Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

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

Informal argument
So, we know that $\mathbb{R}$ with its usual topology is connected, and of course $\mathbb{R}\setminus\{0\}$ is not. Any $``$supposed" isomorphism should grant us a homeomorphism from $\mathbb{R}$ with its usual topology to $\mathbb{R}\setminus\{0\}$ with the topology it inherits from $\mathbb{R}$. So we conclude that they can't be isomorphic.

More Formal Argument
Suppose a bijection $i: \mathbb{R} \to \mathbb{R}\setminus \{0\}$ is such that $\forall x \forall y: x<y \longrightarrow f(x)<f(y)$, and additionally, $\forall a \forall b: a<b \longrightarrow f^{-1}(a)<f^{-1}(b)$. Extract a subsets $\Psi$ and $\xi$ of $\mathbb{R}\setminus\{0\}$ such that: $$\Psi = \{x\in \mathbb{R}\setminus \{0\} \mid x<-\frac{1}{n} \forall n\in \mathbb{N} \}$$ $$ \xi = \{x\in \mathbb{R}\setminus \{0\} \mid x>\frac{1}{n} \forall n \in \mathbb{N}\} $$ It is clear that $\Psi \cup \xi = \mathbb{R}\setminus \{0\}$. It is also the case that $\Psi$, $\xi$ are open in the topology that $\mathbb{R}\setminus \{0\}$. Since $i^{-1}$ is an order preserving bijection, $i^{-1}$ grants us that $\mathbb{R}$ is the union of two disjoint sets: $i^{-1}(\Psi) \cup i^{-1}(\xi)$. It follows that the preimage of $\Psi$, $\xi$ are open in $\mathbb{R}$ and $\mathbb{R}$ is the union of two disjoint open sets, a contradiction.

My question
Is there a way to avoid talking about topological notions of connectedness, or at least phrase connectedness in the language of linear orders? This is an extracurricular problem from my course in set theory and logic and it feels as if I'm cheating.

For example, we have the notion elementarily equivalence. If two linear orders are isomorphic then they are elementarily equivalent. The contrapositive gives us if two linear orders are not elementarily equivalent, then they are not isomorphic. So I guess, my real question is: does there exist a sentence $\phi$ in the language of linear orders that says something about connectedness? I mean surely $(\mathbb{R},<)$ satisfies the axioms of dense linear orders, does $(\mathbb{R}\setminus \{0\}, <)$ $\models \forall x \forall y \exists z (x<z<y)$? (I believe it does). What sort of sentence should I use?

We desire to use the fact that $\mathbb{R}$ is a complete linear order, in that every subset that is bounded above, has a least upper bound. $\mathbb{R}\setminus\{0\}$ does not have this property, as Goos's answer shows. Thus, they are not isomorphic as linear orders.

share|cite|improve this question
You shouldn't need to bring topology in to it. I think the notion you are looking for is completeness - in $\mathbb R$, every bounded-below set has a greatest lower bound. – Anthony Carapetis May 5 '14 at 6:31
@Alex: Your edit overran mine; and I think [order-theory] is better than [logic]+[elementary-set-theory]. – Asaf Karagila May 5 '14 at 6:39
@AsafKaragila Makes sense, I agree. – Alex Becker May 5 '14 at 6:40
@AnthonyCarapetis Psh! damn. >_< (face palms) – Rustyn May 5 '14 at 6:40
As ordered sets, $\mathbb R$ and $\mathbb R\setminus\{0\}$ are elementarily equivalent to each other and to $\mathbb Q$. Every countable densely ordered set without greatest or least element is isomorphic to $\mathbb Q$ (Cantor), and every densely ordered set without greatest or least element is elementarily equivalent to $\mathbb Q$. – bof May 5 '14 at 7:10
up vote 10 down vote accepted

In $\mathbb{R} \setminus \{0\}$, the set $[-1, 0)$ has an upper bound, but no least upper bound. No such set exists in the linearly ordered $\mathbb{R}$.

share|cite|improve this answer
Hey thanks, much appreciated. I should have realized this-- ARG – Rustyn May 5 '14 at 6:42
I'm guessing this is a second order property? Do you also have an example of two uncountable dense linear orders without endpoints that doesn't satisfy the same first order properties? – mvcouwen Jun 23 '14 at 22:06
@mvcouwen I am not sure about whether it is first or second order, but I also guess it is second order. I suggest you ask a new question. – 6005 Jun 24 '14 at 0:57
@mvcouwen I posted your question here:… – Rustyn Sep 30 '14 at 20:18

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.