tinytheorems - about groups - rings - modules - models atom/rss

Exploration 3: Representation theory

.

Questions:


Last exploration, we studied the properties of individual RR-matrices. This time we’re studying RR-matrices collectively.

First, we define HomR(M,N)\Hom_R(M,N) as the set of all RR-module homomorphisms M→NM\to N. When M,NM,N are free RR-modules, the elements of HomR(M,N)\Hom_R(M,N) are RR-matrices. Generally this means HomR(M,N)\Hom_R(M,N) is a noncommutative ring, given that matrix multiplication is not commutative in general.

Let’s start by representing groups using subsets of Hom\Hom.

For instance, we can express the group C×\CC^\times (nonzero complex numbers under multiplication) as 2×22\times 2 real matrices. We define a map ρ:C×→HomR(R2,R2)\rho:\CC^\times\to\Hom_\RR(\RR^2,\RR^2):

ρ(1)=[1001]ρ(i)=[01−10]ρ(a+bi)=[ab−ba] \rho(1)=\left[\begin{matrix}1&0\\0&1\end{matrix}\right]\quad \rho(i)=\left[\begin{matrix}0&1\\-1&0\end{matrix}\right]\quad \rho(a+bi)=\left[\begin{matrix}a&b\\-b&a\end{matrix}\right]

so that ρ((a+bi)(c+di))=ρ(a+bi)ρ(c+di)=[ab−ba][cd−dc]=[ac−bdad+bc−ad−bcac−bd]=ρ((ac−bd)+(ad+bc)i)\begin{aligned} \rho((a+bi)(c+di)) &=\rho(a+bi)\rho(c+di)\\ &=\left[\begin{matrix}a&b\\-b&a\end{matrix}\right] \left[\begin{matrix}c&d\\-d&c\end{matrix}\right]\\ &=\left[\begin{matrix}ac-bd&ad+bc\\-ad-bc&ac-bd\end{matrix}\right]\\ &=\rho((ac-bd)+(ad+bc)i) \end{aligned} and ρ((a+bi)−1)=(ρ(a+bi))−1=[ab−ba]−1=det([ab−ba])adj([ab−ba])=1a2+b2[a−bba]=ρ(a−bia2+b2)\begin{aligned} \rho((a+bi)^{-1}) &=(\rho(a+bi))^{-1}\\ &=\left[\begin{matrix}a&b\\-b&a\end{matrix}\right]^{-1}\\ &=\det\left(\left[\begin{matrix}a&b\\-b&a\end{matrix}\right]\right)\adj\left(\left[\begin{matrix}a&b\\-b&a\end{matrix}\right]\right)\\ &=\frac{1}{a^2+b^2}\left[\begin{matrix}a&-b\\b&a\end{matrix}\right]\\ &=\rho\left(\frac{a-bi}{a^2+b^2}\right) \end{aligned}

Since it preserves product and inverse, these matrices (homomorphisms) define a group, isomorphic to C×\CC^\times. This is just one of many encodings of complex numbers into matrices. The key is that i2=−Ii^2=-I, and that 1,i1,i don’t interact with each other.

Here we found that HomR(R2,R2)\Hom_\RR(\RR^2,\RR^2) was a suitable subgroup of RR-module homomorphisms. In general, to represent a group, we must select the subgroup of HomR(M,N)\Hom_R(M,N) that consists of invertible RR-module homomorphisms. In practice, MM and NN are free modules RmR^m and RnR^n, and invertibility implies m=nm=n. Thus, when representing a group, we can choose among the invertible matrices in HomR(Rn,Rn)\Hom_R(R^n,R^n), also written as Aut(Rn)\Aut(R^n), the group of automorphisms on RnR^n, whose elements we express as invertible n×nn\times n RR-matrices.

So we assign to each group element an invertible n×nn\times n RR-matrix. A matrix RR-representation of a group is a group homomorphism G→Aut(Rn)G\to\Aut(R^n).

