Two ways to define differential forms

108 Views Asked by At

Let $U \subseteq \mathbb{R}^n$ be an open subset.
Then there are two ways to define a differential form:

  • algebraically as in this previous post, were one considers the real algebra $A$ of smooth functions $A$, the module of real derivations and then its exterior algebra;
  • geometrically, as smooth maps $$\omega \colon U \to \operatorname{Alt}^p(\mathbb{R}^n).$$

This second approach is used in most differential geometry books, such as Henri Cartan's Differential forms.
One possible advantage in the second approach is that one can consider $C^k$ maps $$\omega \colon U \subseteq E \to \operatorname{Alt}^p(E, F)$$ where $E$ and $F$ are Banach spaces. The apparent disadvantage is that one has to do a lot of work to give an algebraic structure to those maps.

I have never seen a discussion on the relation between those two definitions so here are some natural questions:

  1. Is it trivial to go from algebraic to geometric definition and back
    (in the particular case of $E = \mathbb{R}^n$, $F = \mathbb{R}$)?
  2. Is there an equivalence / adjunction?
  3. Has this something to do with the relation between bundles and lf. modules of sheaves (the algebro-geometric duality)?

If i understand correctly, the relation should be $$ \Lambda^p(\mathcal{D}^1(C^{\infty}(U, F))) = C^{\infty}(U, \operatorname{Alt}^p(E, F)) $$ where one shows that the second gadget has a well defined wedge product.
The first gadget represents the alternating functor, hence $$ \operatorname{Alt}^p(\mathcal{D}^1(C^{\infty}(U, F)), N) = \operatorname{Hom}(C^{\infty}(U, \operatorname{Alt}^p(E, F)), N) $$ for any $A$-module $N$ where $A = C^{\infty}(U, F)$.