How is the Kähler form decomposed in terms of the metric?

153 Views Asked by At

In countless textbooks and lecture notes (e.g. eqn 4.9 of Lectures on Riemannian Geometry, Part II: Complex Manifolds by Stefan Vandoren), the Kähler (1,1)-form, $\omega$, is written in terms of the Riemannian metric: \begin{equation} \omega = 2 i g_{a\bar{b}}dz^a \wedge d\bar{z}^{\bar{b}}\,. \end{equation} where $dz^a$ and $d\bar{z}^a$ are holomorphic and antiholomorphic coordinates, respectively. My question is simple - how does this make sense? The metric is symmetric $g_{a\bar{b}}=g_{\bar{b}a}$ and the wedge product is of course antisymmetric suggesting to me that $\omega$ should be zero. Also my understanding was that a differential p-form, $C_p$, was defined: \begin{equation} C_p=\frac{1}{p!}C_{m_1\dots m_p}dx^{m_1}\wedge \dots \wedge dx^{m_p} \end{equation} where the components, $C_{m_1\dots m_p}$, are completely antisymmetric in the indices. Since $i g_{a\bar{b}}$ is clearly not antisymmetric, how is this possible? Do the components of a (1,1)-form not have to be totally antisymmetric?

I realise there must just be a simple misunderstanding here on my part so I apologise for the simplicity of the question but I would very much appreciate any help. Thank you.

1

There are 1 best solutions below

0
On

The summation is only through holomorphic indices. So $g_{\bar b a}$ term does not appear in the sum and hence does not cancel out. Its best to write out the definition $w(.,.)=g(J.,.)$ and figure out the coefficients by plugging in co-ordinate vector fields.

If $w=w_{A\bar{B}}dz^{A}\wedge d\bar z^{\bar B}$ then by plugging in the vectors $\dfrac{\partial}{\partial z^{a}},\;\dfrac{\partial}{\partial \bar z^{\bar b}}$ in the above definition, we get that

$$w_{a\bar b}-w_{\bar b {a}}=ig_{a\bar b}$$

where $A,B$ goes through all indices, but $a,b$ takes only holomorphic indices (without bar). In the author's notation, they used antisymmetry of wedge product to combine $w_{a\bar b}dz^a\wedge d\bar z^{\bar b}+w_{\bar b {a}} d\bar z^{\bar b}\wedge dz^a$ into $(w_{a\bar b}-w_{\bar b {a}})dz^a\wedge d\bar z^{\bar b}$ and then used the formula above.

Also, what you have written about anti-symmetry of coefficients of forms is not true. It's true only if you have tensor products instead of wedges in that equation.