Lemma 4.7.
Let be irreducible simply-laced distinct from and let be the index set of an -subdiagram of . Then the spaces and are homeomorphic.
Proof.
From [GHKW17] (exact references below) it follows that the kernel of the covering map coincides with the kernel of the covering map and is equal to the group (for the definition of , see below). This is a consequence of the following facts regarding an irreducible simply-laced diagram (all referring to [GHKW17]):
- •
There is an epimorphism with kernel (see [Theorem 6.8]).
- •
In , all elements coincide (see [Lemma 11.7]).
- •
Let for an arbitrary pair . Then (see [Corollary 11.16]).
- •
is a 2-fold central extension of (see [Theorem 11.17]).
Hence, the 2-fold covering map induces a continuous bijective map . One has a commutative diagram
[TABLE]
Since is open as a covering map and is open by Lemma 3.2, it follows that is a homeomorphism. ∎
Lemma 4.8.
Let be irreducible simply-laced distinct from and . Then is simply connected.
Proof.
is connected since is generated by connected groups isomorphic to . Hence by Lemma 4.3 it is a non-trivial cover of of degree . The claim now follows from Corollary 3.16 and Corollary 4.1. ∎
The following proposition provides our main result in the simply laced case.
Proposition 4.9.
Let be irreducible simply-laced distinct from . Then is simply connected with respect to the Kac–Peterson topology. In particular, .
Proof.
By [Hus94, 4.2.4], for a closed subgroup of a topological group , the projection is a principal -bundle. By Lemma A.13, this bundle is locally trivial if is a (closed) Lie group (note that, by [HR63, Theorem 5.11], every locally compact subgroup of a topological group is closed). Since locally trivial bundles admit local cross sections, [Ste99, Corollary in Section 7.4] implies that, if is a closed Lie group, then is a fibre bundle with fibre . This yields a locally trivial fibre bundle
[TABLE]
By [Hat02, Chapter 4], this yields the homotopy long exact sequence
[TABLE]
from which one extracts the exact sequence
[TABLE]
By Lemmas 4.7 and 4.8 one has and so by exactness .
The second assertion follows from the fact that by Corollary A.15 and the fact that is a 2-fold central extension of by [GHKW17, Theorem 11.17]. ∎
We will now return to the case of a general irreducible Dynkin diagram .
Notation 4.10.
For a subset let
[TABLE]
Lemma 4.11.
Let be the index set of a connected component of . Then the following hold:
- (a)
If has colour , then . 2. (b)
If has colour , then and . 3. (c)
If has colour , then .
Proof.
(a): If has colour , then there exist with and . This implies and which yields . Now, if is an edge in , then . Multiplying these expressions shows that implies . Since is connected, this yields for each . Commutativity then follows from the relations of .
(b): By definition, nodes of colour are isolated in .
(c): Let be the simply laced Dynkin diagram with vertex set and edge set . Let where denotes the standard maximal torus of . Then by Lemma 4.3 and Proposition 4.9, is a universal covering map where has degree and has degree according to [GHKW17, Theorem 11.17]. Since by Theorem 3.15 and Lemma 4.1, this implies . ∎
Proposition 4.12.
Let be the index sets of the connected components of . If the Bruhat decomposition satisfies the conclusion of Proposition 3.7, then
[TABLE]
Proof.
By Theorem 3.15, where
[TABLE]
as defined in 4.10. For , let
[TABLE]
the set of relators of . Let
[TABLE]
the set of commutators of pairs of generators from different connected components of . Then
[TABLE]
Let and be the canonical homomorphisms from the free group to and , respectively. It suffices to show that and . It is clear that a relator with and in a common connected component is contained in , so let with and in different connected components of . Then one has . If , then , so let and . Then is contained in a connected component of colour , and by Lemma 4.11, .
This implies that has order in , hence , the normal closure of in the free group.
Since , one obtains . Since and , one therefore has
[TABLE]
Conversely, it is clear that , so let with and in different connected components. As above, we can assume that and . Since , this implies
[TABLE]
This proves the assertion. ∎
Theorem 4.13**.**
Let be an irreducible Dynkin diagram such that satisfies the conclusions of Proposition 3.7 and of Theorem A.15. Let and be the number of connected components of of colour and , respectively. Then
[TABLE]
In particular, this statement holds in the symmetrizable case.
