Existence of a certain near-metric map on an ordered divisible abelian group
$begingroup$
Let $mathcal{M}=(M,0,+,<)$ be a linearly ordered divisible abelian group. Let's define an $mathcal{M}$-metric on $M$ to be a map $d:Mtimes M rightarrow M$ such that
(1) $forall x,yin M,, d(x,y)=0 Leftrightarrow x=y$,
(2) $forall x,yin M,, d(x,y)=d(y,x)$,
(3) $forall x, y, zin M, , d(x,y) leq d(x,z) + d(z,y)$.
Note that one such map defines a topology on $M$ in the usual way that metrics do.
Let $tau$ be the right half-open interval topology in $M$. Namely the topology with basis ${ [x,x+varepsilon) : , x, varepsilonin M, varepsilon>0}$.
In there some $mathcal{M}$-metric on M that generates $tau$?
If $mathcal{M}=(mathbb{R},0,+,<)$ then the answer is no since the topological space described would be separable and not second countable, hence not metrizable.
Any ideas/intuitions regarding the question, even if they are not full answers, are apreciated.
general-topology metric-spaces abelian-groups model-theory ordered-groups
$endgroup$
add a comment |
$begingroup$
Let $mathcal{M}=(M,0,+,<)$ be a linearly ordered divisible abelian group. Let's define an $mathcal{M}$-metric on $M$ to be a map $d:Mtimes M rightarrow M$ such that
(1) $forall x,yin M,, d(x,y)=0 Leftrightarrow x=y$,
(2) $forall x,yin M,, d(x,y)=d(y,x)$,
(3) $forall x, y, zin M, , d(x,y) leq d(x,z) + d(z,y)$.
Note that one such map defines a topology on $M$ in the usual way that metrics do.
Let $tau$ be the right half-open interval topology in $M$. Namely the topology with basis ${ [x,x+varepsilon) : , x, varepsilonin M, varepsilon>0}$.
In there some $mathcal{M}$-metric on M that generates $tau$?
If $mathcal{M}=(mathbb{R},0,+,<)$ then the answer is no since the topological space described would be separable and not second countable, hence not metrizable.
Any ideas/intuitions regarding the question, even if they are not full answers, are apreciated.
general-topology metric-spaces abelian-groups model-theory ordered-groups
$endgroup$
$begingroup$
So just to be clear: you're looking for a characterization of those ordered divisible abelian groups with the property that they are metrizable when equipped with the half-open interval topology?
$endgroup$
– Alex Kruckman
Jun 22 '17 at 14:17
$begingroup$
@AlexKruckman yes. Upfront the question is whether there are any at all.
$endgroup$
– Anguepa
Jun 22 '17 at 16:19
$begingroup$
@AlexKruckman A totally ordered divisible group $(M,0,+,<)$ is always densely ordered since, for every $x, y in M$, if $x<y$ then $x<frac{x+y}{2}<y$. Or maybe you meant that it could be not totally ordered? I was asumming that it was. I'll edit the question to make this explicit.
$endgroup$
– Anguepa
Jun 22 '17 at 19:25
$begingroup$
Oh, I missed that you were only asking about divisible groups! Sorry
$endgroup$
– Alex Kruckman
Jun 22 '17 at 19:44
add a comment |
$begingroup$
Let $mathcal{M}=(M,0,+,<)$ be a linearly ordered divisible abelian group. Let's define an $mathcal{M}$-metric on $M$ to be a map $d:Mtimes M rightarrow M$ such that
(1) $forall x,yin M,, d(x,y)=0 Leftrightarrow x=y$,
(2) $forall x,yin M,, d(x,y)=d(y,x)$,
(3) $forall x, y, zin M, , d(x,y) leq d(x,z) + d(z,y)$.
Note that one such map defines a topology on $M$ in the usual way that metrics do.
Let $tau$ be the right half-open interval topology in $M$. Namely the topology with basis ${ [x,x+varepsilon) : , x, varepsilonin M, varepsilon>0}$.
In there some $mathcal{M}$-metric on M that generates $tau$?
