Recall that a differential $k$-form $\alpha$ on a smooth manifold $M$ is called weakly closed if

$$\int_M \alpha \wedge d\beta = 0,$$

for all smooth forms $\beta$ of degree $n-k-1$, where $n = \dim M$. My question is:

If $\alpha$ is a weakly closed continuous 1-form on a closed manifold $M$, can we conclude that $\alpha$ is the sum of a smooth closed 1-form and $d\phi$, where $\phi$ is a $C^1$ function on $M$?

This appears to follow from the properties of the de Rham regularization operator(s) and associated homotopy operator(s) on forms; see, for instance, Theorem 12.5 in V. Gol'dshtein and M. Troyanov, Sobolev inequalities for differential forms and $L_{q,p}$-cohomology, Journal of Geometric Analysis, vol. 6, no. 4, 2006.

Any thoughts on this would be appreciated.

  • $\begingroup$ This is true for $\phi \in L^1_{\rm loc}$ though. $\endgroup$
    – Henri
    Jun 20, 2019 at 7:25
  • $\begingroup$ Thanks! But wouldn't $\phi \in L^1_{loc}$ and $d\phi$ continuous (being equal to a continuous form minus a smooth form) imply that on a compact manifold, $\phi \in W^{1,\infty}$? If so, then $\phi$ is Lipschitz, hence differentiable a.e. in the ordinary sense and therefore (since $d\phi$ is $C^0$) $\phi$ is a.e. equal to a $C^1$ function? $\endgroup$ Jun 20, 2019 at 17:28

1 Answer 1


The answer is no. If $\varphi$ is a continuous function on $\mathbb{R}^n$, then the form $\alpha=\varphi\, dx_1\wedge\ldots\wedge dx_n$ is weakly closed since it is an $n$-form. Then using classical notation solving the equation $d\omega=\alpha$ is the same as solving the equation $\operatorname{div}\Phi=\varphi$. However it was proved in [1] that the equation $\operatorname{div}\Phi=\varphi$ may have no $C^1$ solutions.

[1] D. Preiss, Additional regularity for Lipschitz solutions of PDE. J. Reine Angew. Math. 485 (1997), 197-207.

  • $\begingroup$ Thanks! It looks like the "no" answer can be obtained for 1-forms (at least in $\mathbb{R}^n$) as follows. Let $\alpha$ be a continuous weakly closed 1-form. Pick any smooth form $\theta$ of degree $n-2$. Then $\alpha \wedge d\theta$ is trivially weakly closed. If $\alpha$ can be written as $\alpha_\ast + d\phi$, for some smooth closed 1-form $\alpha_\ast$ and a $C^1$ function $\phi$, then $\alpha \wedge d\theta = d(\pm \alpha_\ast \wedge \theta +\phi d\theta)$. Since this works for any continuous weakly closed $\alpha$, we get a result which contradicts [1]. Is this correct? $\endgroup$ Jun 20, 2019 at 17:43

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.