Proof.
By Theorem A.15, , so it suffices to prove that is of the given isomorphism type; note that Theorem A.15 has only been established in the symmetrizable case. Let . The diagram
[TABLE]
with all maps being the respective canonical maps, commutes. Since the maps are continuous by Lemma 3.2, one obtains a commutative diagram of induced homomorphisms
[TABLE]
where and are injective, because and are covering maps (see Lemma 4.3). By Theorem 3.15 and Lemma 4.1, and can be identified with and , respectively ( as in (4) in the above proof), where corresponds to the canonical homomorphism between these groups as the proof of Theorem 3.15 shows.
For the index set of a connected component of , let . Then by Proposition 4.12,
[TABLE]
Summing up, one obtains a commutative diagram
[TABLE]
having replaced and from above with the corresponding monomorphisms.
By Lemma 4.3, the covering has degree This implies that is a subgroup of of index . The isomorphism type of is uniquely determined by this index and Lemma 4.11: One has
[TABLE]
Again by Lemma 4.3, the covering has degree , so is a subgroup of index of . The commutative diagram above implies that . Since this holds for the index set of every connected component of , one has . But the latter is a subgroup of index of , so equality holds. This proves the assertion. ∎
Theorem 4.14**.**
Let be an irreducible Dynkin diagram such that satisfies the conclusion of Proposition 3.7. Let be the number of connected components of of colour . Let be the number of connected components of on which takes the value 1 and which have colour . Then
[TABLE]
In particular, this statement holds in the -spherical and the symmetrizable case.
Proof.
By [GHKW17, Theorem 17.1], the map is a -fold central extension. Let be the index set of a connected component of and let . Let .
Since , one has a continuous induced map making the following diagram commute, where and denote the respective canonical maps:
[TABLE]
Each fiber of has cardinality
[TABLE]
Since is open as a covering map and is open by Lemma 3.2, it follows from Lemma 4.2 that is a covering map.
From here the proof is analogous to the proof of Theorem 4.13, after extending the commutative diagram at the beginning of the latter proof:
[TABLE]
One obtains that where each is a subgroup of index of
[TABLE]
Since is the union of all connected components except , one has , depending on whether is constant or on . This implies
[TABLE]
This proves the assertion. ∎
Now all theorems from the introduction have been proved.
Appendix A Maximal unipotent subgroups of Kac–Moody groups and applications to Kac–Moody symmetric spaces (by Tobias Hartnick and Ralf Köhl)
Throughout this appendix we fix a symmetrizable generalized Cartan matrix with underlying diagram . We consider the corresponding algebraically simply-connected semisimple split real Kac–Moody group as given by Definition 1.1. As in Section 2 we also denote by the fixed point subgroup of the Cartan–Chevalley involution and set . We equip all of these groups with the restrictions of the Kac–Peterson topology.
The goal of this appendix is to relate the topology of to the topology of . Our main result (see Theorem A.15 below) asserts that the inclusion is a weak homotopy equivalence. This implies in particular that and thus allows the computation of by the methods presented in the main part of the article.
In the spherical case the subgroup is even a deformation retract and hence the inclusion is a homotopy equivalence, as a consequence of the topological Iwasawa decomposition of . This decomposition also implies that the associated Riemannian symmetric space is contractible.
While real Kac–Moody groups also possess an Iwasawa decomposition, it is currently unknown whether this decomposition is topological. To establish our main result we thus have to work with a certain central quotient of , for which the topological Iwasawa decomposition was established in [FHHK20]. We will show that the image of in is a strong deformation retract and that the reduced Kac–Moody symmetric space is contractible. Since the finite-dimensional central extension is a Serre fibration by a classical result of Palais [Pal61], this will allow us to deduce the desired result about and .
A.1 The topological Iwasawa decomposition
Let us denote by and the adjoint representations of and respectively. We recall from [FHHK20] that the quotient map factors as
[TABLE]
where is uniquely determined by the fact that is a torus and has finite kernel. The group is referred to as the semisimple adjoint quotient of , and we equip it with the quotient topology with respect to the Kac–Peterson topology on . We will denote by the positive, respectively negative maximal unipotent subgroup of as introduced in Section 2. Also recall from Section 2 that and set .