If $mathcal{M}=(mathbb{R},0,+,<)$ then the answer is no since the topological space described would be separable and not second countable, hence not metrizable.
Any ideas/intuitions regarding the question, even if they are not full answers, are apreciated.
general-topology metric-spaces abelian-groups model-theory ordered-groups
$endgroup$
Let $mathcal{M}=(M,0,+,<)$ be a linearly ordered divisible abelian group. Let's define an $mathcal{M}$-metric on $M$ to be a map $d:Mtimes M rightarrow M$ such that
(1) $forall x,yin M,, d(x,y)=0 Leftrightarrow x=y$,
(2) $forall x,yin M,, d(x,y)=d(y,x)$,
(3) $forall x, y, zin M, , d(x,y) leq d(x,z) + d(z,y)$.
Note that one such map defines a topology on $M$ in the usual way that metrics do.
Let $tau$ be the right half-open interval topology in $M$. Namely the topology with basis ${ [x,x+varepsilon) : , x, varepsilonin M, varepsilon>0}$.
In there some $mathcal{M}$-metric on M that generates $tau$?
If $mathcal{M}=(mathbb{R},0,+,<)$ then the answer is no since the topological space described would be separable and not second countable, hence not metrizable.
Any ideas/intuitions regarding the question, even if they are not full answers, are apreciated.
general-topology metric-spaces abelian-groups model-theory ordered-groups
general-topology metric-spaces abelian-groups model-theory ordered-groups
edited Jun 22 '17 at 19:29
Anguepa
asked Jun 22 '17 at 0:19
AnguepaAnguepa
1,406819
1,406819
$begingroup$
So just to be clear: you're looking for a characterization of those ordered divisible abelian groups with the property that they are metrizable when equipped with the half-open interval topology?
$endgroup$
– Alex Kruckman
Jun 22 '17 at 14:17
$begingroup$
@AlexKruckman yes. Upfront the question is whether there are any at all.
$endgroup$
– Anguepa
Jun 22 '17 at 16:19
$begingroup$
@AlexKruckman A totally ordered divisible group $(M,0,+,<)$ is always densely ordered since, for every $x, y in M$, if $x<y$ then $x<frac{x+y}{2}<y$. Or maybe you meant that it could be not totally ordered? I was asumming that it was. I'll edit the question to make this explicit.
$endgroup$
– Anguepa
Jun 22 '17 at 19:25
$begingroup$
Oh, I missed that you were only asking about divisible groups! Sorry
$endgroup$
– Alex Kruckman
Jun 22 '17 at 19:44
add a comment |
$begingroup$
So just to be clear: you're looking for a characterization of those ordered divisible abelian groups with the property that they are metrizable when equipped with the half-open interval topology?
$endgroup$
– Alex Kruckman
Jun 22 '17 at 14:17
$begingroup$
@AlexKruckman yes. Upfront the question is whether there are any at all.
$endgroup$
– Anguepa
Jun 22 '17 at 16:19
$begingroup$
@AlexKruckman A totally ordered divisible group $(M,0,+,<)$ is always densely ordered since, for every $x, y in M$, if $x<y$ then $x<frac{x+y}{2}<y$. Or maybe you meant that it could be not totally ordered? I was asumming that it was. I'll edit the question to make this explicit.
$endgroup$
– Anguepa
Jun 22 '17 at 19:25
$begingroup$
Oh, I missed that you were only asking about divisible groups! Sorry
$endgroup$
– Alex Kruckman
Jun 22 '17 at 19:44
$begingroup$
So just to be clear: you're looking for a characterization of those ordered divisible abelian groups with the property that they are metrizable when equipped with the half-open interval topology?
$endgroup$
– Alex Kruckman
Jun 22 '17 at 14:17
$begingroup$
So just to be clear: you're looking for a characterization of those ordered divisible abelian groups with the property that they are metrizable when equipped with the half-open interval topology?
$endgroup$
– Alex Kruckman
Jun 22 '17 at 14:17
$begingroup$
@AlexKruckman yes. Upfront the question is whether there are any at all.
