2
$\begingroup$

I would like some help understanding one of the equalities in the following proof:

Let $f:\mathbb{R}^n\rightarrow \mathbb{R}^m$ be a differentiable function. If $\alpha^1,\cdots ,\alpha^k$ are $1$-forms in $\mathbb{R}^m$, prove that $f^*(\alpha^1\wedge\cdots \wedge\alpha^k)=f^*(\alpha^1)\wedge\cdots \wedge f^*(\alpha^k)$

Proof:

\begin{align}f^*(\alpha^1\wedge\cdots \wedge\alpha^k)(v_1,\cdots ,v_k)&=(\alpha^1\wedge\cdots \wedge \alpha^k)(df(v_1),\cdots ,df(v_k))\\&=\det(\alpha^i(df(v_j))\\&=\det(f^*\alpha^i(v_j))\\ &=f^*(\alpha^1)\wedge\cdots \wedge f^*(\alpha^k)(v_1,\cdots ,v_k) \end{align}

My question is, why is the second equality true. I know that the wedge product can be defined using determinants, but if we are dealing with differentials what does $\alpha^i(df(v_j))$ actually mean and why would we write out that way?

The definition I am given for the wedge product is:

$$\alpha^1\wedge\cdots \wedge \alpha^k=\sum_{i_1<...<i_k}^{}\frac{\partial (\alpha^1,\cdots,\alpha^k)}{\partial(e^{i^1},\cdots,e^{i_k})}e^{i_1}\wedge\cdots\wedge e^{i_k}$$

$\endgroup$
3
$\begingroup$

So as always in differential geometry, you have to keep track of what type your objects are.

Here notice that $df$ is a linear function $\mathbb{R}^n \rightarrow \mathbb{R}^m$ so it spits vectors of $\mathbb{R}^m$.

On the other hand $\alpha^i$ is a 1-forms so it takes a vector in $\mathbb{R}^m$ as argument. Thus it shoud be clear (at least formally) what $\alpha^i(df(v_j))$ means.

You can show (by induction for example, knowing that $(\alpha \wedge \beta) (V_1,V_2)=\alpha(V_1)\beta(V_2)-\alpha(V_2)\beta(V_1)$ for any vectors, that for any $V_1, \ldots,V_k$ in $\mathbb{R}^m$, $$\alpha^1 \wedge \ldots \wedge \alpha^k (V_1, \ldots ,V_k)=det(\alpha^i(V_j)).$$

(To me this was pretty much the definition of wedge products).

Hope this was helpful somehow.

$\endgroup$
3
$\begingroup$

It really just comes from this result: Let $\alpha^1,\cdots\alpha^k$ be linear functionals on a vector space $V$, and $v_1,\cdots v_k\in V.$ Then, \begin{align*} (\alpha^1\wedge\cdots\wedge \alpha^k)(v_1,\cdots, v_k)&=A(\alpha^1\otimes\cdots\otimes\alpha^k)(v_1,\cdots, v_k)\\ &=\sum\limits_{\sigma\in S_k}(\text{sgn }\sigma)\alpha^1(v_{\sigma(1)})\cdots\alpha^k(v_{\sigma(k)})\\ &=\det [\alpha^i(v_j)] \end{align*} Here, $A$ is the alternating operator $$Af=\sum\limits_{\sigma\in S_k}(\text{sgn }\sigma)\sigma f,$$ and, by definition, $$f\wedge g=\frac{1}{k!\ell!}A(f\otimes g)$$ for a $k$-form $f$ and $\ell$-form $g$. By $\sigma f$, we mean $$(\sigma f)(v_1,\cdots, v_k)=f(v_{\sigma(1)},\cdots, v_{\sigma(k)}).$$

Also, since $\alpha^i$ is a one-form and $df(v_j)$ is a tangent vector, $\alpha^i(df(v_j))$ is perfectly well-defined.

$\endgroup$

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.