Abstract Nonsense

Crushing one theorem at a time

Subrepresentations, Direct Sum of Representations, and Irreducible Representations (Pt. II)

Point of post: This post is a continuation of this one.

Invariant Subspaces and Subrepresentations


Let \mathscr{V} be a finite-dimensional pre-Hilbert space and G a finite group. Let \rho:G\to\mathcal{U}\left(\mathscr{V}\right) be a representation. Then, we call \mathscr{W}\leqslant \mathscr{V} invariant under \rho if \mathscr{W} is invariant under \rho_g in the normal sense for each g\in G. It’s clear then that the map


\omega:G\to\mathcal{U}\left(\mathscr{W}\right):g\mapsto \left(\rho_g\right)_{\mid\mathscr{W}}


is a representation of G on \mathscr{W}. Such a representation is called a subrepresentation and is denoted \rho_{\mid\mathscr{W}} or \rho\upharpoonright \mathscr{W}. Our first result concerning invariant subspaces and subrepresentations is a fundamental one. Namely:


Theorem: Let \mathscr{V} be a finite dimensional pre-Hilbert space with inner product \langle\cdot,\cdot\rangle and \rho:G\to\mathcal{U}\left(\mathscr{V}\right) a representation of G on \mathscr{V}. Suppose that \mathscr{W}\leqslant \mathscr{V} is invariant under \rho then so is


\mathscr{W}^{\perp}=\left\{v\in\mathscr{V}:\langle v,w\rangle=0\text{ for all }w\in\mathscr{W}\right\}


Moreover, \rho\simeq \rho_{\mid\mathscr{W}}\oplus\rho_{\mid\mathscr{W}^{\perp}}.

Proof: Fix x\in\mathscr{W}^{\perp}, and let w\in\mathscr{W} and g\in G be arbitrary. We note then that


\displaystyle \begin{aligned}\left\langle \rho_g(x),w\right\rangle &= \left\langle x,\rho_g^\ast(w)\right\rangle\\ &= \left\langle x,\rho_g^{-1}(w)\right\rangle\\ &= \left\langle x,\rho_{g^{-1}}(w)\right\rangle\\ &= 0\end{aligned}


and since w,g were arbitrary it follows that \rho_g(x)\in\mathscr{W}^{\perp} for every g\in G. But, by the arbitrariness of x we may conclude that \rho_g(x)\in\mathscr{W}^{\perp} for every g\in G and x\in \mathscr{W}^{\perp}. Said differently \mathscr{W}^{\perp} is invariant under \rho.


Now, to prove that \rho\simeq \rho_{\mid\mathscr{W}}\oplus\rho_{\mid\mathscr{W}^{\perp}} we define


U:\mathscr{V}\to\mathscr{W}\oplus\mathscr{W}^{\perp}:w+w'\mapsto (w,w')


where w+w' is the unique representation of every element of \mathscr{V} as the sum of an element of \mathscr{W} and an element of \mathscr{W}^{\perp} (this comes from the elementary fact that \mathscr{V}=\mathscr{W}\oplus\mathscr{W}^{\perp} where now the \oplus denotes the internal direct sum). This is evidently a unitary map since (recalling that \langle w_1,w'_2\rangle=\langle w'_1,w_2\rangle=0 by definition)


\begin{aligned}\left\langle U\left(w_1+w'_1\right),U\left(w_2+w'_2\right)\right\rangle &= \left\langle (w_1,w'_1),(w_2,w'_2)\right\rangle\\ &= \left\langle w_1,w_2\right\rangle +\left\langle w_1,w'_2\right\rangle+\left\langle w'_1,w_2\right\rangle+\left\langle w'_1,w'_2\right\rangle\\ &= \left\langle w_1+w'_1,w_2+w'_2\right\rangle\end{aligned}