$endgroup$
– Anguepa
Jun 22 '17 at 16:19
$begingroup$
@AlexKruckman yes. Upfront the question is whether there are any at all.
$endgroup$
– Anguepa
Jun 22 '17 at 16:19
$begingroup$
@AlexKruckman A totally ordered divisible group $(M,0,+,<)$ is always densely ordered since, for every $x, y in M$, if $x<y$ then $x<frac{x+y}{2}<y$. Or maybe you meant that it could be not totally ordered? I was asumming that it was. I'll edit the question to make this explicit.
$endgroup$
– Anguepa
Jun 22 '17 at 19:25
$begingroup$
@AlexKruckman A totally ordered divisible group $(M,0,+,<)$ is always densely ordered since, for every $x, y in M$, if $x<y$ then $x<frac{x+y}{2}<y$. Or maybe you meant that it could be not totally ordered? I was asumming that it was. I'll edit the question to make this explicit.
$endgroup$
– Anguepa
Jun 22 '17 at 19:25
$begingroup$
Oh, I missed that you were only asking about divisible groups! Sorry
$endgroup$
– Alex Kruckman
Jun 22 '17 at 19:44
$begingroup$
Oh, I missed that you were only asking about divisible groups! Sorry
$endgroup$
– Alex Kruckman
Jun 22 '17 at 19:44
add a comment |
1 Answer
1
active
oldest
votes
$begingroup$
To manage expectations, let me start with a simple observation. $M$ does not admit a translation-invariant $mathcal{M}$-metric. If $rho$ is a translation-invariant $mathcal{M}$-metric, then $rho(x, 0) = rho(-x, 0)$
for all $x$, which means that any open ball about $0$ is symmetric. Hence
the open set $[0, rightarrow)$ cannot contain such a ball. This suggests
that even if a metric exists, it may not be very useful.
The same reasoning applied to $mathbb{R}$ can be applied to many other cases. If $M$ admits an $mathcal{M}$-metric, then $d(M) = |M|$, i.e. $M$ does not have a dense subset of lower cardinality. Since the balls with centre and radius in a dense subset form a base for the topology, we have
$w(M) le d(M)d(M) = d(M)$. On the other hand we have $w(M) ge |M|$ for
the same reason as in $mathbb{R}$. (Note that here we do not have to distinguish between order-dense, dense in the
order topology and dense in the half-open interval topology, as they are
equivalent for a dense linear order.)
This necessary condition can be sharpened by an observation that is useful
for non-Archimedean groups. If $mathcal{N}$ is a nontrivial convex
subgroup of $mathcal{M}$, the following are equivalent:
- $M$ admits an $mathcal{M}$-metric
- $N$ admits an $mathcal{M}$-metric
- $N$ admits an $mathcal{N}$-metric
- $M$ admits an $mathcal{N}$-metric
That 4. implies 1. is trivial and 1. implies 2. because the restriction of a metric to a subspace induces the subspace topology.
The implication $2. implies 3.$ can be obtained by taking a positive $b inmathcal{N}$ and defining $sigma(x, y) = min{rho(x,y), b}$ as usual. Since $N$ is open in $M$, the quotient $M/N$ is discrete and
$M$ is homeomorphic to $(M/N) times N$. Hence, given an
$mathcal{N}$-metric on $N$, we can make $M$ locally isometric to $N$ to
arrive at 4.
Giving a useful necessary and sufficient condition is probably very hard,
but there is a sufficient condition that is not too difficult. If $mathcal{M}$ has a nontrivial countable convex subgroup, $M$ admits an
$mathcal{M}$-metric. By the previous result, we need only consider the
case $|M| = aleph_0$. Let $B$ be the set of real numbers between 0 and 1
whose binary expansions have finitely many ones. This is also a countable
dense linearly ordered set with no minimum or maximum, so it is order-isomorphic to $M$. It will suffice then to find an $mathcal{M}$-metric
that induces the half-open interval topology on $B$.