Lemma A.1 (Iwasawa decomposition).
Multiplication induces continuous bijections
[TABLE]
Proof.
This follows from [KP85, Proposition 5.1(a)]. ∎
A more refined statement has been established in [FHHK20] for the semisimple adjoint quotient of . To state this result, denote by
[TABLE]
the canonical quotient maps from (5) and set , , and . Equip these groups with their respective quotient topologies and note that restricts to a bijection between and .
Theorem A.2** (Topological Iwasawa decomposition, [FHHK20, Theorem 3.23]).**
Multiplication induces homeomorphisms
[TABLE]
Since is contractible, in order to show that is a deformation retract of it will suffice to show that is contracible. We thus need to understand the topology induced by the Kac–Peterson topology on the standard unipotent subgroups.
A.2 The Kac–Peterson topology on
We now turn to the study of the restriction of the Kac–Peterson topology to the standard maximal unipotent subgroups and . Recall from Section 2 that the Weyl group is a Coxeter group, so elements of can be represented by reduced words in the generators . Given such a reduced word in with corresponding simple roots we define positive roots by
[TABLE]
We then set and define a map
[TABLE]
It is established in [CR09, Section 5.5, Lemma] that the map is a bijection for every reduced word , and that its image depends only on the Weyl group element represented by , but not on the chosen reduced expression. Since is a topological group, the bijection is continuous. In fact, one can show that is a homeomorphism. A proof of this fact was sketched in [HKM13, Lemma 7.25]; since openness of the maps is crucial for everything that follows, we fill in the details of this sketch here:
Lemma A.3.
For every reduced word the map is a homeomorphism onto its image.
Proof.
We argue by induction on the length of and observe that the case holds by definition. Since the linear functionals are linearly independent, there exists an element (see Section 2) such that and for all . It follows that and for all . Indeed, since the word is reduced, none of the positive real roots equals , and since is not a root for any (cf. [Kac90, Proposition 5.1]), each of them contains at least one other positive simple root as a summand. Now for and we have , and thus
[TABLE]
We conclude that if , then
[TABLE]
where the convergence is uniform on compacta. This shows that the map
[TABLE]
is continuous, and hence the map
[TABLE]
is continuous. Now let and let , …, . Now by Axiom (RGD2) of an RGD system (see [AB08, Chapter 8]) there exists an element such that for all , and by induction hypothesis we have a homeomorphism
[TABLE]
Conjugating the inverse of this homeomorphism by we obtain a homeomorphism
[TABLE]
Composing this homeomorphism with the map (7) now provides the desired continuous inverse to . ∎
To describe the topology on we recall that there exist several distinct but related partial orders on which in different places in the literature are referred to as the Bruhat order on . In the sequel we will consider the following version; here denotes the length function with respect to the generating set .
Definition A.4.
The weak right Bruhat order on is the partial order defined as
[TABLE]
According to [CR09, p. 44] we have if and only if there exists a reduced word for such that .
Recall that for the strong Bruhat order one has if there exists a reduced word for and a reduced word for such that is a substring of (not necessarily consecutive). By definition,
[TABLE]
but the converse is not true. An important difference between the weak right Bruhat order and the strong Bruhat order is that contains a cofinal chain, i.e., a totally ordered subset such that for every there exists such that , whereas for the weak right Bruhat order, such a cofinal chain does not exist. In fact, given there will in general not exist an element with and .
Note that if , then we can choose a reduced word for such that . Thus if we define as above then we have a commuting diagram
[TABLE]
where the horizontal maps are inclusions, and the vertical maps are homeomorphisms. In particular, we have a continuous inclusion , hence we may form the colimit
[TABLE]
in the category of topological spaces. We emphasize that in view of the previous remark the system is not directed, hence this colimit is not a direct limit.
Proposition A.5.
The -space is given by the colimit
[TABLE]
both in the category of topological spaces and in the category of -topological spaces.
Proof.
The corresponding statement in the category of sets is established in [CR09, Theorem 5.3]. For the topological statement see [HKM13, Proposition 7.27]. ∎
In view of the applications to Kac–Moody symmetric spaces that we have in mind we recall that are subgroups of the commutator subgroup of , in particular we can consider their images under the map from (5). In this context we will need the following fact:
Proposition A.6.
The map induces homeomorphisms .
