Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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

The introduction to Fulton's Intersection Theory reads in part:

To give an idea of the main thrust of the text, we sketch what we call the basic construction. [...] To a closed regular embedding $i : X \to Y$ of codimension $d$, and a morphism $f : V \to Y$, with $V$ a $k$-dimensional variety. [...] this construction produces a rational equivalence class of $(k-d)$-cycles on $W = f^{-1}(X)$. This intersection class $X \cdot_Y V$ can be defined as follows: Since $i$ is a regular embedding, the normal cone to $X$ in $Y$ is a vector bundle; let $N$ denote the pull-back of this bundle to $W$. The normal cone $C$ to $W$ in $V$ is a $k$-dimensional closed subscheme of $N$. [...]

I'm confused as to why we can regard $C$ as being contained in $N$. In particular, I can see why we should expect there to be a map from $C$ to $N$ -- heuristically, if we move a little ways in $V$ in a direction normal to $W$ then we can send this over to $Y$ and we shouldn't be moving in a direction parallel to $X$, as $W$ is the preimage of $X$ -- but I can't see why this map ought to be injective.

share|cite|improve this question

First of all, let us fix the notation. Let $I\subset \mathcal O_Y$ be the ideal of $X$ in $Y$ and $J\subset \mathcal O_V$ the ideal of $W$ in $V$. For a regular embedding $X\to Y$, the normal cone coincides with the normal sheaf (which is a bundle over $X$, of rank $d$). In other words, the natural surjection $\textrm{Sym }I/I^2\to \oplus_{n\geq 0}I^n/I^{n+1}$ is an isomorphism. Thus $$C_{X/Y}=\textbf{Spec }\bigl(\oplus_{n\geq 0}I^n/I^{n+1}\bigr)\,\,\textrm{and}\,\,N_{X/Y}=\textbf{Spec }\textrm{Sym }I/I^2\,\,\textrm{coincide.}$$ We have the rank $d$ vector bundle $N$ on $W$: $$N=f^\ast N_{X/Y}\cong \textbf{Spec }f^\ast\bigl(\oplus_{n\geq 0}I^n/I^{n+1}\bigr) \cong \textbf{Spec } \textrm{Sym }f^\ast(I/I^{2}).$$

The claim is then that $$C=C_{W/V}=\textbf{Spec }\bigl(\oplus_{n\geq 0}J^n/J^{n+1}\bigr)$$ has a closed immersion into $N$. This happens if and only if there is a surjection $$\textrm{Sym }f^\ast(I/I^{2})\twoheadrightarrow \oplus_{n\geq 0}J^n/J^{n+1}.$$

I will not do this in detail, but just set up the local calculation: Let us assume that $$X=\textrm{Spec }B\hookrightarrow Y=\textrm{Spec }R\,\,\textrm{and}\,\, V=\textrm{Spec }A.$$ According to our previous notation, we have $$B=R/I \,\,\textrm{and}\,\, W=\textrm{Spec }A/J,\,\,\textrm{with}\,\,A/J=A/IA=A\otimes_RB.$$ Now, $X\to Y$ being a regular embedding of codimension $d$ implies that $I/I^2$ is a free $B$-module of rank $d$ (generated by the images $x_1,\dots,x_d\in I/I^2$ of a regular sequence $\{a_1,\dots,a_d\}\subset I$ of length $d$. Hence, if we identify $$\textrm{Sym }f^\ast(I/I^2)=A/J[f^\ast x_1,\dots,f^\ast x_d],$$ it should be clear that it surjects onto $\oplus_{n\geq 0}J^n/J^{n+1}$. (Maybe it is useful to write $J^n/J^{n+1}$ as $J^n\otimes_A(A/J)$.)

In general. The result you are after is that if $X\to Y$ is a regular embedding and $f:Z\to Y$ is any morphism, then $f^\ast \mathcal C_{X/Y}$ has a natural sheaf surjection onto $\mathcal C_{f^{-1}(X)/Z}$. (Here $\mathcal C$ denotes the conormal sheaf.) Moreover, this is an isomorphism when $f$ is flat.

share|cite|improve this answer

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.