MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $X$ be a smooth projective variety over $\mathbb C$ with $d = \dim X \geq 1$. Let $CH(X)$ denotes the total Chow group of (cycles modulo rational equivalences of) $X$ and $CH(X)_{\mathbb Q} = CH(X)\otimes_{\mathbb Z} \mathbb Q$. My question is

Suppose $CH(X)_{\mathbb Q}$ is finite-dimensional as a $\mathbb Q$-vector space. When can we conclude that $H^d(X, \mathcal O_X) =0$?

I suspect that the answer might be always yes: in dimension one, such varieties have to be rational. In dimension two, I think it follows from a paper of Mumford. Also, it is true for flag manifolds.

More vaguely, I am also quite interested in:

Suppose $CH(X)_{\mathbb Q}$ is finite-dimensional as a $\mathbb Q$-vector space. What can we say about the geometry of $X$?

Any partial answers would be appreciated. Thanks!

share|cite|improve this question
I believe that Srinivas has extended Roitman's result to $X$ normal and not necessarily smooth – aginensky Sep 30 '11 at 18:20
up vote 9 down vote accepted

As you suspect, the answer to your first question is yes.

A. Roĭtman in "Rational equivalence of zero-dimensional cycles". Math. USSR-Sb. 18 (1974), 571--588, generalised Mumford's theorem to show that if $h^q(X, \mathcal{O}_X) > 0$ for any $q>0$ then $CH_0(X)$ is infinite dimensional. If $q > 1$ then there is no surjection from the points of a variety over $\mathbb{C}$ onto $CH_0(X)$ so it is infinite dimensional in a very strong sense.

In general, it is conjectured -- generalizing Bloch's conjecture for surfaces -- that $CH(X)$ is finite dimensional iff $h^p(X, \Omega^q) = 0$ for $p \neq q$. In this case one expects that the cycle class map $CH(X) \to H(X)$ is an isomorphism, where $H(X)$ denotes the total singular cohomology with $\mathbb{Q}$-coefficients.

The difficult direction is proving finite dimensionality -- this is not even known for surfaces -- but I think the other direction is also not known in full generality. However, if the cycle class map is injective then Jannsen has proved the above vanishing of the Hodge numbers (in Motivic sheaves and filtrations on Chow groups. Motives (Seattle, WA, 1991), 245–302, Proc. Sympos. Pure Math., 55, Part 1, Amer. Math. Soc., Providence, RI, 1994.)

share|cite|improve this answer
Dear ulrich, thank you very much, this sounds fascinating. I do not have access to Roitman's paper, so please allow me two extra questions: Do we know something about $H^d(X, \mathcal O_X(i))$ for $i>0$? And what is a reference for your third and fourth paragraphs? Thanks a ton. – Hailong Dao Sep 25 '11 at 14:26
Dear Hailong, the $H^q(X, \mathcal{O}_X)$ are summands of the Hodge structure on $H^q(X)$ and control its "level" which is why they influence the Chow groups. As far as I know, $H^d(X,\mathcal{O}_X(i))$ for $i>0$ have no motivic meaning so I don't think one can say much about them from knowledge of the Chow groups. – ulrich Sep 25 '11 at 15:39
Dear Hailong, I added a reference for Jannsen's result. I do not know a precise reference for the conjecture in the third paragraph; it is a consequence of the generalised Hodge conjecture, Roitman's theorem and the (conjectured) faithfulness of the map from the category of mixed motives to the category of mixed Hodge structures. I will try to find a reference and add it later. – ulrich Sep 25 '11 at 15:47
Roitman, and also Bloch, only treated the torsion which is prime to the characteristic. The $p$-torsion in caracteristic $p$ has been treated in a paper of J. S. Milne (Compositio Math., 1982). Numdam link: – ACL Sep 25 '11 at 22:54
Dear ulrich and ACL, thank you very much. – Hailong Dao Sep 26 '11 at 2:14

For a smooth proper complex variety $X$, the condition that $CH^*(X)_{\mathbb Q}$ is a finite dimensional vector space is equivalent to the Chow motive of $X$ being a direct sum of Tate motives. In particular, the singular cohomology of $X$ (with $\mathbb Q$ coefficients) coincides with $CH^*(X)_{\mathbb Q}$ (and thus the Hodge decomposition is rather trivial). This can be proved using the technique of decomposition of the diagonal.

A version of this argument (for an arbitrary Chow motive) appears in an article of Kimura:

MR2562457 (2010j:14014), Surjectivity of the cycle map for Chow motives. in ``Motives and algebraic cycles'', 157–165, Fields Inst. Commun., 56, Amer. Math. Soc., Providence, RI, 2009.

Actually, one can even make do without the ``supported correspondences'', by reasoning as in Kapil Paranjape's article

MR1283872 (95g:14008), Cohomological and cycle-theoretic connectivity. Ann. of Math. (2) 139 (1994), no. 3, 641–660.

But they do give a clean way to write the proof.

There is a follow-up by Charles Vial (again, there is no real need to use ``birational motives'' in the proofs, for this result):

MR2738925 (2012c:14010), Pure motives with representable Chow groups, C. R. Math. Acad. Sci. Paris 348 (2010), no. 21-22, 1191–1195.

share|cite|improve this answer
Thank you very much, and welcome to Mathoverflow! – Hailong Dao Mar 27 '13 at 19:28

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.