Proof.
By [HKM13, Proposition 7.27] the map is a homeomorphism and the kernel of is contained in . The latter implies that restricts to a continuous bijection , and the former implies that this bijection is open. ∎
A.3 Dilation structures on
Definition A.7.
Let be a topological group. By a dilation structure on we mean a family of maps with the following properties:
- (a)
Each is a continuous automorphisms of the topological group . 2. (b)
is a -parameter group, i.e. and for all . 3. (c)
If we define by , then the map
[TABLE]
is continuous.
Remark A.8.
Note that if a topological group admits a dilation structure, then it is in particular contractible. Indeed, if we define , then
[TABLE]
is continuous with and , hence a contraction to the identity.
Dilation structures on finite-dimensional simply-connected nilpotent Lie groups play a major role in conducting analysis on such groups, see e.g. [Goo76]. Not every finite-dimensional simply-connected nilpotent Lie group admits a dilation structure, but if is the unipotent radical of a minimal parabolic subgroup of a semisimple Lie group, then such a dilation structure always exists. The methods of [Kum02] allow one to extend this result to the Kac–Moody setting.
Following [Kac90, §3.12], we define the fundamental chamber of as
[TABLE]
Since the family is linearly independent, there exists
[TABLE]
Indeed, by the linear independence of the solution space for the system of linear equations has strictly larger dimension than the solution space for the system of linear equations .
We now define a -parameter subgroup of by and denote by
[TABLE]
the associated automorphism of the Lie algebra . Similarly we denote by
[TABLE]
the restriction of the conjugation-action of on to . Note that if is ad-locally finite then
[TABLE]
From (1) and the defining property of one deduces that for every positive root with height
[TABLE]
It follows that for all positive roots one has
[TABLE]
(see [Tit87, (4), p. 549]), where is the root subgroup of corresponding to the root space . As a consequence, if one endows each of the root subgroups with the natural topology of , then contracts each of them. We are now in a position to reproduce the following result and proof by Kumar:
Theorem A.9** ([Kum02, Proposition 7.4.17]).**
The family defines a dilation structure on .
Proof.
Let be a reduced word and write with corresponding simple roots . Recall that multiplication induces a homeomorphism
[TABLE]
where the roots are given by
[TABLE]
Given an element by (8) one has
[TABLE]
Setting for all , we deduce that the map
[TABLE]
is continuous and that . Combining this with Proposition A.5 one deduces that the map
[TABLE]
is continuous, hence a dilation structure. ∎
Recall that is isomorphic to under the Cartan–Chevalley involution of , which maps to . Thus if we define then we obtain:
Corollary A.10.
The family defines a dilation structure on .∎
Combining this with Remark A.8 and Proposition A.6 we can record:
Corollary A.11.
The topological groups and are contractible. Consequently, the groups and are contractible.∎
A.4 Homotopy groups of real-split semisimple Kac–Moody groups
Corollary A.12.
The subgroup is a deformation retract. In particular the inclusion is a homotopy equivalence and thus induces isomorphisms for all .
Proof.
We have established in Corollary A.11 that is contractible, and is contractible since it is homeomorphic to . The assertion now follows from Theorem A.2. ∎
Since it is currently unknown whether the Iwasawa decomposition of is also a topological decomposition, the strategy of the above proof can not be applied to . However, using the following result of Palais [Pal61, Section 4.1, Corollary], one can still obtain an isomorphism between the fundamental groups of and .
Proposition A.13 (Palais).
Let be a topological group and let be a subgroup which is homeomorphic to a Lie group. Then the fibration is locally trivial, in particular a Hurewicz fibration, hence there is a long exact sequence of homotopy groups
[TABLE]
Recall that the kernel of the quotient map is homeomorphic to . In particular it has connected components, whereas its higher homotopy groups vanish. Applying Proposition A.13 to the diagram of fibrations
[TABLE]
we thus obtain:
Corollary A.14.
There is a commutative diagram with exact rows
[TABLE]
Moreover, for there are isomorphisms and .
