I'm looking to find a way to derive the Fubini-Study metric on $\Bbb CP^n$ and the corresponding Kähler form, but I cannot find a proper derivation for this. In most references I found they just state that the Kähler form is defined either as $$\omega_{FS} = i\partial\bar{\partial}\log \|s\|^2$$ for a section $s$ or something like

$$ \omega_{FS} ={\frac {i}{2}}\partial {\bar {\partial }}\log |\mathbf {Z} |^{2} $$

as wikipedia does. How are these things derived they seem very mysterious?

  • 1
    $\begingroup$ Here are a few bits of intuition: You want a $U(n+1)$-invariant metric on $\Bbb CP^n = U(n+1)/(U(n)\times U(1))$. Secondly, in your formula, $s$ is a nowhere-zero (local) holomorphic section of $\mathscr O(-1)$, the tautological line bundle; any two sections differ by multiplication by a nonzero holomorphic function $f$. Note that $\log\|fs\|^2 = \log |f|^2 + \log \|s\|^2$, and $\partial\bar\partial\log|f|^2 = 0$. If you're comfortable with Maurer-Cartan forms on Lie groups, I can give you a different derivation. $\endgroup$ Commented May 18 at 17:30

2 Answers 2


$\newcommand\C{\mathbb{C}}$Here is a brief explanation but you'll need to fill in all of the details: First, denote \begin{align*} \pi: \C^{n+1}\backslash\{0\} &\rightarrow \C P^n\\ Z &\mapsto [Z] \end{align*} and for each $Z \in \C^{n+1}\backslash\{0\}$, let $$ \pi_Z: T_z\C^{n+1} \rightarrow T_{[Z]}\C P^n $$ be the pushforward map of $\pi$ at $Z$.

First, observe that given any metric $g$ on $\C^{n+1}\backslash\{0\}$, the restriction of $\pi_Z$, $$ \pi_Z: T^\perp_Z \rightarrow T_{[Z]}\C P^n, $$ where $T^\perp_Z \subset T_Z\C^{n+1}$ is the subspace orthogonal to $\pi^{-1}([Z])$, is an isomorphism. So the idea is to define a metric $\hat{g}$ on $\C^{n+1}\backslash\{0\}$ that is invariant under rescaling by a complex scalar and define the metric on $T_{[Z]}\C P^n$ to be $\hat{g}$ restricted to $T^\perp_Z$.

The Euclidean metric $dZ|^2$ on $\C^{n+1}\backslash\{0\}$ is not invariant under scaling. However, the metric $$ \frac{|dZ|^2}{|Z|^2} $$ is. This metric restricted to $T^\perp_Z$ and pushed forward to $T_{[z]}\C P^n$ is the Fubini-Study metric.


I find it helps to zoom out a bit. Let $V$ be a complex vector space of dimension at least one and let $P(V)$ be the projective space of lines in it. We have the trivial vector space $V \to P(V)$, and the tautological line bundle $\mathcal O(-1) \subset V$. These fit into a short exact sequence $$ 0 \to \mathcal O(-1) \to V \to V / \mathcal O(-1) \to 0. $$ Now pick a Hermitian inner product $h$ on $V$. This defines a flat Hermitian metric on the trivial bundle $V$, and thus a Hermitian metric on the line bundle by restriction. That metric will have negative curvature by the Codazzi—Griffiths equations, and its negative is the Fubini—Study metric.

If you pick a basis of $V$ and the standard inner product this unpacks to the unmotivated definition in Wikipedia and other places.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .