3
$\begingroup$

I tried this question twice on math.stackexchange but got no answer so I decided to move it here.

Let $M$ be a smoth manifold. Then $$C^\infty(M):=\{f:M\longrightarrow \mathbb R; f\ \textrm{is smooth}\},$$ is an $\mathbb R$-algebra with the pointwise product.

If you don't know anything about smooth manifolds it doesn't matter, all that you need to know is that $C^\infty(M)$ is an $\mathbb R$-algebra for all that follows is purely algebraic.

Let us define the space of vector fields on $M$ by: $$\mathfrak{X}(M):=\mathsf{der}_{\mathbb R}\ C^\infty(M):=\{X\in \mathsf{End}_{\mathbb R}(C^\infty(M)): X(fg)=fX(g)+X(f)g\}.$$ This is a $C^\infty(M)$-module with $$(f\cdot X)(g):=f X(g),$$

where $fX(g)$ is the pointwise product of the functions $f$ and $X(g)$.

We can then define the space of $p$-forms on $M$ as the $C^\infty(M)$-module:

$$\Omega^p(M):=\mathsf{Hom}_{C^\infty(M)}(\Lambda^p \mathfrak{X}(M), C^\infty(M)).$$ The De Rham differential on $M$ is the degree one operator $$d: \Omega^p(M)\longrightarrow \Omega^{p+1}(M)$$ given by $$\begin{eqnarray*} d\varepsilon(X_{1}, \ldots, X_{p+1})&&:=\sum_{\sigma\in\mathsf{Sh}(1, p)} \mathsf{sgn}(\sigma) X_{\sigma(1)}(\varepsilon(X_{\sigma(2)}, \ldots, X_{\sigma(p+1)})\\ &&+\sum_{\sigma\in\mathsf{Sh}(2, p-1)}\mathsf{sgn}(\sigma) \varepsilon([X_{\sigma(1)}, X_{\sigma(2)}], X_{\sigma(3)}, \ldots, X_{\sigma(p+1)}), \end{eqnarray*}$$ where $[\cdot, \cdot]$ is the commutator of vector fields which is defined by $$[X, Y]:=X\circ Y-Y\circ X.$$

Notation: For integers $p, q\geq 1$ let us write $S(p, q)$ as the subset of permutations $\sigma$ of the set $\{1, \ldots, p+q\}$ such that $\sigma(1)<\ldots< \sigma(p)$ and $\sigma(p+1)<\ldots< \sigma(p+q)$. The elements of $\mathsf{Sh}(p, q)$ are known as $(p, q)$-shufles for obvious resons.

Now, we can define a product $$\wedge: \Omega^p(M)\times \Omega^q(M)\longrightarrow \Omega^{p+q}(M)$$ setting $$(\varepsilon\wedge \eta)(X_1, \ldots, X_{p+q}):=\sum_{\sigma\in\mathsf{Sh}(p, q)} \mathsf{sgn}(\sigma) \varepsilon(X_{\sigma(1)}, \ldots, X_{\sigma(p)}) \eta(X_{\sigma(p+1)}, \ldots, X_{\sigma(p+q)}).$$ Can anyone help me to prove $$d(\varepsilon\wedge \eta)=d\varepsilon\wedge \eta+(-1)^p \varepsilon\wedge d\eta,$$ for every $\varepsilon\in \Omega^p(M)$ and $\eta\in \Omega^q(M)$?

I know the property I want to show is a classical one but I can't seem to find the proof using this algebraic formulation. I've already asked this question before and got no answer.

However, by that time things were more obscure so I decided to update with this improved version hoping someone could help me.

Remark: 1) The cardinality of the set $\mathsf{Sh}(p, q)$ is $\binom{p+q}{q}$. Hence we easily see there are bijections:

$$\mathsf{Sh}(1, p+q)\times \mathsf{Sh}(p, q)\longrightarrow \mathsf{Sh}(p+1, q)\times \mathsf{Sh}(1, p)$$

and

$$\mathsf{Sh}(1, p+q)\times \mathsf{Sh}(p, q)\longrightarrow \mathsf{Sh}(p, q+1)\times \mathsf{Sh}(1, q).$$

So the problem boils do to writing those bijections explicitly.

$\endgroup$
8
  • 3
    $\begingroup$ Have you tried induction on degrees? $\endgroup$
    – Bernie
    Dec 29, 2017 at 1:01
  • 3
    $\begingroup$ Also, isn't it easier to write everything in terms of the full symmetric group instead of shuffles? $\endgroup$
    – Deane Yang
    Dec 29, 2017 at 5:33
  • $\begingroup$ I guess one can write a rather tedious direct proof, starting from the observation that in the first term of the definition of $d$, $X_{\sigma(1)}$ either differentiates the first or the second factor (because it is a derivation). And in the second term, inserting the commutator happens in the first or in the second factor (because inserting a vector field is a derivation, too). Grouping the terms in $d(\varepsilon\wedge\eta)$ correspondingly should prove the claim. $\endgroup$ Dec 29, 2017 at 11:14
  • $\begingroup$ @DeaneYang maybe, I'll give it a try, thanks for the hint =) $\endgroup$
    – PtF
    Dec 29, 2017 at 14:35
  • $\begingroup$ The problem boils down to establishing the bijections$\mathsf{Sh}(1, p+q)\times \mathsf{Sh}(p, q)\simeq \mathsf{Sh}(p+1, q)\times \mathsf{Sh}(1, p)$ and $\mathsf{Sh}(1, p+q)\times \mathsf{Sh}(p, q)\simeq \mathsf{Sh}(p, q+1)\times \mathsf{Sh}(1, q)$ and this really. $\endgroup$
    – PtF
    Dec 29, 2017 at 18:03

0

Your Answer

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