Combining this with Corollary A.12 we deduce:
Theorem A.15**.**
For every the inclusion induces isomorphisms
[TABLE]
hence is a weak homotopy equivalence. In particular, . ∎
A.5 Kac–Moody symmetric spaces and causal contractions
We conclude this appendix with an application to the results obtained so far to Kac–Moody symmetric spaces. It was established in [FHHK20] that the homogeneous spaces and carry the natural structure of topological reflection spaces, and the same holds for their quotients and . The topological reflection space is called the unreduced Kac–Moody symmetric space of type , and the topological reflection space is called the reduced Kac–Moody symmetric space of type .
Corollary A.16.
The reduced symmetric space is contractible.
Proof.
In view of the topological Iwasawa decomposition the orbit map at the basepoint
[TABLE]
is a homeomorphism. Since and are contractible, this implies contractability of . ∎
The proof of Theorem A.9 can be used to provide an explicit contraction for , using the contraction by conjugation with suitable elements of the torus on the group and the standard contraction on the finite-dimensional real vector space . It turns out that this contraction has interesting additional properties. Recall from [FHHK20, Section 7] that the symmetric space admits future and past boundaries and that both carry a simplicial structure which turns them in the geometric realizations of the positive and negative halves of the twin building of . Following [FHHK20, Section 7], a causal ray is a geodesic ray of whose parallelity class equals a point in and a piecewise geodesic causal curve is the concatenation of a finite set of segments of causal rays that can be parametrized in such a way that the walking direction always points towards the future boundary. Given we say that causally preceeds (in symbols ) if there exists a piecewise geodesic causal curve from to .
Since both conjugation by elements of and the standard contraction of the vector space preserve geodesic rays and the future and past boundaries (cf. [FHHK20, Section 7]), the set of piecewise geodesic causal curves of , and hence the causal pre-order , are invariant under the given contraction.
Corollary A.17.
The reduced symmetric space is causally contractible, i.e., it admits a contraction that preserves .
Appendix B The Bruhat decomposition is a CW decomposition (by Julius Grüning and Ralf Köhl)
Let be a Kac–Moody group endowed with the Kac–Peterson topology and let be the standard maximal torus and , the standard unipotent subgroups. [KP83, Theorem 4(a)] asserts without proof that the multiplication map
[TABLE]
is a homeomorphism with respect to the Kac–Peterson topology. In this note we provide a proof in the symmetrizable case that makes use of this fact in the two-spherical case ([HKM13, Proposition 7.31]), of the embedding of Kac–Moody groups constructed in [Mar19, Theorem 3.15(2)], and of the fact that the Kac–Peterson topology is . Among the various consequences of this result is that the Bruhat decomposition of a symmetrizable topological Kac–Moody group is a CW decomposition.
Recall that a -space (alternatively: compactly generated space) is a topological space in which a set is closed if and only if its intersection with any compact subset of is compact. That is, a -space is a topological space whose topology is coherent with the family of all compact subspaces of . A -space is a topological space whose topology is coherent with respect to a countable ascending family of compact subspaces. By (3) of [FST77] any -space is a -space.
Proposition B.1 ([Pal70, Corollary]).
A continuous proper map from a topological space to a -space is closed. In particular, a continuous injection into a -space with compact such that for each the pre-image is also compact is a topological embedding, i.e., it is a homeomorphism onto its image.
Proof.
The first statement is exactly [Pal70, Corollary]. The second statement is an immediate consequence of the first, since a -space is a -space in which any compact subset of is contained in some of the ascending family of compact subsets (statement (3) of [FST77]). ∎
Remark B.2.
The authors thank Tobias Hartnick and Stefan Witzel for various lively discussions concerning the correct formulation and application of Proposition B.1. Moreover, they thank Stefan Witzel for suggesting to make use of the concept of proper maps.
A subgroup of a Kac–Moody group is bounded if it lies in the intersection of two spherical parabolic subgroups of opposite signs. In other words, it is bounded if and only if it stabilises a point the Davis CAT(0) realization of each half of its twin building. The maximal bounded subgroups of a Kac–Moody group have been determined in [CM06, Theorem 4.1].
Proposition B.3.
Let be a split real Kac–Moody group. Then the Kac–Peterson topology on equals the finest group topology on such that the embeddings of the maximal bounded subgroups, each endowed with its Lie group topology, are continuous.
Proof.
