A classical result in algebraic geometry states that every irreducible component of a variety defined by $r$ polynomials in affine $n$-space has dimension not less than $n-r$. This is a special case of Krull's height theorem which states that in a noetherian ring every prime ideal minimal above an ideal generated by $r$ elements has height at most $r$. I would like to know if this bound holds for all associated primes, i.e., also for the embedded primes.

To be precise, the question is the following: Let $k$ be a field and $f_1,\ldots,f_r\in k[x_1,\ldots,x_n]$ polynomials in $n$-variables ($r< n$). Assume that the ideal $I=(f_1,\ldots,f_r)$ generated by the $f_i$'s does not contain $1$. Is it true that $\dim(k[x_1,\ldots,x_n]/\mathfrak{p})\geq n-r$ for every associated prime ideal $\mathfrak{p}$ of $k[x_1,\ldots,x_n]/I$?

It seems a rather natural question to me and so I find it hard to believe that I am the first one to think about it. However, so far, I could not find anything in the literature. This seems to indicate that it might not be true. At the same time it appears not too easy to find a counterexample. Note that in a Cohen-Macaulay ring every ideal of height $r$ generated by $r$ elements is unmixed, i.e., has no embedded primes. In particular the answer is "yes" if $\dim(k[x_1,\ldots,x_n]/I)=n-r$.


No, this is not true. A counter example is the ideal $\langle xz, yw, xw+yz \rangle$, which has $3$ generators, but has an associated prime of codimension 4 at the origin. I'm not sure how much it explains, but I remember this example as being the ideal of pairs of linear forms whose product is zero, i.e. the coefficients of $(xt + y)(zt + w)$, considered as a polynomial in $t$. This at least lets you see right away that the minimal primes are $\langle x, y\rangle$ and $\langle z, w \rangle$.

  • 1
    $\begingroup$ Thanks for the example. Indeed a primary decomposition is given by the ideals $\langle x,y\rangle$, $\langle z,w\rangle$, $\langle xz, yw, xw+yz, x^2, y^2, z^2, w^2)$. $\endgroup$ – anonymous Mar 4 '12 at 15:19

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.