MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

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

I am having some trouble finding and/or understanding a general definition of subvarieties intersecting transversally. Assume that $Z_1,\ldots,Z_k$ are closed, irreducible subvarieties of a nonsingular algebraic variety $Y$.

Intuitively, I would say that these subvarieties intersect transversally if all varieties have pairwise intersection multiplicity one, i.e. $i(W,Z_i\cdot Z_j; Y)=1$ along any component $W$ of $Z_i\cap Z_j$ for $i\ne j$.

In our scenario, we should have $i(W,Z_i\cdot Z_j;Y)=\mathop{\mathrm{length}}_{\mathscr{O}_{Y,W}}\left(\mathscr{O}_{Z_i\cap Z_j, W}\right)$, if I am not mistaken.

Another intuitive definition would be that the tangent sheaves $\mathscr{T}_{Z_i}$ form a direct sum inside $\mathscr{T}_{Y}$. This seems to agree with the definition in this paper (in 5.1.2), but the author gives an equivalent definition which looks interesting:

For any $y\in Y$, there exists

  1. a system of parameters $x_1,\ldots,x_n$ on $Y$ at $y$ that are regular on an affine neighborhood $U$ of $y$ such that $y$ is defined by the maximal ideal $(x_1,\ldots,x_n)$ as well as
  2. integers $0=r_0 \le r_1 \le \cdots \le r_k \le n$ such that the subvariety $Z_i$ is defined by the ideal $I_i=\left(x_{r_{i-1}+1},\ldots,x_{r_i}\right)$ for all $1\le i\le k$.

Just for the record, what precisely does "defined by" mean here? I assume it means that $U\cap Z_i = Z(I_i)$ ... or do we actually get that $I_i=I(Z_i\cap U)$?

I would like to know if and how these three definitions are equivalent - there is no (general) treatment of this in Hartshorne or even in Fulton's book on Intersection Theory, which befuddled me greatly.

share|cite|improve this question
"defined by" must be scheme-theoretic, ie, $I_i=I(Z_i\cap U)$ – quim May 20 '11 at 9:03
Well, that is at least something. Maybe I will just work with that, it's so comfortably local. – Jesko Hüttenhain May 23 '11 at 15:18
up vote 4 down vote accepted

This naive answer is strictly intended to provoke a knowledgeable answer from someone. Transversal intersection at p should mean that all subvarieties are smooth at p, all contain p, and the codimension of the intersection of the tangent spaces equals the sum of the codimensions of the individual tangent spaces. ????

share|cite|improve this answer
The only correction I'd make is: the codimension of the intersection of the Zariski tangent spaces equals the sum of the codimensions of the subvarieties at the point (so, intersections at the singular points won't be allowed). – quim May 20 '11 at 9:01
I thought my requirement that the points be smooth includes this? – roy smith May 22 '11 at 3:38
Well. Yea, this is yet another intuitive thing you would want a transversal intersection to satisfy, but as you already pointed out, it's not the answer to my question. I am somewhat startled that noone posted a follow-up. Thanks anyway! – Jesko Hüttenhain May 23 '11 at 15:18

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.