Satake diagram

In the mathematical study of Lie algebras and Lie groups, Satake diagrams are a generalization of Dynkin diagrams that classify involutions of root systems that are relevant in several contexts. They were introduced in Template:Harvtxt and were originally used to classify real simple Lie algebras. Additionally, they also classify symmetric pairs $(𝔤,𝔨)$ of Lie algebras, where $𝔤$ is semisimple.

More concretely, given a complex semisimple Lie algebra $𝔤$, the Satake diagrams made from $𝔤$'s Dynkin diagram classify the involutions of $𝔤$'s root system that extend to an anti-linear involutive automorphism Template:Mvar of $𝔤$.Template:Sfn The fixed points ${𝔤}^{\sigma }$ are then a real form of $𝔤$. The same Satake diagrams also classify the involutions of $𝔤$'s root system that extend to a (linear) involutive automorphism Template:Mvar of $𝔤$.Template:Sfn The fixed points $𝔨$ form a complex Lie subalgebra of $𝔤$, so that $(𝔤,𝔨)$ is a symmetric pair.

More generally, the Tits index or Satake–Tits diagram of a reductive algebraic group over a field is a generalization of the Satake diagram to arbitrary fields, introduced by Template:Harvtxt, that reduces the classification of reductive algebraic groups to that of anisotropic reductive algebraic groups.

Satake diagrams are distinct from Vogan diagrams although they look similar.

Definition

Let Template:Mvar be a real vector space. A σ-root systemTemplate:Sfn $(R,\sigma )$ consists of a root system $R\subset V$ that spans Template:Mvar and a linear involution Template:Mvar of Template:Mvar that satisfies $\sigma (R)=R$.

Let $R_{\bullet }\subset R$ be the set of roots fixed by Template:Mvar and let $$\mathrm{\Sigma }:=\{\frac{\alpha -\sigma (\alpha )}{2}:\alpha \in R\setminus R_{\bullet }\}.$$ Template:Mvar is called the restricted root system.

The Satake diagram of a σ-root system $(R,\sigma )$ is obtained as follows: Let ${\alpha }_1,\dots ,{\alpha }_n$ be simple roots of Template:Mvar such that ${\alpha }_{n+1-p},\dots ,{\alpha }_n$ are simple roots of $R_{\bullet }$. We can define an involution Template:Mvar of $\{1,\dots ,n-p\}$ by having $$\sigma ({\alpha }_i)={\alpha }_{\tau (i)}+\mathbb{Z}{R}_{\bullet }(i=1,\dots ,n-p).$$ The Satake diagram is then obtained from the Dynkin diagram describing Template:Mvar by blackening the vertices corresponding to ${\alpha }_{n+1-p},\dots ,{\alpha }_n$, and by drawing arrows between the white vertices that are interchanged by Template:Mvar.

Satake diagram of a real semisimple Lie algebra

Let ${𝔤}_{ℝ}$ be a real semisimple Lie algebra and let $𝔤={𝔤}_{ℝ}\otimes ℂ$ be its complexification. Define the map $$\sigma :𝔤\to 𝔤,X\otimes z\mapsto X\otimes \bar{z}.$$ This is an anti-linear involutive automorphism of real Lie algebras and its fixed-point set is our original ${𝔤}_{ℝ}$.

Let $𝔥\le 𝔤$ be a Cartan subalgebra that satisfies $\sigma (𝔥)=𝔥$ and is maximally split, i.e. when we split $𝔥$ into Template:Mvar-eigenspaces, the $-1$-eigenspace has maximal dimension. Template:Mvar induces an anti-linear involution Template:Mvar on ${𝔥}^{\ast }$: $${\sigma }^{\ast }(\lambda )(v)=\overline{\lambda (\sigma (v))}(\lambda \in {𝔥}^{\ast },v\in 𝔥).$$ If $X\in {𝔤}_{\alpha }$ is a root vector, one can show that $\sigma (X)\in {𝔤}_{{\sigma }^{\ast }(\alpha )}$. Consequently, Template:Mvar preserves the root system Template:Mvar of $𝔤$. We thus obtain a σ-root system $(R,{\sigma }^{\ast })$ whose Satake diagram is the Satake diagram of ${𝔤}_{ℝ}$.

Satake diagram of a symmetric pair

Let $(𝔤,𝔨)$ be a symmetric pair of complex Lie algebras where $𝔤$ is semisimple, i.e. let Template:Mvar be an involutive Lie algebra automorphism of $𝔤$ and let $𝔨$ be its fixed-point set. It is shown in Template:Harvtxt that these symmetric pairs (even for $𝔤$ an infinite-dimensional Kac-Moody algebra), or equivalently these involutive automorphisms, can be classified using so-called admissible pairs. These admissible pairs describe again a σ-root system that can be obtained from the automorphism Template:Mvar, and the Satake diagrams that arise this way are exactlyTemplate:Sfn the ones listed in Template:Harvtxt and the Satake diagrams obtained by blackening all vertices.