Theorem: There is a one-to-one correspondence between RR-module automorphisms and matrix RR-representations of a finite group GG.
  • Every representation G→Aut(Rn)G\to\Aut(R^n) assigns each g∈Gg\in G to a specific permutation matrix in Aut(Rn)\Aut(R^n).
  • But this implicitly defines an action of GG on RnR^n. The action is given by matrix muliplication in RnR^n, where each gg acts by multiplying by its corresponding matrix.
  • Conversely, by Cayley’s theorem every group is isomorphic to a permutation group, and therefore can act on R∣G∣R^{|G|} by permuting the ∣G∣|G| basis RR-vectors. But permutations are automorphisms, so this essentially assigns each element g∈Gg\in G an automorphism over R∣G∣R^{|G|}, i.e. constructing a matrix RR-representation G→Aut(R∣G∣)G\to\Aut(R^{|G|}).

The representation G→Aut(R∣G∣)G\to\Aut(R^{|G|}) (mentioned in the above proof) is known as the regular representation of a finite group GG. The main idea of this representation is to have GG permute the indices of R∣G∣R^{|G|} according to the permutation group isomorphic to GG.

In this section, we show the module analog of group actions.

In the same manner that adjoining an indeterminate symbol xx to RR gives you a polynomial ring R[x]R[x], we can adjoin a whole group to RR to get a group ring R[G]R[G]. Elements look like RR-linear combinations of elements of GG: ∑g∈Grgg\sum_{g\in G}r_gg

Theorem: Every group ring R[G]R[G] is a free module.

Since every element in R[G]R[G] is by definition uniquely represented as a linear combination of elements g∈Gg\in G, we know that GG spans R[G]R[G]. Like in polynomial rings, the only linear combination equal to zero is the one where every coefficient is zero, so the elements of GG are linearly independent and therefore form a basis of R[G]R[G]. Having a basis means the elements of R[G]R[G] can be represented by RR-vectors of coefficients, and therefore the group ring R[G]R[G] is a free module by construction.

Thus a group ring is simultaneously a group and a free RR-module.

Theorem: The group ring R[G]R[G] is exactly the regular representation of GG over RR.

Since R[G]R[G] is a free RR-module, it has a basis whose elements are precisely the elements of the group GG. It is a property of groups that any element g∈Gg\in G permutes the elements of gg by left-multiplication. Permuting the elements is an automorphism on GG and therefore an automorphism on R∣G∣R^|G|, thus R[G]R[G] is exactly the regular representation G→Aut(R∣G∣)G\to\Aut(R^{|G|}).

Therefore: Every group action of GG on an RR-module MM can be encoded by an appropriate group ring R[G]R[G].

A representation G→Aut(M)G\to\Aut(M) assigns an automorphism on the RR-module MM to every element g∈Gg\in G.

Since the automorphisms on RR-modules are exactly the linear transformations by definition, every g∈Gg\in G is assigned a linear transformation on GG. In other words, the action of gg on MM must be linear in MM.

But R[G]R[G] contains exactly every linear combination of GG with coefficients in RR, and therefore contains every possible action of GG on MM.

This means that every action defined for GG on some module MM can be linearly extended to an action of R[G]R[G] on MM. Specifically, if the group action g⋅mg\cdot m is defined for all g∈G,m∈Mg\in G, m\in M, then we define (∑g∈Grgg)⋅m(\sum_{g\in G}r_gg)\cdot m as ∑g∈Grg(g⋅m)\sum_{g\in G}r_g(g\cdot m). Such a module MM is known as a R[G]R[G]-module, a module over a group ring.

Theorem: Every group ring R[G]R[G] is an R[G]R[G]-module.

R[G]R[G] is a free RR-module, and GG has a natural group action on R[G]R[G] by left-multiplication in the ring. Thus it is a R[G]R[G]-module.

To preserve the R[G]R[G]-module structure, R[G]R[G]-submodules of R[G]R[G]-modules must be GG-invariant: just as how R[G]R[G]-modules are closed under the action of R[G]R[G], R[G]R[G]-submodules WW must be closed under the action of R[G]R[G] in the sense that for all g∈R[G],w∈Wg\in R[G],w\in W, gw∈Wgw\in W.

