3
$\begingroup$

I want to prove the following statement. Please help!

Given any semialgebraic set $A$, consider its real Zariski closure $V_{\mathbb{R}}$ (which always has the same real dimension of $A$). Now consider the complex ideal $I_{\mathbb{C}}(V_{\mathbb{R}})$, and its complex solution set $V_{\mathbb{C}}$. Note that $V_{\mathbb{R}}$ is the restriction to $\mathbb{R}^{n}$ of $V_{\mathbb{C}}$ (which we can assume irreducible). Prove that the complex dimension of $V_{\mathbb{C}}$ equals the real dimension of $V_{\mathbb{R}}$.

One avenue is to emulate Tao's solution to Every real variety contains non-singular points and prove that the complex dimension of $V_{\mathbb{C}}$ equals the real dimension of $V_{\mathbb{R}}$ by working with smooth points, as follows.

If $V_{\mathbb{C}}$ is not invariant by complex conjugation, then one may intersect this variety with its complex conjugate, dropping the dimension of this variety without affecting $V_{\mathbb{R}}$ as a set. Thus, by induction on dimension, one may assume that $V_{\mathbb{C}}$ is invariant by complex conjugation, that is is definable over the reals.

Now, we're left by induction with some smooth real point of $V_{\mathbb{C}}$. At such point of $V_{\mathbb{C}}$, the complex tangent space is invariant by complex conjugation, and is thus the complexification of a real space of the same dimension.

However, from here I can't conclude that $V_{\mathbb{R}}$ is smooth at the point, unless by using the equality between dimension which I want to prove in the first place!

Also check out the comments under https://mathoverflow.net/a/262737/501777

$\endgroup$

2 Answers 2

4
$\begingroup$

Let $x$ be a real point of a complex variety $V_\mathbb C \subseteq \mathbb C^n$ that is smooth of dimension $\dim(V_\mathbb C)=m$ and defined by polynomials over the real numbers. Then we can always choose exactly $n-m$ elements $f_1,\dotsc, f_{n-m}$ of the ideal of $V_\mathbb C$ whose derivatives together generate the conormal space to $V_{\mathbb C}$ at $x$. The complex vanishing set of $f_1,\dotsc, f_{n-m}$ is smooth of dimension $m$ at $x$ by the complex intermediate value theorem, and the real vanishing set of $f_1,\dotsc, f_{n-m}$ is smooth of dimension $m$ in a neighborhood of $x$ by the real intermediate value theorem.

If $V_{\mathbb R}$ is not smooth of dimension $m$ at $x$, then it is a proper subset of the real vanishing set of $f_1,\dotsc, f_{n-m}$ in every neighborhood of $x$, so $V_\mathbb C$ is a proper subset of the complex vanishing set of $f_1,\dotsc, f_{n-m}$ in every neighborhood of $x$, contradicting that $V_\mathbb C$ is smooth of dimension $m$ at $x$.

$\endgroup$
2
  • $\begingroup$ Thanks for your answer! Do you think the statement (and some adaptation of your proof) holds where instead of real algebraic varieties and their complex Zariski closure, we consider real analytic varieties and their complex analytic closure (real and complex algebraic dimensions now become real and complex analytic dimensions)? $\endgroup$
    – user86954
    Commented Apr 3, 2023 at 17:18
  • $\begingroup$ @user86954 Yes, I think by almost literally the same proof. $\endgroup$
    – Will Sawin
    Commented Apr 3, 2023 at 17:36
1
$\begingroup$

I think this topic is covered in Whitney, Hassler. "Elementary structure of real algebraic varieties." Annals of Mathematics Second Series, Vol. 66, No. 3 (Nov., 1957), pp. 545-556 (link). See esp. Lemmas 6, 8, and 9.

$\endgroup$
0

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .