4
$\begingroup$

Let $\mathbb{P}^n$ denote the complex projective space of dimension $n$. In many respects, this is the model of (positivity in) complex geometry. There are some well-known characterisations of $\mathbb{P}^n$:

  1. (Most standard characterisation?) $\mathbb{P}^n$ is the parameter space of complex lines through the origin in $\mathbb{C}^{n+1}$.

  2. (Kobayashi-Ochiai). $\mathbb{P}^n$ is the unique compact Kähler manifold $X$ whose first Chern class is $\lambda c_1(\mathscr{F})$, where $\lambda \geq 1+\dim_{\mathbb{C}}(X)$ and $\mathscr{F}$ is a positive line bundle on $X$.

  3. (Mori, Siu-Yau). $\mathbb{P}^n$ is the unique compact Kähler manifold $X$ with a Kähler metric of positive bisectional curvature (this, in fact, characterises $\mathbb{P}^n$ with the Fubini-Study metric, since the biholomorphism here is also an isometry.

In contrast with spheres, $\mathbb{P}^n$ does not appear to be characterised by the sectional curvature of a Riemannian metric on it. Indeed, by the well-known $\frac{1}{4}$-pinching theorem, a compact Riemannian manifold with a Riemannian metric whose sectional curvature $K$ is pinched $\frac{1}{4} < K \leq 1$ is diffeomorphic to a sphere.

Of course, the Fubini-Study metric on $\mathbb{P}^n$ has sectional curvature pinched by $\frac{1}{4} \leq K \leq 1$, but this does characterise $\mathbb{P}^n$ at all: spheres, the Quaternionic projective space, and the Fake Cayley plane also possess metrics with sectional curvature pinched between $\frac{1}{4} \leq K \leq 1$.

I'm very interested in the following question:

What other characterisations of $\mathbb{P}^n$ exist in the literature?

Additionally:

In particular, are there purely Riemannian characterisations of $\mathbb{P}^n$?

$\endgroup$
6
  • 6
    $\begingroup$ Why isn't this a purely Riemannian characterization? Take the canonical circle bundle over the manifold that kills the $2^{nd}$ homotopy group, this is an odd-dimensional sphere and the projective space has the Riemann metric of that quotient space. This is just the higher version of talking about covering spaces of manifolds with $+1$ sectional curvature. You've just replaced covering spaces (i.e. discrete bundles) with circle bundles. $\endgroup$ Commented Dec 30, 2021 at 1:06
  • 2
    $\begingroup$ There is also the characterization by Cho, Miyaoka, and Shepherd-Barron: the unique Fano $n$-fold with Fano pseudoindex equal to $n+1$. $\endgroup$ Commented Dec 30, 2021 at 2:46
  • 4
    $\begingroup$ Though you seem to be more interested by differential-geometric characterizations, let me mention that there are many such characterizations in algebraic geometry: see Cohomological characterizations of projective spaces and hyperquadrics by Araujo-Druel-Kovács (Invent. Math. 174 (2008), no. 2, 233–253), and in particular the introduction. $\endgroup$
    – abx
    Commented Dec 30, 2021 at 7:51
  • 5
    $\begingroup$ It seems like the Seshadri constant of the anticanonical at a smooth point also gives a characterization: link.springer.com/article/10.1007/s00209-017-1941-9 $\endgroup$
    – user347489
    Commented Dec 30, 2021 at 9:52
  • 2
    $\begingroup$ If the Blaschke Conjecture is true then the only Riemannian manifolds satisfying this condition also having an almost complex structure (without any compatibility condition with the metric) are $\mathbb{CP}^n$ (with the Fubini-Study metric) and $S^6$ (with the round metric). If one requires that the metric is Kahler then the exceptional case $S^6$ dissapears. The Blaschke Conjecture is known up to diffeomorphism for complex projective spaces and up to isometry for spheres. Searching "Blaschke Conjecture " should yield the relevant references, $\endgroup$
    – Nick L
    Commented Dec 31, 2021 at 18:15

0

You must log in to answer this question.