Similarly, homomorphisms σ:M→N\sigma:M\to N between R[G]R[G]-modules M,NM,N must be GG-equivariant: they must commute with the action of R[G]R[G] in the sense that for all g∈R[G],m∈Mg\in R[G],m\in M, gσ(m)=σ(gm)g\sigma(m)=\sigma(gm).

Theorem: Given finite GG, we can always construct a GG-equivariant version σ~:M→M\tilde{\sigma}:M\to M of any endomorphism σ:M→M\sigma:M\to M of the R[G]R[G]-module MM.

The trick to get GG-equivariance is to take a sum over all actions of GG: σ~(m)=∑g∈Gg−1σ(gm)\tilde{\sigma}(m)=\sum_{g\in G}g^{-1}\sigma(gm) which we can do because GG is finite. Then we can show that for h∈Gh\in G: hσ~(m)=h∑g∈Gg−1σ(gm)=∑g∈Ghg−1σ(gm)=∑g∈Gh(gh)−1σ((gh)m) since g↦gh simply reorders the sum over G=∑g∈Gg−1σ(g(hm))=σ~(hm)\begin{aligned} h\tilde{\sigma}(m)&=h\sum_{g\in G}g^{-1}\sigma(gm)\\ &=\sum_{g\in G}hg^{-1}\sigma(gm)\\ &=\sum_{g\in G}h(gh)^{-1}\sigma((gh)m)&\text{ since }g\mapsto gh\text{ simply reorders the sum over }G\\ &=\sum_{g\in G}g^{-1}\sigma(g(hm))\\ &=\tilde{\sigma}(hm) \end{aligned} and thus σ~\tilde{\sigma} is GG-equivariant.

Theorem: If GG is finite and ∣G∣|G| is a unit in RR and MM is a free R[G]R[G]-module, then every R[G]R[G]-submodule WW of MM implicitly defines a GG-equivariant projection M→WM\to W.
  • Since MM is free, it has a basis, and so does the submodule WW. We can obtain a projection π:M→W\pi:M\to W by mapping the basis vectors not in WW to basis vectors in WW.
  • Recall that endomorphisms like projections can be made GG-equivariant via a “sum” trick. To ensure that the resulting endomorphism π~:M→M\tilde{\pi}:M\to M is still a projection, we need to ensure that it is idempotent and that its image is WW. It is enough to construct π~\tilde{\pi} as a map that fixes elements w∈Ww\in W (thus ensuring idempotence and that the image is at least WW) and maps other elements in MM to an element in WW (thus ensuring that the image is at most WW).
  • Given that ∣G∣|G| is a unit in RR, the map obtained by the “sum” trick can be modified into an “averaging” trick to ensure the above properties: π~(m)=1∣G∣∑g∈Gg−1σ(gm)\tilde{\pi}(m)=\frac{1}{|G|}\sum_{g\in G}g^{-1}\sigma(gm)
  • This π~\tilde{\pi} fixes w∈Ww\in W: π~(w)=1∣G∣∑g∈Gg−1π(gw)=1∣G∣∑g∈Gg−1π(w′) since W is G-invariant=1∣G∣∑g∈Gg−1w′ since π is a projection onto W=1∣G∣∑g∈Gw since w′=gw=1∣G∣∣G∣w=w\begin{aligned} \tilde{\pi}(w)&=\frac{1}{|G|}\sum_{g\in G}g^{-1}\pi(gw)\\ &=\frac{1}{|G|}\sum_{g\in G}g^{-1}\pi(w')&\text{ since }W\text{ is }G\text{-invariant}\\ &=\frac{1}{|G|}\sum_{g\in G}g^{-1}w'&\text{ since }\pi\text{ is a projection onto }W\\ &=\frac{1}{|G|}\sum_{g\in G}w&\text{ since }w'=gw\\ &=\frac{1}{|G|}|G|w\\ &=w \end{aligned} and maps all m∈Mm\in M to an element in WW, since π(gm)∈W by definition of π  ⟹  g−1π(gm)∈W since W is G-invariant  ⟹  1∣G∣∑g∈Gg−1σ(gm)∈W by linearity of the group action  ⟹  π~(m)∈W\begin{aligned} &\pi(gm)\in W&\text{ by definition of }\pi\\ \implies&g^{-1}\pi(gm)\in W&\text{ since $W$ is $G$-invariant}\\ \implies&\frac{1}{|G|}\sum_{g\in G}g^{-1}\sigma(gm)\in W&\text{ by linearity of the group action}\\ \implies&\tilde{\pi}(m)\in W \end{aligned}
  • Thus π~\tilde{\pi} as defined is a GG-equivariant projection onto WW.

