Let $G=\mathbb R^*$ act on $X=\mathbb R^3\setminus\{0\}$ by pointwise multiplication. That is for any $t\in\mathbb G$ and $(x_1,x_2,x_3)\in X$ we have $$t\cdot(x_1,x_2,x_3)=(tx_1,tx_2,tx_3)$$

Is the resulting quotient a smooth manifold?

I know from this answer that I only need to check that the action is smooth, free and proper. Smoothness and freeness are easy.

How do I show that the action is proper?

I have to show that the map $\theta:G\times X\to X\times X$ given by $\theta(g,x)=(gx,x)$ is proper (inverse image of compact set is compact) or equivalently - for each pair of pints $x,y\in X$ there are neighbourhoods $V_x$ and $V_y$ of $x$ and $y$ respectively such that $H=\{\ g\in G\ |\ gV_x\cap V_y\neq\emptyset\ \}$ is relatively compact in $G$. But I am unable to do this. Any help would be appreciate.

Thank you.

  • 2
    $\begingroup$ Do you understand that the action is a scaling? So it $x, y$ does not lie in the same line, $gx$ will never be $y$, So in this case you can choose small neigborhood $V_x, V_y$ so that $\{ g\in G : gV_x \cap V_y \neq \emptyset\} = \emptyset$. $\endgroup$ – user99914 Jun 2 '16 at 7:08
  • $\begingroup$ @John Ma, Yes I see! But what about when $y=gx$ for some $g$? Can I choose neighbourhoods so that $H=\{g\}$? $\endgroup$ – R_D Jun 2 '16 at 8:04
  • 2
    $\begingroup$ No you cannot, but you need only that $H$ is bounded and bounded away from $0$. Just pick any small $V_x$, and $V_y$. $\endgroup$ – user99914 Jun 2 '16 at 9:57
  • 4
    $\begingroup$ I think you can directly use that in a compact subset of $X\times X$ both components are bounded in norm above and below (i.e. bounded away from zero), This readily implies that the pre-image under $\theta$ is bounded in norm in both directions, which implies that $\theta$ is proper. Just as a sideremark: For the example in question, it is probably easier to directly construct local charts via inhomogeneous coordinates on projective space than the verify the properties of the action. $\endgroup$ – Andreas Cap Jun 2 '16 at 10:44

The following reformulation of properness is often very useful in practice.

Proposition. (Lee, Introduction to smooth manifolds, Proposition 21.5) Let $M$ be a manifold and $G$ a Lie group acting continuously on $M$. The following are equivalent.

  1. The action is proper.
  2. If $(p_i)$ is a sequence in $M$ and $(g_i)$ is a sequence in $G$ such that both $(p_i)$ and $(g_i\cdot p_i)$ converge, then a subsequence of $(g_i)$ converges.

Now, we can easily prove that your action is proper by proving that the second item in this proposition holds.

Let $(x_n,y_n,z_n)\in\Bbb R^3-\{0\}$ and $t_n\in\Bbb R^*$ be such that $(x_n,y_n,z_n)$ and $(t_nx_n,t_ny_n,t_nz_n)$ converge. Then, $$\lim_{n\to\infty}(x_n,y_n,z_n)=(x,y,z)\in\Bbb R^3-\{0\}.$$ Since $(x,y,z)\in\Bbb R^3-\{0\}$, one of the coordinates is non-zero, and hence we may assume without loss of generality that $x\neq 0$. Hence $x_n\to x\neq 0$ as $n\to\infty$. In particular, $x_n\neq 0$ for all $n$ large enough, so $$\lim_{n\to\infty}\frac{1}{x_n}=\frac{1}{x}.\tag{1}$$ (By removing enough of the first elements of the sequence.) Now, we are given that $(t_nx_n,t_ny_n,t_nz_n)$ also converges, so $$\lim_{n\to\infty}t_nx_n=a,\tag{2}$$ for some $a\in\Bbb R$. Putting $(1)$ and $(2)$ together, we get $$\lim_{n\to\infty}t_n=\lim_{n\to\infty}t_nx_n\frac{1}{x_n}=\left(\lim_{n\to\infty}t_nx_n\right)\left(\lim_{n\to\infty}\frac{1}{x_n}\right)=\frac{a}{x}\in\Bbb R.$$ The second equality is justified since both limit exist. So $t_n$ converges in $\Bbb R$. The last thing we need to check is that the limit is actually in $\Bbb R^*$. But this must be the case as otherwise $$\lim_{n\to\infty}(t_nx_n,t_ny_n,t_nz_n)=(0\cdot x,0\cdot y,0\cdot z)=(0,0,0)\notin\Bbb R^3-\{0\}$$ (because $(x_n,y_n,z_n)$ converges to $(x,y,z)$). This shows that the action is proper and hence the quotient $(\Bbb R^3-\{0\})/\Bbb R^*$ is a manifold.

(The manifold that you constructed with this quotient is called a real projective space, and is usually denoted $\Bbb{RP}^3$.)


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.