26
$\begingroup$

I was always bothered by the definition of the cross product given in e.g. a calculus course because it's never made clear how one would go about defining the cross product in a coordinate-free manner. I now know, not one, but two ways of doing this, and I can't quite see how they're related:

  • The cross product is the Lie bracket in the Lie algebra of $\text{SO}(3)$.
  • The cross product is the Hodge star map $\Lambda^2(V) \to V$ where $V$ is an oriented $3$-dimensional real inner product space.

Okay, so there's one obvious relation here: $V$ has automorphism group $\text{SO}(3)$. But for some reason I can't figure out where to go from here. A good starting point would be to exhibit a canonical isomorphism between an oriented $3$-dimensional inner product space $V$ and the Lie algebra of $\text{Aut}(V)$. Maybe this is obvious. In any case, I would appreciate some clarification.

$\endgroup$
3
  • 10
    $\begingroup$ The Lie algebra $so(V)$ is naturally isomorphic to $\Lambda^2(V).$ $\endgroup$ Jul 30, 2010 at 7:41
  • $\begingroup$ Thanks, Victor. I guess I should've thought about this a little more. $\endgroup$ Jul 30, 2010 at 17:24
  • $\begingroup$ The ways in which I first heard the cross-product defined was coordinate free, albeit deficient in logical rigor: to find $\mathbf a\times\mathbf b,$ let the fingers of your right hand curl in the direction in which you rotate when going from $\mathbf a$ to $\mathbf b$, and then the direction your thumb points tells you which of two directions orthogonal to $\mathbf a$ and $\mathbf b$ is the right one, and then the norm is the area of the rectangle whose sides are $\mathbf a$ and $\mathbf b.$ $\endgroup$ Oct 22, 2017 at 19:46

3 Answers 3

23
$\begingroup$

To expand on Victork Protsak's comment, if $V$ is an $n$-dimensional real vector space with inner-product, the inner-product gives an isomorphism $V\to V^*$ and hence $V\otimes V \to \mathrm{End}(V)$. Under this isomorphism, $\Lambda^2(V)$ is identified with skew-adjoint endomorphisms of $V$, which is precisely the Lie algebra $\mathfrak{so}(V)$.

In the case $\dim V =3,$ the Hodge star gives an isomorphism $\Lambda^2(V) \to V$ and so in total we see that $V$ is canonically isomorphic to $\mathfrak{so}(V)$. A more direct way to see this isomorphism is to send the vector $v \in V$ to the generator of the right-handed rotation about the axis in the direction of $v$ with speed $|v|$.

The use of the phrase "right-handed" makes it clear that in order to identify $V$ and $\mathfrak{so}(V)$ we have used an orientation on $V$; indeed, you need that for the Hodge star. What is interesting is that if you reverse the orientation on $V$, the map to $\mathfrak{so}(V)$ changes sign. This means that what ever orientation you chose on $V$, the push-forward to $\mathfrak{so}(V)$ is the same. Conclusion: $\mathfrak{so}(3)$ is naturally oriented. This is analogous to the natural orientation on $\mathbb{C}$. A more prosaic way to describe the orientation is to pick two independent elements $x,y \in \mathfrak{so}(3)$ and then use $[x,y]$ to complete them to an oriented basis. (Of course, you then need to check that this doesn't depend on your choice of $x,y$.)

$\endgroup$
2
  • $\begingroup$ No worries. The fact that so(3) has a natural orientation actually came up in something I was working on recently. Some objects I was looking at had a sign associated to them and it turned out that hidden in the background was an isomorphism between two bundles of so(3) lie algebras; the sign was just whether this isomorphism preserved or reversed the orientation. $\endgroup$
    – Joel Fine
    Jul 30, 2010 at 19:04
  • 1
    $\begingroup$ Why is this answer accepted? Sure, it does explain how $so(V)$ is canonically isomorphic to $V$ for a 3-dimensional $V$ as vector spaces. But why is the Lie bracket in $so(V)$ the same as the wedge product followed by the Hodge star in $V$? $\endgroup$
    – Kostya_I
    Jul 6, 2018 at 22:31
18
$\begingroup$

Let $\varepsilon( )$ be the volume form in $\mathbb R^3$. For given vectors ${\bf p}$ and ${\bf q}$ the function $f:{\bf x}\mapsto\varepsilon({\bf p},{\bf q},{\bf x})$ is a linear functional and so is represented by a vector ${\bf r}\in\mathbb R^3$, i.e., one has $f({\bf x})=\langle{\bf r},{\bf x}\rangle$. This vector ${\bf r}$ depends in a skew bilinear way from ${\bf p}$ and ${\bf q}$ and is called the $vector\ product$ of ${\bf p}$ and ${\bf q}$.

$\endgroup$
1
  • 8
    $\begingroup$ Yes, I think this answer has the advantage that it uses precisely the assumptions present in a calculus course, and no extra machinery. This means that you can (more or less) explain it to calculus students. I am teaching a multi-variable calc class this summer, and have them derive the formula for $2\times 2$ and $3\times 3$ determinant from signed area and volume. From here, one can give an honest definition of cross product (rather I gave them the computational one earlier, so they could do their physics homework, and now I can revisit that with a proper explanation) $\endgroup$ Jul 30, 2010 at 12:37
3
$\begingroup$

Geometric (Clifford) algebra provides yet another way to understand the cross product. If a vector represents a 1D subspace containing the origin, the wedge (outer) product is a bilinear form on 2 vectors ($a \wedge b$) that represents the subspace containing both vectors. The vector has basis $e_1, e_2, ..., e_n$ ($n$ elements), so the wedge product has basis $e_1\wedge e_2, e_1\wedge e_3, \ldots, e_{n-1}\wedge e_n$ ($\binom{n}{2}$ elements). The cross product is a way to represent this subspace (using its dual, the normal vector). This works only for 3D, where the basis sizes match, using $e_1 e_2 \to e_3,\, e_2 e_3 \to e_1,\, e_3 e_1 \to e_2$.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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