MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

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

For a Kahler manifold $M$ we have two well-known Laplacians: the de Rham Laplacian $\Delta_{\text{d}} = ($d$ + $d$^\ast)^2$, and the Dolbeault Laplacian $\Delta_{\overline{\partial}} = (\overline{\partial} + \overline{\partial}^\ast)^2$. Now on smooth functions, these two operators are related by the well-known formula $$ \Delta_{\text{d}} = 2\Delta_{\overline{\partial}} $$ Now both these operators act on the exterior algebra. Does there exist a similar formula in this more general setting?

share|cite|improve this question
Yes, the identity holds for differential forms. This is what's behind the Hodge decomposition. See your favourite book on Kahler manifolds (Griffiths-Harris, Wells...) for an explanation. – Donu Arapura Mar 24 '11 at 18:20
(I don't think you want the superscript $2$ in your displayed formula.) On a Kähler manifold, the three laplacians associated with $d$, $\partial$ and $\bar\partial$ satisfy $$\Delta_d = 2 \Delta_{\partial} = 2 \Delta_{\bar\partial}$$ not just on functions, but also on forms. This is why on a compact Kähler manifold, one has the decomposition of the de Rham cohomology in terms of the Dolbeault cohomology groups. This can be found, e.g., in Well's Differential analysis on complex manifolds. – José Figueroa-O'Farrill Mar 24 '11 at 18:24
He does want superscripts, since what he calls Laplacian is not... He is writing the Dirac operator instead... – diverietti Mar 24 '11 at 18:28
Sorry, about that, superscripts in the wrong place. Yes, it still works out, but best to stick to convention. – Jean Delinez Mar 24 '11 at 18:31
up vote 9 down vote accepted

If $(X,\omega)$ is Kähler, then it is always true that $$ \Delta'=\Delta''=\frac 12\Delta, $$ where these three Laplacians are with respect, in order, to $\partial$, $\bar\partial$ and $d$. This is valid when they act on any space of complex-valued differential forms.

More generally, you can look to differential forms with values in a hermitian vector bundle $E\to X$. In this case, take $D_E$ to be the (unique) Chern connection of $E$ and let $D_E=D'_E+D''_E$ its decomposition in the $(1,0)$ and $(0,1)$ part (then, by definition $D''_E=\bar\partial$).

In this case, you can again compare $\Delta'_E$ and $\Delta''_E$. They no longer coincide, but differ by a order zero operator which is expressed in terms of the curvature $\Theta(E)=D^2_E$ and the (formal) adjoint $\Lambda_\omega$ of the operator $L_\omega=\omega\wedge\bullet$ of wedge product with $\omega$. The relation is $$ \Delta''_E=\Delta'_E+[i\Theta(E),\Lambda_\omega], $$ where $[\bullet,\bullet]$ is the (graded) commutator.

On the other hand, coming back to complex-valued differential forms, if you merely suppose your manifold to be hermitian than the relation between the three Laplacians is a little bit more complicated. You have to introduce the torsion operator $$ \tau=[\Lambda_\omega,\partial\omega] $$ which is of type $(1,0)$ and order zero (observe that if $\omega$ is Kähler then $\partial\omega=0$). With these notations, you have $$ \Delta''=\Delta'+[\partial,\tau^*]-[\bar\partial,\bar\tau^*] $$ and $$ \Delta=\Delta'+\Delta''-[\partial,\bar\tau^*]-[\bar\partial,\tau^*], $$ so that $\Delta'$, $\Delta''$ and $\frac 12\Delta$ no longer coincide but they differ by linear differential operators of order $1$ only.

share|cite|improve this answer

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.