DefinitionTemplate:Sfn Given a Dynkin diagram with vertex set Template:Mvar, an admissible pair $(I_{\bullet },\tau )$ consists of a subset $I_{\bullet }$ of finite type and a diagram automorphism Template:Mvar satisfying

• ${\tau }^2=\mathrm{id}$
• $\tau (I_{\bullet })=I_{\bullet }$
• The permutation $\tau {|}_{I_{\bullet }}$ coincides with $-w_{\bullet }$ (where $w_{\bullet }$ is the longest element of the Weyl group generated by the vertices in $I_{\bullet }$)
• For $j\in I\setminus I_{\bullet }$ with $\tau (i)=i$, we have ${\alpha }_j({\rho }_{\bullet }^{\vee })\in \mathbb{Z}$, where $${\rho }_{\bullet }^{\vee }=\frac{1}{2}\sum _{\alpha \in R_{\bullet }^{+}}{\alpha }^{\vee }.$$

Given an admissible pair $(I_{\bullet },\tau )$, we can define a σ-root system by equipping the root system Template:Mvar of Template:Mvar with the involution $\sigma =-w_{\bullet }\circ \tau$

Classification of Satake diagrams

In Template:Harvtxt it is proven that every Satake diagram arising from a real semisimple Lie algebra (equivalently: symmetric pair $(𝔤,𝔨)$ with $𝔤$ semisimple) is a disconnected union of

• two times the same Dynkin diagram (with white vertices), with arrows matching the vertices
• one of the following diagrams:

Connected irreducible Satake diagrams

Name	Diagram	Type of restricted root system	Example real semisimple Lie algebra	Example symmetric pair ($ℂ$ omitted)
${𝖠𝖨}_n$		${𝖠}_n$	$𝔰𝔩(n+1)$	$(𝔰𝔩(n+1),𝔰𝔬(n+1))$
${𝖠𝖨𝖨}_{2p+1}$		${𝖠}_p$	${𝔰𝔲}^{\ast }(2p+2)$	$(𝔰𝔩(2p+2),𝔰𝔭(2p))$
${𝖠𝖨𝖨𝖨}_{n,p}(4\le 2p\le n)$		${𝖡𝖢}_p$	$𝔰𝔲(p,n+1-p)$	$(𝔰𝔩(n+1),𝔰(𝔤𝔩(p)\oplus 𝔤𝔩(n+1-p)))$
${𝖠𝖨𝖨𝖨}_{2p-1,p}$		${𝖢}_p$	$𝔰𝔲(p,p)$	$(𝔰𝔩(2p),𝔰(𝔤𝔩(p)\oplus 𝔤𝔩(p)))$
${𝖠𝖨𝖵}_n$		${𝖡𝖢}_1$	$𝔰𝔲(n,1)$	$(𝔰𝔩(n+1),𝔤𝔩(n))$
${𝖡𝖨}_{n,p}(2\le p\le n)$		${𝖡}_p$	$𝔰𝔬(p,2n+1-p)$	$(𝔰𝔬(2n+1),𝔰𝔬(p)\oplus 𝔰𝔬(2n+1-p))$
${𝖡𝖨𝖨}_n$		${𝖠}_1$	$𝔰𝔬(2n,1)$	$(𝔰𝔬(2n+1),𝔰𝔬(2n))$
${𝖢𝖨}_n$		${𝖢}_n$	$sp(2n)$	$(𝔰𝔭(2n),𝔤𝔩(2n))$
${𝖢𝖨𝖨}_{n,p}(2\le 2p\le n-1)$		${𝖡𝖢}_p$	$𝔰𝔭(2p,2n-2p)$	$(𝔰𝔭(2n),𝔰𝔭(2p)\oplus 𝔰𝔭(2n-2p))$
${𝖢𝖨𝖨}_{2p,p}$		${𝖢}_p$	$𝔰𝔭(2p,2p)$	$(𝔰𝔭(4p),𝔰𝔭(2p)\oplus 𝔰𝔭(2p))$
${𝖣𝖨}_{n,p}(2\le p\le n-2)$		${𝖡}_p$	$𝔰𝔬(p,2n-p)$	$(𝔰𝔬(2n),𝔰𝔬(p)\oplus 𝔰𝔬(2n-p))$
${𝖣𝖨}_{n,n-1}$		${𝖡}_{n-1}$	$𝔰𝔬(n-1,n+1)$	$(𝔰𝔬(2n),𝔰𝔬(n-1)\oplus 𝔰𝔬(n+1))$
${𝖣𝖨}_{n,n}$		${𝖣}_n$	$𝔰𝔬(n,n)$	$(𝔰𝔬(2n),𝔰𝔬(n)\oplus 𝔰𝔬(n))$
${𝖣𝖨𝖨}_n$		${𝖠}_1$	$𝔰𝔬(2n-1,1)$	$(𝔰𝔬(2n),𝔰𝔬(2n-1))$
${𝖣𝖨𝖨𝖨}_{2p}$		${𝖢}_p$	${𝔰𝔬}^{\ast }(4p)$	$(𝔰𝔬(4p),𝔤𝔩(2p))$
${𝖣𝖨𝖨𝖨}_{2p+1}$		${𝖡𝖢}_p$	${𝔰𝔬}^{\ast }(4p+2)$	$(𝔰𝔬(4p+2),𝔤𝔩(2p+1))$
$𝖤𝖨$		${𝖤}_6$	${𝔢}_{6,6}$	$({𝔢}_6,𝔰𝔭(8))$
$𝖤𝖨𝖨$		${𝖥}_4$	${𝔢}_{6,2}$	$({𝔢}_6,𝔰𝔩(6)\oplus 𝔰𝔩(2))$
$𝖤𝖨𝖨𝖨$		${𝖡𝖢}_2$	${𝔢}_{6,-14}$	$({𝔢}_6,𝔰𝔬(10)\oplus ℂ)$
$𝖤𝖨𝖵$		${𝖠}_2$	${𝔢}_{6,-26}$	$({𝔢}_6,{𝔣}_4)$
$𝖤𝖵$		${𝖤}_7$	${𝔢}_{7,7}$	$({𝔢}_7,𝔰𝔩(8))$
$𝖤𝖵𝖨$		${𝖥}_4$	${𝔢}_{7,-5}$	$({𝔢}_7,𝔰𝔬(12)\oplus 𝔰𝔩(2))$
$𝖤𝖵𝖨𝖨$		${𝖢}_3$	${𝔢}_{7,-25}$	$({𝔢}_7,{𝔢}_6\oplus ℂ)$
$𝖤𝖵𝖨𝖨𝖨$		${𝖤}_8$	${𝔢}_{8,8}$	$({𝔢}_8,𝔰𝔬(16))$
$𝖤𝖨𝖷$		${𝖥}_4$	${𝔢}_{8,-24}$	$({𝔢}_8,{𝔢}_7\oplus 𝔰𝔩(2))$
$𝖥𝖨$		${𝖥}_4$	${𝔣}_{4,4}$	$({𝔣}_4,𝔰𝔭(6)\oplus 𝔰𝔩(2))$
$𝖥𝖨𝖨$		${𝖡𝖢}_1$	${𝔣}_{4,-20}$	$({𝔣}_4,𝔰𝔬(9))$
$𝖦$		${𝖦}_2$	${𝔤}_{2,2}$	$({𝔤}_2,𝔰𝔩(2)\oplus 𝔰𝔩(2))$

