Question. Is there a continuous function $f:\mathbb R^\omega\to\mathbb R$ whose restriction $f|\mathbb Q^\omega$ is injective?
-
2$\begingroup$ What topology is put on $\mathbb R^\omega$? $\endgroup$– Fan ZhengCommented Sep 3, 2017 at 3:34
-
2$\begingroup$ @FanZheng $\mathbb R^\omega$ carries the Tychonoff product topology. $\endgroup$– Taras BanakhCommented Sep 3, 2017 at 3:43
1 Answer
It looks like no.
Assume the contrary. We may start with two distinct rationals $q_1,p_1$ such that the sets $f(q_1\times \mathbb{R}^{\omega-1})=f(\{(q_1,\cdot,\cdot,\dots)\})$ and $f(p_1\times \mathbb{R}^{\omega-1})$ intersect. Indeed, for any $p_1$ the set $f(p_1\times \mathbb{R}^{\omega-1})$ has non-empty interior, choose reals such that $f(p_1,x_2,x_3,\dots)$ is interior point of this set. After that any $q_1$ close enough to $p_1$ works, since $f(q_1,x_2,\dots)$ is close to $f(p_1,x_2,\dots)$.
After that choose small intervals $\Delta_2,\Delta_3,\dots$ and $\delta_2,\delta_3,\dots$ such that the sets $f(q_1\times \Delta_2\times \Delta_3\times \dots)$ and $f(p_1\times \delta_2\times \delta_3\times \dots)$ intersect and have diameter less than 1. After that choose $q_2\in \Delta_2$ and $p_2\in \delta_2$ such that $f(q_1\times q_2\times \Delta_3\times \dots)$ and $f(p_1\times p_2\times \delta_3\times \dots)$ intersect. Proceeding this way, we get two rational sequences with the same value of $f$.
-
$\begingroup$ Thank you for the nice and short solution. I also arrived to negative answer and started to write it (using Baire Theorem), but your solution is much shoter and nice. $\endgroup$ Commented Sep 3, 2017 at 6:47
-
$\begingroup$ Wait, why does for any $p_1$ the set $f(p_1\times\mathbb{R}^{\omega-1})$ have non-empty interior? Isn't the first projection $f\colon(p_n)_{n\in\omega}\,\mapsto p_1$ a counterexample? Or are you using injectivity of $f$ on $\mathbb{Q}^\omega$, but how? (I remember convincing myself that your argument was correct, but now I all confused.) $\endgroup$– Gro-TsenCommented Sep 17, 2018 at 14:59
-
$\begingroup$ @Gro-Tsen by injectivity it contains two different points and by continuity on the segment between them we get the whole segment $\endgroup$ Commented Sep 17, 2018 at 15:38
-
$\begingroup$ Oh yes of course, silly question. Thanks. $\endgroup$– Gro-TsenCommented Sep 17, 2018 at 17:26