By [Mar15, Lemma 4.3], the Kac–Peterson topology on induces the Lie group topology on its maximal bounded subgroups. A fundamental is bounded and, in particular, embeds as a closed subgroup into a maximal bounded subgroup. Therefore its subspace topology equals its Lie group topology; by [HKM13, Proposition 7.21] the topology equals the finest group topology on such that the embeddings of the fundamental Lie subgroups is continuous, whence is finer than or equal to the final group topology with respect to the embedded maximal bounded subgroups. Again, since by [Mar15, Lemma 4.3] the Kac–Peterson topology on induces the Lie group topology on its maximal bounded subgroups, the two described topologies actually coincide. ∎
Corollary B.4.
Let be a split real Kac–Moody group endowed with the Kac–Peterson topology and let be a finite family of Lie-subgroups of such that each fundamental is contained in at least one of the . Then the Kac–Peterson topology on equals the finest group topology on such that the embeddings of the , each endowed with its Lie group topology, are continuous.
Proposition B.5 (cf. [HKL15, 1.5, 1.10], [Mar19, Theorem 3.15(2)]).
Any symmetrizable topological Kac–Moody group endowed with the Kac–Peterson topology admits a continuous injective group homomorphism into a simply laced topological Kac–Moody group with closed image with respect to the Kac–Peterson topology.
Proof.
By [Mar19, Theorem 3.15(2)] for any symmetrizable Kac–Moody group there is an injective group homomorphism into a simply laced Kac–Moody group embedding each fundamental rank- subgroup diagonally into the direct product
[TABLE]
of a suitable (finite) family of fundamental rank- subgroups of .
The restriction of this map to any fundamental rank- subgroup of is continuous with respect to the Lie group topology on and the Kac–Peterson topology on . Hence, by universality (see [HKM13, Proposition 7.21]), the map is continuous with respect to the Kac–Peterson topology on both and .
One has
[TABLE]
where is the automorphism of given by
[TABLE]
for some , where acting by permuting the factors of the direct product . Since the automorphisms are continuous with respect to the Kac–Peterson topology on , the group is a closed subgroup of . ∎
The embedding corresponds to an embedding of the twin building of into the twin building of such that (with the now considered as twin building automorphisms) and the additional property that two chambers of are opposite in if and only if they are opposite in .
Indeed, this is immediate from an argument along the lines of descent in buildings (cf. [MPW15]). The automorphisms act on the twin apartment defined by the fundamental chambers , of and, by definition, the fixed substructure is isometric to a twin apartment of . The claim then follows from the fact that acts transitively on the twin apartments of .
In particular, this embedding
[TABLE]
of twin buildings induces an embedding of opposite geometries
[TABLE]
Specialising to the embedding of a fundamental rank- subgroup of diagonally into the direct product
[TABLE]
of a suitable (finite) family of fundamental rank- subgroups of , one obtains an embedding of the real projective line (the building of type ) diagonally into a suitable product of real projective lines (the building of type ).
This in turn yields an embedding of the corresponding opposites geometries of pairs of distinct points of with adjacency relation given by the complete relation (the opposite geometry of type of diameter ), respectively of -tuples of pairs of distinct points of with adjacency relation given by equality in all up to at most one component (the opposite geometry of type of diameter ).
Refer to [Gra09, Section 4.3] for more details, some examples, and applications of the opposite geometry. The most striking application of the opposite geometry is a proof of [Tit74, Theorem 13.32] via its simple connectedness and Mühlherr’s generalization to Kac–Moody groups333A manuscript that has never been published and unfortunately seems to be lost. To the second author’s dismay he has lost his copy that he once owned.; see also [AM97].
The following result follows immediately from the preceding discussion:
Proposition B.6.
Let be the injective group homomorphism from Proposition B.5, let be the induced embedding of twin buildings, and the resulting embedding of opposite geometries. Given , for all exists such that the intersection of with the ball of radius in around is contained in the ball of radius in around .
Corollary B.7.
Let be a topological Kac–Moody group endowed with the Kac–Peterson topology. If it is two-spherical or symmetrizable, then the multiplication map is a homeomorphism onto its image.
Proof.
The two-spherical case is [HKM13, Proposition 7.31]. In the symmetrizable case note that Proposition B.1 is applicable since the Kac–Peterson topology is by [HKM13, Proposition 7.10]. Consequently, the injection from Proposition B.5 yields a topological embedding , provided one can find -decompositions and such that each intersection lies in some . (Indeed, is closed by continuity of , so it is compact once it lies inside some compact set , which is equivalent to .)