Examples

• Compact Lie algebras correspond to the Satake diagram with all vertices blackened. This corresponds to the symmetric pair $(𝔤,𝔤)$
• Split Lie algebras correspond to the Satake diagram with only white (i.e., non blackened) and unpaired vertices.
• A table can be found at Template:Harv.

Differences between Satake and Vogan diagrams

Both Satake and Vogan diagrams are used to classify semisimple Lie groups or algebras (or algebraic groups) over the reals and both consist of Dynkin diagrams enriched by blackening a subset of the nodes and connecting some pairs of vertices by arrows. Satake diagrams, however, can be generalized to any field (see above) and fall under the general paradigm of Galois cohomology, whereas Vogan diagrams are defined specifically over the reals. Generally speaking, the structure of a real semisimple Lie algebra is encoded in a more transparent way in its Satake diagram, but Vogan diagrams are simpler to classify.

The essential difference is that the Satake diagram of a real semisimple Lie algebra $𝔤$ with Cartan involution θ and associated Cartan pair $𝔤=𝔨\oplus 𝔭$ (the +1 and −1 eigenspaces of θ) is defined by starting from a maximally noncompact θ-stable Cartan subalgebra $𝔥$, that is, one for which $\theta (𝔥)=𝔥$ and $𝔥\cap 𝔨$ is as small as possible (in the presentation above, $𝔥$ appears as the Lie algebra of the maximal split torus S), whereas Vogan diagrams are defined starting from a maximally compact θ-stable Cartan subalgebra, that is, one for which $\theta (𝔥)=𝔥$ and $𝔥\cap 𝔨$ is as large as possible.

The unadorned Dynkin diagram (i.e., that with only white nodes and no arrows), when interpreted as a Satake diagram, represents the split real form of the Lie algebra, whereas it represents the compact form when interpreted as a Vogan diagram.

Generalisation: Satake—Tits diagrams

Suppose that G is an algebraic group defined over a field k, such as the reals. We let S be a maximal split torus in G, and take T to be a maximal torus containing S defined over the separable algebraic closure K of k. Then G(K) has a Dynkin diagram with respect to some choice of positive roots of T. This Dynkin diagram has a natural action of the Galois group of K/k. Also some of the simple roots vanish on S. The Satake–Tits diagram is given by the Dynkin diagram D, together with the action of the Galois group, with the simple roots vanishing on S colored black. In the case when k is the field of real numbers, the absolute Galois group has order 2, and its action on D is represented by drawing conjugate points of the Dynkin diagram near each other, and the Satake–Tits diagram is called a Satake diagram.

See also

• Relative root system
• List of irreducible Tits indices

References

• Template:Citation
• Template:Citation
• Template:Citation
• Template:Citation
• Template:Citation
• Template:Citation
• Template:Citation
• Template:Citation
• Template:Cite journal
• Template:Cite journal