1
$\begingroup$

Let $ f: \mathbb{P}^n \dashrightarrow (\mathbb{P}^1)^n , [x_0:\dots :x_n] \mapsto ([x_0,x_1],[x_0,x_2], \dots ,[x_0,x_n])$ a birational map. Is $f$ a Crepant birational map?

$\endgroup$
2
$\begingroup$

Kontsevich proved in '95 that the hodge polynomial is independent of crepant resolution over a field of characteristic zero. He created the field of motivic integration to prove this fact. Here are the respective hodge polynomials of the two spaces in question where I repress the fact that we are working over the field $k$ just for notational simplicity:

$$H([\mathbb{P}^n], u,v) = \sum_{i=0}^{n} H([\mathbb{A}^i], u,v) =\sum_{i=0}^{n} (uv)^i $$

and

$$H([\prod_{i=1}^{n}\mathbb{P}^{1}], u,v) = \prod_{i=1}^{n} (H([\mathbb{A}^1], u,v)+1) = \prod_{i=1}^{n} (uv+1) = (uv+1)^n = \sum_{i=0}^{n} {n \choose i} (uv)^i $$

where we denote by $[X]$ the equivalence class of the variety $X$ in the grothendieck ring of varieties. Therefore, there are no crepant birational maps between $\mathbb{P}^n$ and $\prod_{i=1}^{n} \mathbb{P}^1$.

Note, actually, here $[X]$ lives in the image of the Grothendieck ring of varieties localized at $[\mathbb{A}^1]$ under the completion homomorphism given by the dimension filtration. But, the hodge polynomial ring homomorphism `'$$H(X,u,v) = \sum_{i,j} (-1)^{i+j} dim(H^i(X, \Omega_{X}^{j}))u^iv^j $$' factors through the ring homomorphism $X \mapsto [X]$.

I am only sure that this argument works over a field of characteristic zero for two reasons: 1) there is not currently a resolution of singularities over fields of positive characteristic, and 2) I don't know how to define the hodge polynomial in positive characteristic.

Note that this result is sometimes known as Batyrev's Theorem because he originally proved a similar result in '95 involving the beta numbers via p-adic integration and the weil conjectures.

$\endgroup$
7
  • $\begingroup$ Sorry, i make a mistake in the statement. I will edit my post. Obviously there are no crepant birational maps between $\mathbb{P}^n$ and $(\mathbb{P}^1)^n$ because they have a distinct number of generators of their Picard Groups. $\endgroup$ Mar 6 '13 at 6:30
  • 1
    $\begingroup$ That is annoying that you completely changed the question. $\endgroup$ Mar 6 '13 at 6:39
  • 1
    $\begingroup$ anyway, this argument I posted still works. $f$ cannot be a birational map or else the hodge polynomials would be the same (by the same work of Kontsevich mentioned above). next time you should either add to your question or start a new question instead of asking a completely different question after someone has answered you. $\endgroup$ Mar 6 '13 at 6:48
  • $\begingroup$ Really sorry, i get confused and wrote the wrong question cause both had the birational morphism f. $\endgroup$ Mar 6 '13 at 6:49
  • $\begingroup$ I will edit it or will get very confusing. $\endgroup$ Mar 6 '13 at 6:50

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.