I am trying to understand the relations and differences between integral of differential form and integral of measure. From Wikipedia:

On a general differentiable manifold (without additional structure), differential forms cannot be integrated over subsets of the manifold; this distinction is key to the distinction between differential forms, which are integrated over chains, and measures, which are integrated over subsets.

  1. Isn't a chain a manifold, and therefore a subset of a manifold? Why is that "differential forms cannot be integrated over subsets of the manifold"?
  2. Is integral of a differential form defined in terms of Lebesgue integral via parameterization of the chain, as like here

    $$\int_S \omega =\int_D \sum a_{i_1,\dots,i_k}(S({\mathbf u})) \frac{\partial(x^{i_1},\dots,x^{i_k})}{\partial(u^{1},\dots,u^{k})}\,du^1\ldots du^k $$

    So can one say that integral of a differential form is not a different integration method from Lebesgue integral?

Thanks and regards!

  • 4
    $\begingroup$ Related but not identical: math.stackexchange.com/q/49641/1543 $\endgroup$ – Willie Wong Aug 15 '11 at 15:48
  • 2
    $\begingroup$ @Tim: A chain is neither a manifold nor a submanifold. Rather, it is a formal sum of oriented submanifolds, possibly mutually intersecting (or with other pathologies). $\endgroup$ – Zhen Lin Aug 15 '11 at 16:04
  1. Differential forms cannot be integrated over all subsets of a manifold. You cannot integrate a 2-form over a curve, or a 1-form over a surface. More precisely: integration in measure space is defined over a $\sigma$-algebra. So if integration is defined on two sets $A$ and $B$, the integral is also defined on $A\cap B$. But the same is not true for integration of differential forms. Given a two form $\omega$ in $\mathbb{R}^3$, and take $S$ and $T$ be two smooth surfaces that intersects transversally. $\int_S\omega$ and $\int_T\omega$ make sense, but not $\int_{S\cap T}\omega$.

  2. Generally a differential form, by definition, when taken in a local coordinate representation, is given by functions that are at least continuous. So in fact you don't need the machinery of Lebesgue integration: all the functions you are going to deal with can be treated with Riemann integration.


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.