For and choose -decompositions making use of Corollary B.4 and -decompositions of the fundamental subgroups of and the corresponding subgroups of into which the embed diagonally, endowed with their Lie group topology. That is,
[TABLE]
where each of the is the ball of radius around of the maximal bounded subgroup endowed with some suitable metric inducing its Lie group topology, with and lower index taken modulo the total number of maximal bounded subgroups.
By construction, each intersects in some compact subset of a fundamental subgroup of with respect to the Lie group topology. In other words, each lies in some . Forming finite products of such sets and using Proposition B.6 one concludes that lies in some suitable product ; that is, the injective homomorphism indeed is a topological embedding.
Since restricts to maps {\left.\kern-1.2pt\iota\vphantom{\big{|}}\right|_{U_{+}^{G}}}:U_{+}^{G}\to U_{+}^{H}, {\left.\kern-1.2pt\iota\vphantom{\big{|}}\right|_{U_{-}^{G}}}:U_{-}^{G}\to U_{-}^{H}, {\left.\kern-1.2pt\iota\vphantom{\big{|}}\right|_{T^{G}}}:T^{G}\to T^{H}, one can conclude that the diagram
[TABLE]
commutes, which proves that the map is a homeomorphism onto its image, since is a homeomorphism onto its image by [HKM13, Proposition 7.31]. ∎
Corollary B.8.
Let be a topological Kac–Moody group endowed with the Kac–Peterson topology. If it is two-spherical or symmetrizable, then the associated twin building with the quotient topology is a strong topological twin building.
Proof.
The two-spherical case is [HKM13, Theorem 1]. In the symmetrizable case it follows by replacing [HKM13, Proposition 7.31] with Corollary B.7; cf. the discussion after [HKM13, Theorem 1]. ∎
Corollary B.9.
Let be a topological Kac–Moody group endowed with the Kac–Peterson topology. If it is two-spherical or symmetrizable, then the Bruhat decomposition of a symmetrizable Kac–Moody group is a CW decomposition.
Proof.
This is a restatement of Proposition 3.7 from the main text. Its proof heavily relies on Corollary B.8. ∎
Corollary B.10.
Let be a topological Kac–Moody group endowed with the Kac–Peterson topology. If it is two-spherical or symmetrizable, then the coset model, the group model, and the involution model of the reduced Kac–Moody symmetric space are pairwise homeomorphic with respect to their internal topologies.
Proof.
The two-spherical case is [FHHK20, Proposition 4.19]. In the symmetrizable case it follows from [FHHK20, Proposition 4.19] and Corollary B.7. ∎
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[AB 08] Peter Abramenko and Kenneth S. Brown, Buildings. Theory and applications , Springer, New York, 2008.
- 2[AM 97] Peter Abramenko and Bernhard Mühlherr, Présentations de certaines B N 𝐵 𝑁 BN -paires jumelées comme sommes amalgamées , C. R. Acad. Sci. Paris Sér. I Math. 325 (1997), 701–706.
- 3[CM 06] Pierre-Emmanuel Caprace and Bernhard Mühlherr, Isomorphisms of Kac-Moody groups which preserve bounded subgroups , Adv. Math. 206 (2006), 250–278.
- 4[CR 09] Pierre-Emmanuel Caprace and Bertrand Rémy, Groups with a root group datum , Innov. Incidence Geom. 9 (2009), 5–77.
- 5[DMGH 09] Tom De Medts, Ralf Gramlich, and Max Horn, Iwasawa decompositions of split Kac-Moody groups , J. Lie Theory 19 (2009), 311–337.
- 6[Dyn 46] Evgenii Dynkin, Classification of the simple Lie groups , Rec. Math. [Mat. Sbornik] N. S. 18(60) (1946), 347–352.
- 7[FHHK 20] Walter Freyn, Tobias Hartnick, Max Horn, and Ralf Köhl, Kac–Moody symmetric spaces , Münster J. Math. 13 (2020), 1–114.
- 8[FST 77] Stanley P. Franklin and Barbara V. Smith Thomas, A survey of k ω subscript 𝑘 𝜔 k_{\omega} -spaces , Topology Proceedings 2 (1977), 111–124.
