21

I'm beginning a course that uses representation theory, but I do not really understand what that is about. In the text I am following, I have the following definition:

A representation of the Lie group $G$ on the vector space $V$ is a continuous mapping $\cdot \colon G \times V \to V$ such that

  1. for each $g \in G$, the translation $T_{g} \colon V \to V$ given by $T_g(v) = g \cdot v$, $v \in V$, is a linear map;
  2. $T_{e} = \mathrm{Id}$ where $e$ is the identity element of $G$;
  3. $T_{gh} = T_{g} T_{h}$ for $g, h \in G$. We call the pair $(V,\cdot)$ a real (resp. complex) representation and $V$ the representation space.

What is the motivation behind this sort of definition? From my google searches I have seen different definitions, but I still don't really know why what I am reading is defined that way. Why a Lie group and not a regular group? etc.

Hanul Jeon
  • 27,376
diffGeoLost
  • 223
  • 2
  • 4
  • I feel like that is a hard definition to get your head around. Two equivalent definitions are: 1) A representation is a (continuous) homomorphism $G\to \mathrm{GL}(V)$ - i.e. a way of viewing $G$ or some quotient of it as a subgroup of a group of linear maps. Or 2) A (continuous) action of $G$ on a vector space $V$ that preserves the vector space structure. We know that group actions are important - acting on a vector space gives us additional structure. This is your definition, but the fact that this just defines a group action is not made explicit. – Mathmo123 Jan 27 '16 at 00:32
  • My professor 1) Move the class along at a very (insanely) quick pace; 2) Is not known for giving intuitive definitions. Thanks for the helping. Why do you say "or" when listing those two definitions? Do the definitions actually mean the same thing said in different ways or are they really two different definitions with the same name? – diffGeoLost Jan 27 '16 at 00:36
  • They are equivalent. A group action assigns to each $g\in G $ an invertible linear map from $V\to V$ - I.e. an element of $\mathrm{GL}(V)$. If you've never seen representation theory before, it seems strange to start with Lie groups. A first course on representation theory would start with representations of a finite group on a vector space whose underlying field has characteristic $0$. Already in this case, representation theory gives us extra structure on the group, and allows us to prove some thms that we don't know how to prove otherwise. I'd suggest finding a book or some notes on this – Mathmo123 Jan 27 '16 at 00:42
  • 2
    To answer your question "why a Lie group", the answer is that the definition works for any group. In fact, there are some groups which we only really know how to study via their representations. A representation allows us to take a group we may know very little about and study it using the familiar tools of linear algebra. – Mathmo123 Jan 27 '16 at 00:45
  • Thank you for the information! I encourage you to add your comments here as an Answer! – diffGeoLost Jan 27 '16 at 00:47
  • I was planning to write an answer, but it's quite late in my time zone. I'll see how it goes tomorrow morning. – Mathmo123 Jan 27 '16 at 00:48
  • 2
    This is worth reading. – Mathmo123 Jan 27 '16 at 22:38
  • Have you searched this site? This has been discussed before. It is a waste to duplicate effort :-( – Mariano Suárez-Álvarez Jan 27 '16 at 23:17
  • 1
    @MarianoSuárez-Alvarez A cursory look suggests that there are several questions on the applications of rep theory, but none that I can find asking why we would bother to define representations in the first place. This question is the closest, but seems to be asking more from a perspective of looking for applications from a position of knowledge. Of course I may have missed something. – Mathmo123 Jan 27 '16 at 23:20
  • 1
    I tried searching 'motivation representation theory', but what I found didn't really satisfy what I was looking for. The answer given by @Mathmo123 has helped immensely with clearing up my confusion and hopefully will clear up the confusion of my classmates and anyone else. – diffGeoLost Jan 27 '16 at 23:48
  • 1
    @Mathmo123, the asnwers here and IMO the question are subsets of those at the second question you linked. – Mariano Suárez-Álvarez Jan 28 '16 at 00:13
  • This webpage might be useful in answering the question: https://www.maths.ox.ac.uk/node/28902 – vyali Sep 06 '23 at 07:58

2 Answers2

23

One of the most common ways that groups arise "in the wild" is as sets of symmetries of an object. For example

  • The symmetry group $S_n$ is the group of all permutations of $\{1,\ldots n\}$
  • The dihedral group $D_{2n}$ is the group of symmetries of a regular $n$-gon
  • The Lie group $\mathrm{GL}_n(\mathbb R)$ is the group of invertible linear maps on $\mathbb R^n$

More generally, given a general abstract group $G$, we regularly consider the case of $G$ acting on a set $X$, and we might ask the question: given some set $X$, what is its "group of symmetries".

Representation theory asks the converse to this question:

Given a group $G$, what sets does it act on?

Whilst it is possible to attempt to answer this general, a useful starting point is to restrict the sets in question to sets we know an awful lot about: vector spaces.

Definition: Let $G$ be a group, and $V/k$ be a vector space. A representation of $G$ is a group action of $G$ on $V$ that is linear (so preserves the vector space structure of $V$) - i.e. for every $g\in G$, $u,v\in V$, $\mu,\lambda\in k$ $$g(\lambda u+\mu v) = \lambda g(u) + \mu g(v).$$

This is the definition that you have been given. With $V$ as before, an equivalent definition is this:

A representation of $G$ is a group homormophism $$\rho: G\to\mathrm{GL}(V)$$

Indeed, a group action of $G$ on $V$ assigns to each $g$ an invertible linear map. And given a homomorphism $\rho$, $G$ acts on $V$ via $g\cdot v = \rho(g)v$.

In the case that $G$ is a Lie group (or more generally a topological group), then we require this action/representation to be continuous.

Representation theory allows us to translate our viewpoint by viewing (a quotient of) our group as a group of linear maps on a vector space. This allows us to tackle problems in group theory using the familiar and powerful tools of linear algebra. For example, we can take the trace of a linear map, and the identity $\mathrm{tr}(ABA^{-1}) = \mathrm{tr}(B)$ tells us that the trace of a representation (called the character of the representation) is constant on the conjugacy classes of a group. We can also consider determinants, characteristic polynomials, dual vector spaces (or the dual representation), dimension and many more of our favourite concepts from linear algebra.

Representations are certainly powerful:

  • There are theorems (for example, concerning Frobenius groups) whose only known proofs use representation theory.
  • There are groups (such as $\mathrm{Gal}(\overline{\mathbb Q}/\mathbb Q)$) which we only really know how to study via their representations.
durianice
  • 2,624
Mathmo123
  • 23,018
4

Rough answer. Before there was an abstract definition of "Lie group" mathematicians studied groups of matrices. A Lie group is a generalization of a group of matrices. It turns out that one way to try to understand a Lie group is to look at all the ways to "represent" it as a group of matrices. Each mapping in the definition in your question does just that.

Ethan Bolker
  • 95,224
  • 7
  • 108
  • 199