3
$\begingroup$

It is well known that the regular representation of $\mathfrak{so}(3)$ is the so-called "cross product" matrix $A(x)$ which follows $A(x)y = x\times y$, and $x,y\in\mathbb{R}^3$, while the cross product is also connected to the exterior product in $\mathbb{R}^3$. This begs the following two questions:

(1) How are $\mathfrak{so}(3)$ and $\Lambda^2(\mathbb{R}^3)$ related?

(2) Is there a general theory connecting exterior algebras and Lie algebras, or is this just a strange coincidence?

$\endgroup$
3
$\begingroup$

I don't know what you mean by the regular representation of a Lie algebra, but the connection is the following. It's a bit confusing because there are a bunch of $3$-dimensional vector spaces that all get identified.

Let $V$ be a real inner product space. The inner product induces a canonical isomorphism $V \cong V^{\ast}$ giving an isomorphism

$$\mathfrak{gl}(V) \cong V \otimes V^{\ast} \cong V \otimes V.$$

The subalgebra $\mathfrak{so}(V) \subset \mathfrak{gl}(V)$ gets identified with the antisymmetric tensors inside $V \otimes V$, which I'll write as $\Lambda^2(V)$ even though this usually refers to a quotient of $V \otimes V$ and not a subspace. At this level of generality the Lie bracket has nothing to do with wedge products, as far as I know.

However, if $V$ is furthermore $3$-dimensional and equipped with an orientation, then together with the inner product we get a trivialization $\Lambda^3(V) \cong \mathbb{R}$, and hence the wedge product gives a nondegenerate bilinear form

$$V \otimes \Lambda^2(V) \to \Lambda^3(V) \cong \mathbb{R}$$

inducing an isomorphism

$$\Lambda^2(V) \cong V^{\ast} \cong V.$$

(So, the $3$-dimensional vector spaces getting identified in this story: $V$, its dual, $\Lambda^2(V)$, its dual, and $\mathfrak{so}(V)$.) Up to maybe some factors of $2$, these isomorphisms intertwine the cross product, the wedge product, and the Lie bracket on $\mathfrak{so}(V)$.

$\endgroup$
3
  • $\begingroup$ Perfect! Thanks for the lesson. By the regular representation I am following R. Gilmore's terminology. If $\mathfrak{g}$ is a Lie algebra with basis $\{X_1,\dots,X_n\}$ and $Z$ is an arbitrary element in $\mathfrak{g}$, Gilmore refers to the matrix $R(Z)$ which obeys $[Z,X_i] = \sum_{j}R_{ij}X_j$ as the regular representation. $\endgroup$ – JMJ May 6 '16 at 17:24
  • 1
    $\begingroup$ @ALB: a more standard name for that (which does not require choosing a basis) is the adjoint representation. $\endgroup$ – Qiaochu Yuan May 6 '16 at 17:27
  • $\begingroup$ Yet another question of mine answered! @Qiaocho_Yuan, you are a gentleman and a scholar :-) P.S. I am also a fan of your blog. $\endgroup$ – JMJ May 6 '16 at 17:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.