In this section, we consider representations as algebraic structures in their own right.

Note that for our representation for C×\CC^\times in the beginning, each element of the group is assigned a distinct element of Aut(Rn)\Aut(\RR^n), i.e. the representation is injective. When the representation is injective, we call it a faithful representation, and they are the typical ones where each element of GG is represented by a distinct matrix in Aut(Rn)\Aut(R^n).

There are also non-faithful representations. An example of a very not faithful representation is the trivial representation, which represents every element as the identity map in Aut(Rn)\Aut(R^n). Another is the sign representation: Sn→Aut(Z)S_n\to\Aut(\ZZ), which simply takes the sign (11 or −1-1) of every permutation in SnS_n. Another example is det∘ρ\det\circ\rho, where ρ\rho is some matrix representation that we then take the determinant of. In general, a representation is a group homomorphism G→Aut(M)G\to\Aut(M) (where MM is some RR-module).

Theorem: A R[G]R[G]-module MM fully describes a representation G→Aut(M)G\to\Aut(M).

This result comes naturally from the fact that a group ring R[G]R[G] encodes all possible linear group actions and therefore its action on a module MM models every possible automorphism on the RR-modules MM, since automorphisms on RR-modules are linear by definition.

Theorem: A faithful representation ρ:G→Aut(M)\rho:G\to\Aut(M) is one where no g∈Gg\in G acts as the identity action on VV except for the identity element e∈Ge\in G.

A faithful representation is injective, i.e. its domain GG is isomorphic to its image Aut(M)\Aut(M). This means that only one element g∈Gg\in G maps to the identity automorphism, so only one element g∈Gg\in G acts as identity on MM.

Theorem: A faithful representation ρ\rho, when represented as a R[G]R[G]-module MM, is one where each x∈R[G]x\in R[G] has a unique action on MM.

If two elements x,y∈R[G]x,y\in R[G] have the same action on MM, then x−yx-y must be the zero action (sending all elements of MM to zero). But x−y=0x-y=0 implies x=yx=y.

The regular representation ρ:G→Aut(R∣G∣)\rho:G\to\Aut(R^{|G|}) represents elements of GG as automorphisms on the free RR-module R∣G∣R^{|G|} (whose elements are RR-vectors). Recalal that ρ=R[G]\rho=R[G] is another way to fully describe this representation: here the elements g∈Gg\in G act on the elements ∑g∈Grgg∈R[G]\sum_{g\in G}r_gg\in R[G] naturally by multiplication, which is another way to say it’s an automorphism on RR-vectors whose entries are rgr_g. The regular representation is always faithful, since it is composed of distinct permutations of GG, so there is always a distinct matrix representation of any group GG. Can we do better? Ideally, we want to represent elements of a group GG using RR-vectors that are perhaps smaller than ∣G∣|G|, without affecting the faithfulness of ρ\rho.

By using the above fact that any representation ρ\rho is equivalent to some R[G]R[G]-module MM, we can take subrepresentations of ρ\rho as R[G]R[G]-submodules of MM. To begin, given a representation ρ:G→Aut(M)\rho:G\to\Aut(M), we can think about quotienting the corresponding R[G]R[G]-module MM by one of its R[G]R[G]-submodules. However, quotienting might create a representation that isn’t faithful. When does quotienting the underlying module MM of a faithful representation ρ:G→Aut(M)\rho:G\to\Aut(M) preserve faithfulness?

Therefore: Faithful representations are exactly those where the group action is injective.

Recall that preserving the group action means that every element of the group is identified with a distinct action. In particular, that means there is only one element that behaves like the identity action: the identity element e∈Ge\in G. But if the representation ρ:G→Aut(Rn)\rho:G\to\Aut(R^n) maps only ee to the identity in Aut(Rn)\Aut(R^n), then it is injective, i.e. faithful.

