I see that both ${\Bbb R}/{\Bbb Z}$ and ${\Bbb R}/{2\Bbb Z}$ are isomorphic to $S^1$. But when I apply the third isomorphism theorem I get ${\Bbb R}/{\Bbb Z}≈\frac {{\Bbb R}/{2\Bbb Z}}{{\Bbb Z}/{2\Bbb Z}}$ i.e. ${\Bbb R}/{\Bbb Z}≈\frac {{\Bbb R}/{2\Bbb Z}}{\Bbb Z_2}$. So if the claim in the title is true, I get ${\Bbb R}/{\Bbb Z}≈\frac {{\Bbb R}/{\Bbb Z}}{\Bbb Z_2}$, which seems incorrect to me.

Since $\Bbb R$ is Abelian, all the quotient sets are groups.

  • 4
    $\begingroup$ I think the first step is to clarify what $S_1 / \mathbb{Z}_2$ is. To start, ask what is the embedding $\mathbb{Z}_2 \to S_1$. $\endgroup$ – Carl Mummert Mar 12 '18 at 18:55

This is not a problem at all.

Consider the surjective group morphism $f: S^1 \to S^1$ $$f(z)=z^2$$ The kernel is $\{ 1 ; -1\} \cong \Bbb Z_2$ and by the first isomorphism theorem $$S^1 \cong S^1 / \{ 1 ; -1\}$$ which is exactly what you found.


It actually is correct! Taking $S^1$ and identifying antipodes gives $S^1$ again. This explains why $\mathbb{R}P^1$ is homeomorphic to $S^1$, and what you have is the group-theoretic version of this.

You have to be careful about how you are realizing $\mathbb{Z}_2$ as a subgroup of the circle. The sensible thing is as the multiplicative group $\{\pm 1\}$, and the cosets are then antipodal pairs, which gives my first paragraph above.

  • 1
    $\begingroup$ I understand the part where you said {$1,-1$} should be taken into consideration since it is the subgroup of $S^1$. However I am unfamiliar with most of the terms you used. $\endgroup$ – Hrit Roy Mar 12 '18 at 19:00
  • $\begingroup$ Understood, it is background-specific. Some of the other answers get it differently via homomorphisms, which may suit you better. $\endgroup$ – Randall Mar 12 '18 at 19:01
  • $\begingroup$ The circle group has only one element of order 2, so there's only one way to realize $\mathbb{Z}_2$ as a subgroup of the circle group. ​ ​ $\endgroup$ – user57159 Mar 13 '18 at 2:28
  • $\begingroup$ @RickyDemer yeah, you're totally right. $\endgroup$ – Randall Mar 13 '18 at 2:36

Indeed, if you consider $S^1$ as the multiplicative group $\{ z\in\mathbb C : \lvert z\rvert = 1\}$, then $H=\{+1,-1\}$ is a subgroup isomorphic to $\mathbb Z/2\mathbb Z$ and you have an isomorphism \begin{align*} S^1/H &\longrightarrow S^1, \\ [z] &\longmapsto z^2, \end{align*} where $[z]=zH=\{z,-z\}$.


Elements of $\mathbb{R}/2 \mathbb{Z}$ are of the form $x + 2\mathbb{Z}$ for $x \in \mathbb{R}$ and a complete set of representatives is given by choosing $x$ from $[0,2)$, although many other choices are possible.

Elements of $\mathbb{Z}/2 \mathbb{Z}$ are of the form $z + 2\mathbb{Z}$ for $z \in \mathbb{Z}$ and a complete set of representatives is given by choosing $z$ from $\{0,1\}$.

Elements of $G = \frac{\mathbb{R}/2 \mathbb{Z}}{\mathbb{Z}/2 \mathbb{Z}}$ are apparently, then, of the form $x + 2\mathbb{Z} + \mathbb{Z}/2\mathbb{Z}$. We may find all the elements of the equivalence class containing $x$ by \begin{align*} \bigcup_{z \in \mathbb{Z}/2\mathbb{Z}} & x + 2\mathbb{Z} + z \\ &= \left( x + 2\mathbb{Z} + (0 + 2\mathbb{Z}) \right) \cup \left( x + 2\mathbb{Z} + (1 + 2\mathbb{Z}) \right) \\ &= \left( x + 2\mathbb{Z} \right) \cup \left( x + 1 + 2\mathbb{Z} \right) \\ &= x + (2\mathbb{Z} \cup (1 + 2\mathbb{Z}) ) \\ &= x + \mathbb{Z} \text{.} \end{align*}

That is, the equivalence classes of the quotient defining $G$ are the same as those in $\mathbb{R}/\mathbb{Z}$ (and addition goes through without complications). So $G \cong \mathbb{R}/\mathbb{Z}$.


You're approaching the problem from the wrong side, in my opinion. The problem you're facing is due to how you identify $\mathbb{Z}/2\mathbb{Z}$ with a subgroup of $\mathbb{R}/\mathbb{Z}$ after applying the isomorphism.

There's a simpler way to proceed. Consider the (additive) group homomorphism $\mu\colon\mathbb{R}\to\mathbb{R}$ defined by $\mu(x)=x/2$. This is of course an isomorphism. Compose it with the canonical projection $\pi\colon\mathbb{R}\to\mathbb{R}/\mathbb{Z}$.

Then $\pi\circ\mu$ is surjective and $$ \ker(\pi\circ\mu)=\{x\in\mathbb{R}:\mu(x)\in\mathbb{Z}\}=2\mathbb{Z} $$ so the (first) homomorphism theorem provides $$ \mathbb{R}/2\mathbb{Z}= \mathbb{R}/\ker(\pi\circ\mu)\cong \mathbb{R}/\mathbb{Z} $$ If instead of $\mu(x)=x/2$ we use $\mu(x)=x/k$, where $k$ is any nonzero real, we get similarly that $$ \mathbb{R}/k\mathbb{Z}\cong\mathbb{R}/\mathbb{Z} $$


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.