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

Let $F:\mathbb{C}^n\to \mathbb{C}^n$ be a homeomorphism homogeneous of degree 1 (i.e., $F(tx)=tF(x)$, $t>0$) and $g:\mathbb{C}^n\to \mathbb{C}$ a homogeneous polynomial of degree $k$. Let $L$ ($0\in L$) a complex line and $H=F(L)$ such that $g^{-1}(0)\cap H=\{0\}$. It is true that $h=g|_H\circ F|_L$ has topological degree less than or equal to k? This is true if F is linear map!

share|cite|improve this question
What is the topological degree in the noncompact setting? – Alex Degtyarev Apr 3 '14 at 12:18
@AlexDegtyarev it can be defined as the value of the automorphism induced on $2n$-dimensional cohomology with compact support (or homology with infinity support). In that dimension the aforementioned cohomology is $\mathbb Z$. – Fernando Muro Apr 3 '14 at 14:46
@ Fernando: that still requires a properness assumption. Take the constant map from a vector space to itself. It's not proper and it does not induce morphism between cohomologies with compact supports. – Liviu Nicolaescu Apr 3 '14 at 15:01
@FernandoMuro Not any map induces a homomorphism in the homology with compact supports. As stated, the question just doesn't make sense (as the answer brlow points out). – Alex Degtyarev Apr 3 '14 at 15:03
@LiviuNicolaescu and AlexDegtyarev sorry I misread the question and thought he was just talking about homeomorphisms. – Fernando Muro Apr 3 '14 at 15:30

Your question needs to be phrased more carefully.

Take $n=2$, $g(z_1,z_2)=z_1z_2$, $F=$ the identity map $\mathbb{C^2}\to\mathbb{C^2}$ and $L$ is the subspace

$$L=\lbrace(z_1,0),\;\;z_1\in\mathbb{C}\rbrace\subset \mathbb{C}^2. $$

Then $g\circ F|_L=0 $ and $g\circ F|_L$ does not have a degree.

In general for a map $g: \mathbb{C}\to \mathbb{C}$ to have a degree it has to be a proper map. The zero set of a homogeneous polynomial is a conical set in $\mathbb{C}^n$ and so is the image $F(L)$. If $F(L)$ intersects $Z_g:=g^{-1}(0)$ at a point $p_0\neq 0$, then the the intersection $Z_g\cap F(L)$ contains the whole ray $tp_0$, $t\geq 0$ and thus $g\circ F|_L$ cannot be proper and thus it does not even have a well defined degree.

It is not hard to see that $g\circ F|_L$ is proper if and only if $Z_g\cap F(L)=\lbrace 0\rbrace$. Assume that $L$ is the first coordinate line in $\mathbb{C}^n$. The degree of $g\circ F|_L$ is then the winding number of the loop

$$ [0,2\pi]\ni \theta \mapsto g\bigl(\, F(e^{i\theta},0,\dotsc, 0)\bigr)\in\mathbb{C}\setminus 0. $$


Edit. The computation of the above winding number could be tricky. Here is an example that shows that the topological degree need not be equal to the algebraic degree. Assume $n=2$, $g(z_1,z_2)=z_1$, and $L=\{(z_1,0)\}\subset \bC^2$. Set

$$\Sigma=\bigl\{\, (z_1,z_2);\;\;|z_1|^2+|z_2|^2 =1\,\bigr\}\subset \bC^2. $$

We will construct degree $1$ homogeneous homeo $\bC^2\to\bC^2$ by fixing a homeo

$$\phi: \Sigma\to \Sigma. $$

We get a homeo $F_\phi: \bC^2\to\bC^2$ by setting

$$F_\phi(tp)=t\phi(p),\;\;\forall p\in\Sigma,\;\;t>0. $$

Take a tiny unknotted circle $K\subset \Sigma$ contained in a small neighborhood of the point $(1,0)\in \Sigma$. In particular, the projection of $K$ on the first coordinate axis $L$ is a small loop $K'$ with winding number $0$ because $K'$ is contained in a tiny disk of $L$ disjoint from the origin. The circle $\newcommand{\ve}{{\varepsilon}}$

$$ C =L\cap\Sigma=\bigl\{ (z_1,0),\;\;\;|z_1|=1\,\bigr\}\subset \Sigma. $$

describes another unknotted circle on $\Sigma$. In particular, it is isotopic to the circle $K$ and thus there exists a homeomorphism $\phi: \Sigma\to\Sigma$ such that $\phi(C)=K$.

The loop

$$ g\circ \phi: C\to \bC\setminus 0 $$

coincides with the loop $K'$ whose winding number is zero. This shows that

$$ \deg g\circ F_\phi|_L=0\neq \deg g=1. $$

share|cite|improve this answer
Thanks and sorry, suppose $ H \cap \{g = 0 \} = \{0 \} $. Then consider $ L \cong \mathbb {C} \cong \mathbb {R} ^2 $. Hence, define the topological degree of $ g |_H \circ F |_L $ to be the winding number of $ g|_{H \setminus \{0 \}} \circ F |_{L \setminus \{0 \}}: L \setminus \{0 \} \to \mathbb {C} \setminus \{0 \} $. Mr. Liviu Nicolaescu keeps going for my conjecture! Under the above conditions, with $ T $ the topological degree and $ k $ the algebraic degree, then $ | T | \leq k $. The Mr. can give a counter example? – Edson Apr 7 '14 at 14:32
@ Edson. I fixed an omission in my proof. You can get more interesting answers if in my proof you choose $K$ to be a knot disjoint from $C$ and having large linking number with $C$. I believe that $T$ as defined by you is equal to this likinking number. – Liviu Nicolaescu Apr 7 '14 at 18:57
Very hard to be true! For $\mathbb{C}^2$ in the above problem is equivalent to Zakiski's conjecture, that is true in this case! However I think your idea is very useful if I try to adapt for dimension $> 2 $. – Edson Apr 9 '14 at 16:29

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.