We then note that for every g\in G and w+w'\in\mathscr{V} we have that


\begin{aligned}U\rho_g(w+w') &=U(\rho_g(w)+\rho_g(w'))=U(\left(\rho_g\right)_{\mid\mathscr{W}}(w)+\left(\rho_g\right)_{\mid\mathscr{W}^{\perp}}(w')\\ &=\left(\left(\rho_g\right)_{\mid\mathscr{W}}(w),\left(\rho_{g}\right)_{\mid\mathscr{W}^{\perp}}(w')\right)\\ &=\left(\left(\rho_g\right)_{\mid\mathscr{W}}\oplus\left(\rho_g\right)_{\mid\mathscr{W}^{\perp}}\right)U\left(w+w'\right)\\ &= \left(\rho_{\mid\mathscr{W}}\oplus\rho_{\mid\mathscr{W}^{\perp}}\right)_gU\left(w,w'\right)\end{aligned}


from where it follows that \rho\simeq\rho_{\mid\mathscr{W}}\oplus\rho_{\mid\mathscr{W}^{\perp}}. \blacksquare



We also have somewhat of a converse of this theorem:


Theorem: Let \mathscr{V} be a finite dimensional pre-Hilbert space and G a finite group. Then, if \rho:G\to\mathcal{U}\left(\mathscr{V}\right) is a representation of G on \mathscr{V} such that \rho\simeq \omega_1\oplus\omega_2 where








then if U:\mathscr{W}\oplus\mathscr{U}\to\mathscr{V} is the guaranteed unitary isomorphism such that U^{-1}\rho_gU=\left(\omega_1\oplus\omega_2\right)_g for every g\in G then U\left(\mathscr{W}\times\{e\}\right) is invariant under \rho.

Proof: Let U((w,e))\in U\left(\mathscr{W}\times\{e\}\right) then we have that \rho_g(U(w,e))=U\left(\left(\omega_1\oplus\omega_2\right)_g(w,e)\right)\in U\left(\mathscr{W}\times\{e\}\right). Since g\in G and U((w,e))\in U\left(\mathscr{W}\times\{e\}\right) were arbitrary the conclusion follows. \blacksquare



1.Simon, Barry. Representations of Finite and Compact Groups. Providence, RI: American Mathematical Society, 1996. Print.

2. Serre, Jean Pierre. Linear Representations of Finite Groups. New York: Springer-Verlag, 1977. Print.


January 18, 2011 - Posted by | Algebra, Representation Theory | , , , , ,


  1. […] Point of post: This post is a continuation of this one. […]

    Pingback by Representation Theory: Subrepresentations, Direct Sum of Representations, and Irreducible Representations (Pt. III) « Abstract Nonsense | January 18, 2011 | Reply

  2. […] hopes would be that if and are irreps that is also an irrep, but this is not true. But, by first principles we know […]

    Pingback by Representation Theory: The Tensor Product and the Tensor Product of Representations « Abstract Nonsense | February 14, 2011 | Reply

  3. […] of with characters and respectively. Then, if is the tensor product of and then . If is the direct sum of and then […]

    Pingback by Representation Theory: General Characters and the Uniqueness of Decomposition Into Irreps « Abstract Nonsense | February 27, 2011 | Reply

  4. […] are two important things we make note of. Namely, since the proof for the complex representation case did not actually depend on the fact that representations were complex and can easily be repeated […]

    Pingback by Representation Theory: Representations on Real and Quaternionic Vector Spaces (Examples and Basics) « Abstract Nonsense | March 28, 2011 | Reply

  5. […] Said more concretely for each we define to be equal to and consider the space given the usual inner product structure on direct sum spaces with a normalization factor of –we will think of the elements of this […]

    Pingback by Another Way of Looking at Induced Representations (Pt. I) « Abstract Nonsense | April 24, 2011 | Reply

  6. […] let be any group and to representations on . We call and disjoint if they have no equivalent subrepresentations and we denote this property by . This notation is supposed to be provocative since it’s […]

    Pingback by Mackey Irreducibility Criterion « Abstract Nonsense | May 6, 2011 | Reply

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: