1

Calculate $\pi_{1}$ of $X = S^2 / \thicksim $ where a point on the equator is identified with its antipodal point. I know how to calculate $\pi_{1}(S^2),$ but I do not know how to calculate $\pi_{1}$ of this space. could anyone help me in doing so?

I will edit my question with the hints I got within 4 hours.

EDIT: I think at that moment there is no need to write the hints I got because the hints in the comments are much more valuable.

EDIT2 Here is a picture of the original question from the book:

enter image description here

Emptymind
  • 1,901
  • 2
    $X$ is actually the projective plane. You can find an answer here : https://math.stackexchange.com/q/383537/701389 – Arthur Breton Dec 11 '19 at 10:23
  • Remember that the closed upper hemisphere with antipodal identification on the equator is the projective plane. Then, you should be able to conclude with Van Kampen – Popyaitte Dec 11 '19 at 10:25
  • @ArthurBreton could you please explain (maybe by drawing in an answer) why it is the real projective plane? – Emptymind Dec 11 '19 at 13:53
  • could you please explain (maybe by drawing in an answer) why it is the real projective plane?@Popyaitte – Emptymind Dec 11 '19 at 13:53
  • 2
    Do you mean that one point is identified with its antipodal point, or all points are identified with their antipodal point? The hints you've gotten so far assume the latter. In this case, the fact that this is the real projective plane is basically the definition of the real projective plane. – Mees de Vries Dec 11 '19 at 13:55
  • 4
    As I read it, it actually is that we have a fixed equator (a great circle) and each point on it is identified with its antipodal. This is not the real projective plane. Either way some explanation is needed. – freakish Dec 11 '19 at 14:10
  • 1
    @Emptymind To answer your question about how to see the upper closed hemisphere with antipodal identification on the equator is the projective plane. Denote $H$ the open upper half sphere, $\bar{H}$ the closed one, $E$ the equator. $\bar{H}\setminus H = E$. The projection $p: \mathbb{R}^3 \longrightarrow \mathbb{RP}^2$, restricted to $\bar{H}$ is still surjective and injective on its restriction to $H$. By quotienting out by the antipodal relation on $E$ it is clear that we have a bijection with the $\mathbb{RP}^2$. The continuity follows from the definition of quotient topology. – Popyaitte Dec 11 '19 at 16:51
  • 2
    I agree with @freakish, although it could be that just ONE point on the equator is identified with its antipode (although in that case, I'd have said "the north and south poles are identified). If it IS that latter case, then the answer's $\Bbb Z$; if it's what freakish described, it's two projective planes with an essential circle in each identified with an essential circle in the other; my quick computation says $\Bbb Z/2Z$ for that, by Seifert-van Kampen. – John Hughes Dec 11 '19 at 16:56
  • @MeesdeVries I will edit the question with a photo of the question from the book .... because I do not know the answer to your question. – Emptymind Dec 11 '19 at 19:43
  • @MeesdeVries I think the answer of your question (according to the statement of the problem) is that all points are identified with their antipodal point ..... am I correct? – Emptymind Dec 11 '19 at 19:47

1 Answers1

4

To make my notation clear I will state the problem like this:

Let $S^2 \subset \mathbb{R}^3$ be the unit sphere, and consider the equator $E = \{(x, y, z) \in S^2\ |\ z = 0\}\cong S^1$. Then we construct the quotient space $X = S^2 / \sim$, where $v \sim w$ iff $v,w\in E$ and $w = -v$. Compute $\pi_1(X)$.

Just like John Hughes suggests in his hint, the result is actually two copies of $\mathbb{R}P^2$ attached along essential loops, and we can use Seifert-van Kampen to compute $\pi_1$.

Describing $X$ in terms of $\mathbb{R}P^2$'s:

Let $q\colon S^2 \to X$ denote the quotient map, and let $S= q(E)$. Then $S = \mathbb{R}P^1 \cong S^1$.

If we restrict the quotient map $q\colon S^2 \to X$ to the upper hemisphere $D_+ = \{ (x, y, z) \in S^2\ | \ z \geq 0\}$, then topologically we're looking at a copy of $D^2$ whose boundary is glued to itself via the antipodal map, and therefore results in $\mathbb{R}P^2$. Let $q_+\colon H_+ \to X_+ \cong \mathbb{R}P^2$ denote the restricted quotient map, and notice $S= q_+(E)$. We can model this as a CW complex by attaching a $D^2$ to $S^1$ (representing $S$) by a map of degree $2$ on the boundary. If we restrict the quotient map to the lower hemisphere $q_-\colon D_- \to X_-$ then again we get a copy of $\mathbb{R}P^2$, and we have $X = X_+\cup X_-$ and $X_+ \cap X_- = q(E) \cong \mathbb{R}P^1$. In other words, $X$ is a union of two subspaces homeomorphic to $\mathbb{R}P^2$ whose intersection is the common subspace $\mathbb{R}P^1$. The CW model for $X$ then looks like a circle representing $S$ and then two disks attached to it via maps of degree $2$ on their boundaries.

Computing $\pi_1(X)$:

If $N \in X$ denotes the image of the north pole, then we can decompose $X$ into open sets $U_1 = q(int(D_+))\simeq\ast$ and $U_2 = X \setminus \{N\}\simeq q(D_-) \cong \mathbb{R}P^2$, where $U_1 \cap U_2 \simeq S^1$ (see Fig 1). Therefore the Seifert-van Kampen theorem describes $\pi_1(X)$ as the free product

$$ \pi_1(X) \cong \pi_1(\ast)\ast_{\pi_1(S^1)} \pi_1(\mathbb{R}P^2) \cong 1\ast_{\mathbb{Z}}\mathbb{Z}/2 , $$

where the homomorphism $\pi_1(S^1) \to \pi_1(\mathbb{R}P^2)$ is trivial because a generator of $\pi_1(U_1\cap U_2)$ wraps twice around $S$, and this loop is null-homotopic in $q(D_-)$ since $\pi_1(\mathbb{R}P^2)\cong \mathbb{Z}/2$. Therefore the free product is

$$ \pi_1(X) \cong \mathbb{Z}/2.$$

Fig 1.

schematic diagram of quotient map, U_1 and U_2

William
  • 9,310
  • And what is the meaning of the word essential in "essantial loops"? – Emptymind Dec 12 '19 at 00:20
  • I do not understand your $U_{1}$ and $U_{2}$ could you draw them please? – Emptymind Dec 12 '19 at 01:36
  • 1
    I added a schematic drawing of $X$, along with what my open sets look like. An "essential" loop in a topological space is just a loop which is not null-homotopic. For example, up to homotopy $\mathbb{R}P^2$ has exactly one essential loop, and it is represented by $\mathbb{R}P^1\subset \mathbb{R}P^2$ – William Dec 12 '19 at 02:04
  • https://math.stackexchange.com/questions/3473107/may-be-a-typo-in-the-solution-of-problem-10-parta-in-section-2-2-of-at?noredirect=1#comment7140275_3473107 could you please justify this solution for me? – Emptymind Dec 12 '19 at 02:21
  • 1
    Thanks for fleshing out the half-baked S-vK proof I had in mind, and the nice picture. :) – John Hughes Dec 12 '19 at 12:49