Therefore, quotienting MM preserves faithfulness of ρ\rho exactly when the quotient M/WM/W preserves the injectivity of the group action on MM. There are two requirements to this:

So we can quotient MM by any R[G]R[G]-submodule WW to get a subrepresentation of M=ρM=\rho, but the resulting subrepresentation is only faithful if WW intersects trivially with ker ρ\ker\rho. In theory, by quotienting MM repeatedly by a maximal R[G]R[G]-submodule (whether it intersects trivially with ker ρ\ker\rho or not), we can “factor” MM into a direct sum of submodules known as irreducible representations, or irreps.

Lemma: For a finite group GG, if WW is a GG-invariant submodule of the R[G]R[G]-module MM, then M/WM/W is another GG-invariant submodule with M≅M/W⊕WM\iso M/W\oplus W, assuming char R∤∣G∣\char R\nmid |G|.
  • Recall that if a projection exists on MM, then MM is isomorphic to ker σ⊕im σ\ker\sigma\oplus\im\sigma. So it is enough to define a projection σ:M→M\sigma:M\to M where ker σ≅M/W\ker\sigma\iso M/W and im σ=W\im\sigma=W. Let’s see what those conditions imply.
    • To ensure that ker σ≅M/W\ker\sigma\iso M/W, we need to ensure that every element in ker σ\ker\sigma differs by some element in WW.
    • The requirement im σ=W\im\sigma=W implies that σ\sigma maps MM to the GG-invariant submodule WW of MM. Thus we need σ\sigma to be GG-equivariant (gσ(m)=σ(gm)g\sigma(m)=\sigma(gm)) so that it preserves GG-invariance.
    • Finally, to be a projection, σ\sigma must be idempotent.
  • We’ll start with the requirement that σ\sigma is GG-equivariant. In order to get GG-equivariance, one trick is to take the sum of all products with gg: σ~(m)=∑g∈Ggm\tilde{\sigma}(m)=\sum_{g\in G}gm Then we can show that for h∈Gh\in G: hσ~(m)=h∑g∈Ggm=∑g∈Ghgm by linearity in R[G]-modules=∑g∈Ggm since g↦hg is an automorphism on G=σ~(hm)\begin{aligned} h\tilde{\sigma}(m)&=h\sum_{g\in G}gm\\ &=\sum_{g\in G}hgm&\text{ by linearity in }R[G]\text{-modules}\\ &=\sum_{g\in G}gm&\text{ since }g\mapsto hg\text{ is an automorphism on }G\\ &=\tilde{\sigma}(hm) \end{aligned} where ∑g∈Ghgm=∑g∈Ggm\sum_{g\in G}hgm=\sum_{g\in G}gm because the act of multiplying every element of the sum by hh just permutes the order of the sum. Therefore, σ~\tilde{\sigma} commutes with the action of GG on MM, and is therefore GG-equivariant.
  • Next up is ensuring that the elements of ker σ\ker\sigma differ by elements of WW. One way is to send every gmgm to WW via some projection π:M→W\pi:M\to W. We redefine σ~\tilde{\sigma}: σ~(m)=∑g∈Gπ(gm)\tilde{\sigma}(m)=\sum_{g\in G}\pi(gm) Then σ~\tilde{\sigma} is a linear combination of elements of WW. This means that im σ~\im\tilde{\sigma} is a subset of WW, and we can show that two elements of ker σ~\ker\tilde{\sigma} differ by π(gm)−π(gn)∈W\pi(gm)-\pi(gn)\in W: σ~(m)−σ~(n)=∑g∈Gπ(gm)−∑g∈Gπ(gn)=∑g∈Gπ(gm)−π(gn)\begin{aligned} \tilde{\sigma}(m)-\tilde{\sigma}(n)&=\sum_{g\in G}\pi(gm)-\sum_{g\in G}\pi(gn)\\ &=\sum_{g\in G}\pi(gm)-\pi(gn) \end{aligned}
  • Finally, σ\sigma must be a projection. Therefore, it must be idempotent, and every element of WW must be mapped to. Without losing the previous properties, we can have σ(w)=w\sigma(w)=w for every w∈Ww\in W by taking the inverse action of GG, and dividing by ∣G∣|G| (which works since char R∤∣G∣\char R\nmid |G|): σ(m)=1∣G∣∑g∈Gg−1π(gm)\sigma(m)=\frac{1}{|G|}\sum_{g\in G}g^{-1}\pi(gm) because σ(w)=1∣G∣∑g∈Gg−1π(gw)=1∣G∣∑g∈Gg−1π(w′) since W is G-invariant=1∣G∣∑g∈Gg−1w′ since π is a projection onto W=1∣G∣∑g∈Gw since w′=gw=1∣G∣∣G∣w=w\begin{aligned} \sigma(w)&=\frac{1}{|G|}\sum_{g\in G}g^{-1}\pi(gw)\\ &=\frac{1}{|G|}\sum_{g\in G}g^{-1}\pi(w')&\text{ since }W\text{ is }G\text{-invariant}\\ &=\frac{1}{|G|}\sum_{g\in G}g^{-1}w'&\text{ since }\pi\text{ is a projection onto }W\\ &=\frac{1}{|G|}\sum_{g\in G}w&\text{ since }w'=gw\\ &=\frac{1}{|G|}|G|w\\ &=w \end{aligned} Therefore σ(w)=w\sigma(w)=w for all w∈Ww\in W, meaning W⊆im σW\subseteq\im\sigma. By definition, σ(m)\sigma(m) remains a linear combination of elements of WW, and therefore im σ⊆W\im\sigma\subseteq W. Therefore im σ=W\im\sigma=W.
  • Finally, since σ\sigma is an idempotent endomorphism, we know that MM decomposes into ker σ⊕im σ\ker\sigma\oplus\im\sigma. Since ker σ≅M/W\ker\sigma\iso M/W and im σ=W\im\sigma=W by construction, we have M≅M/W⊕WM\iso M/W\oplus W.

