MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

This question is similar to Do chains and cochains know the same thing about the manifold? in the sence that both deal with a natural "comparison" quasi-isomorphism that does not preserve the ring structure.

Let $M$ be a smooth manifold. There is a natural comparison map $Comp$ from the differential forms on $M$ to the smooth singular cochains of $M$ (i.e. the linear dual of the vector space spanned by smooth singular simplices). It is defined as follows: take a form $\omega$ of degree $p$ and set $Comp(\omega)$ to be the cochain $\sigma\mapsto \int_\triangle \sigma^*\omega$ where $\triangle$ is the standard $p$-dimensional simplex and $\sigma:\triangle\to M$ is a smooth singular simplex.

$Comp$ is a map of complexes (Stokes' theorem) and moreover, a quasi-isomorphism (the de Rham theorem). But as simple examples show, it does not preserve the ring structure. However it is natural to ask whether the ring structures, up to quasi-isomorphism, of the differential forms and of the cochains contain the same information about $M$. This translates into the following questions.

  1. Can $Comp$ be completed to a morphism of $A_\infty$-algebras?

  2. If the answer to 1. is positive (it presumably is), what about the $E_\infty$ case?

These questions also have natural rational versions. Namely, we can take an arbitrary polyhedron $X$ instead of $M$ and consider Sullivan's $\mathbf{Q}$-polynomial forms. There is a comparison quasi-isomorphism similar to the one above that will go from the $\mathbf{Q}$-polynomial forms of $X$ to the piecewise linear $\mathbf{Q}$-cochains. Can it be completed to a map of $A_\infty$ or $E_\infty$ algebras?

share|cite|improve this question

This theorem was proven in 1977 in "V. K. A. M. Gugenheim, On Chen’s iterated integrals, Illinois J. Math. Volume 21, Issue 3 (1977), 703–715."

share|cite|improve this answer


Here is one way to see it:

before passing to dg-algebras, let's look at cosimplicial algebras and then later apply the normalized cochain (Moore) complex functor.

Work in a smooth (oo,1)-topos, modeled by simplicial presheaves on a site of smooth loci. In there, we have for every manifold $X$

  • the singular simplicial complex $X^{\Delta_{Diff}^\bullet}$ of smooth singular simplices on $X$,

  • the infinitesimal singular simplicial complex $X^{(\Delta^\bullet_{inf})}$ of infinitesimal singular simplices.

There is a canonical injection $X^{(\Delta^\bullet_{inf})} \to X^{\Delta^\bullet_{Diff}}$. We may take degreewise (internally, i.e. smoothly) functions on these, to get the cosimplicial algebras $[X^{\Delta^\bullet_{inf}},R]$ and $[X^{\Delta^\bullet_{Diff}},R]$.

The normalized cochain complex of chains on $[X^{\Delta^\bullet_{Diff}},R]$ is the complex of smooth singular cochains.

The normalized cochain complex of chains on $[X^{\Delta^\bullet_{inf}},R]$ turns out to be, by some propositions by Anders Kock, to be the deRham algebra, as discussed a bit at differential forms in synthetic differential geometry.

Therefore under the ordinary Dold-Kan correspondence we have a canonical morphism

$$ N^\bullet([X^{\Delta^\bullet_{Diff}},R] \to [X^{\Delta^\bullet_{inf}},R]) = C^\bullet_{smooth}(X) \to \Omega_{dR}^\bullet(X) $$

which is an equivalence of cochain complexes. But there is a refinement of the Dold-Kan correspondence the monoidal Dold-Kan correspondence. And this says that this functor is also a weak equivalence of oo-monoid objects.

share|cite|improve this answer
Thanks, Urs. Do I understand correctly that your answer concerns the $A_\infty$ case? If so, what about the $E_\infty$ case? I think this would translate as whether or not the Moore complex functor is symmetric lax monoidal. Is this translation correct? – algori Jan 6 '10 at 0:23

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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