The Hodge Index Theorem for compact Kaehler manifolds seems to be a big deal in complex geometry. See here for the surface version of the result.


I personally can't see why this result is important, and I would like to know what are some important consequences of this result.

  • 1
    $\begingroup$ The negativity lemma (lemma 3.39 in Koll\'ar-Mori) is very useful in the minimal model program $\endgroup$
    – Hacon
    Oct 19 '17 at 19:31
  • 2
    $\begingroup$ Another consequence: If $A,B$ are nef divisors on a surface and $A\cdot B=0$, then $A$ and $B$ are numerically proportional. This trick is often used in the study of linear systems on surfaces. $\endgroup$
    – byu
    Oct 19 '17 at 22:30
  • $\begingroup$ if you believe the Riemann hypothesis for curves is important, that seems to follow from Hodge (and Riemann -Roch), via Mattuck-Tate, Castelnuovo-Severi, Grothendieck. arxiv.org/pdf/1509.00797.pdf $\endgroup$
    – roy smith
    Oct 20 '17 at 3:50
  • $\begingroup$ @byu Using Hodge index theorem in the situation you described, I can see that $A^2>0$ implies $B$ numerically trivial and, by symmetry, $B^2>0$ implies $A$ numerically trivial. How do you conclude $A,B$ are numerically proportional? $\endgroup$
    – rmdmc89
    May 1 at 21:14

Let $S$ be a smooth projective surface and $H$ a $\mathbb{Q}$-divisor with $H^2>0$ (for instance, an ample divisor).

If $D$ is another $\mathbb{Q}$-divisor such that $HD=0$, then $D^2 \leq 0$ and equality holds if and only if the class of $D$ is zero in $H^2(S, \, \mathbb{Q})$.

  • 3
    $\begingroup$ Isn't this exactly the statement? You mean that the statement itself is the reason why it is important? :p $\endgroup$
    – diverietti
    Oct 19 '17 at 20:08
  • 3
    $\begingroup$ The original statement is about the signature of the intersection form in the Neron-Severi group. The statement I gave is the form in which the result is generally used in the context of surfaces. In fact, in Barth-Peters-Van de Ven book it is explicitly stated as an important corollary. $\endgroup$ Oct 19 '17 at 20:14

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.