By applying this construction inductively, we can represent MM as a direct sum of irreps. Note that the construction requires assuming char R∤∣G∣\char R\nmid |G|, so this is only possible if char R∤∣G∣\char R\nmid |G|.

Maschke’s Theorem: Every representation ρ\rho of a finite group GG (over a field with characteristic not dividing ∣G∣|G|) is a direct sum of irreducible representations.
  • If the R[G]R[G]-module MM corresponding to the representation ρ\rho doesn’t reduce into irreps, there is a proper GG-invariant submodule WW of MM.
  • By the above theorem, which we can use since char R∤∣G∣\char R\nmid |G|, there is some complementary GG-invariant submodule W′W' such that M≅W⊕W′M\iso W\oplus W'.
  • By recursively decomposing proper GG-invariant submodules WW of each of the factors, you get smaller and smaller submodules. Since GG is finite, this process eventually stops when you are left with a direct sum of irreps that is isomorphic to the given representation ρ\rho.

A RR-module that can be written as a direct sum of simple modules is semisimple. Hence:

Corollary: A R[G]R[G]-module is semisimple if GG is finite and char R∤∣G∣\char R\nmid |G|.

In this section, we try to find irreps of any matrix representation.

Now that we know by Maschke’s theorem that we can factor any finite group representation into irreps (assuming char R∤∣G∣\char R\nmid |G|), let’s do so for matrix representations (which are always finite).

A matrix representation G→Aut(Rn)G\to\Aut(R^n) is special compared to more general representations G→Aut(M)G\to\Aut(M) because it implies that the module M=RnM=R^n is a free module. This simplifies finding irreps considerably.


In this section, we find an easier way to discover the eigenvalues in an algebraically closed field.

Find the eigenvalues of a given RR-matrix, where RR is a PID.

I know that the characteristic equation is used to find eigenvalues in vector spaces, which are defined over fields. What happens if we move to the world of modules defined over integral domains?

< Back to category Exploration 3: Representation theory (permalink)
Exploration 2: Module actions Exploration 4: Character theory