Since $M$ is countable, there is a strictly decreasing
sequence ${a_n}$ of positive elements with $inf_n a_n = 0$. Define
$rho(x, y) = 0$ when $x=y$ and otherwise $rho(x, y) = a_m$, where $m$ is the largest integer such that the binary expansions of $x$ and $y$ are
equal to $m$ places. Symmetry and positivity of $rho$ are obvious and it
is easily seen to satisfy the ultrametric triangle inequality
$rho(x, z) le max{rho(x, y),rho(y,z)}$.
To see that the topology induced by $rho$ is finer that the half-open interval topology, note that in the latter a point $x$ has a base of
neighbourhoods of the form $[x, y)$ where $y>x$. There is an $n$ such that
the binary expansions of $x$ and $y$ contain no ones after the $n$th place.
For such $n$, the $rho$-ball of radius $a_{n+1}$ about $x$ is contained in
$[x, y)$. Similarly, the topology induced by $rho$ is coarser than the
half-open interval topology, since every $rho$-ball about $x$ contains a
half-open interval starting at $x$. Thus $rho$ induces the half-open
interval topology on $B$.
Note that under the same hypothesis, $M$ also admits a real-valued metric,
by Urysohn's metrization theorem.
Side note: For topological purposes, it usually makes no difference
whether a ordered abelian group is divisible or just densely ordered. On one hand, in a densely ordered group there is for every $b > 0$ and every
positive integer $n$ an $a > 0$ such that $na <= b$. On the other hand,
in a divisible group there may be no $a > 0$ such that
$lim_{ntoinfty} a/n = 0$.
$endgroup$
$begingroup$
This is great. Thank you. So for example the right half-open interval topology on $(mathbb{Q},<)$ is $mathbb{Q}$-metrizable.
$endgroup$
– Anguepa
Apr 15 '18 at 19:28
$begingroup$
To prove the observation on densely ordered groups in the side note in Niels' answer let $0<a<b$ and note that either $2aleq b$ or $2(b-a)< b$.
$endgroup$
– Anguepa
Dec 18 '18 at 19:21
add a comment |
Your Answer
StackExchange.ifUsing("editor", function () {
return StackExchange.using("mathjaxEditing", function () {
StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix) {
StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
});
});
}, "mathjax-editing");
StackExchange.ready(function() {
var channelOptions = {
tags: "".split(" "),
id: "69"
};
initTagRenderer("".split(" "), "".split(" "), channelOptions);
StackExchange.using("externalEditor", function() {
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled) {
StackExchange.using("snippets", function() {
createEditor();
});
}
else {
createEditor();
}
});
function createEditor() {
StackExchange.prepareEditor({
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: true,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: 10,
bindNavPrevention: true,
postfix: "",
imageUploader: {
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
},
noCode: true, onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
});
}
});
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f2331814%2fexistence-of-a-certain-near-metric-map-on-an-ordered-divisible-abelian-group%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
1 Answer
1
active
oldest
votes
1 Answer
1
active
oldest
votes
active
oldest
votes
active
oldest
votes
$begingroup$
To manage expectations, let me start with a simple observation. $M$ does not admit a translation-invariant $mathcal{M}$-metric. If $rho$ is a translation-invariant $mathcal{M}$-metric, then $rho(x, 0) = rho(-x, 0)$
for all $x$, which means that any open ball about $0$ is symmetric. Hence
the open set $[0, rightarrow)$ cannot contain such a ball. This suggests
that even if a metric exists, it may not be very useful.
The same reasoning applied to $mathbb{R}$ can be applied to many other cases. If $M$ admits an $mathcal{M}$-metric, then $d(M) = |M|$, i.e. $M$ does not have a dense subset of lower cardinality. Since the balls with centre and radius in a dense subset form a base for the topology, we have
$w(M) le d(M)d(M) = d(M)$. On the other hand we have $w(M) ge |M|$ for
the same reason as in $mathbb{R}$. (Note that here we do not have to distinguish between order-dense, dense in the
order topology and dense in the half-open interval topology, as they are
equivalent for a dense linear order.)
This necessary condition can be sharpened by an observation that is useful
for non-Archimedean groups. If $mathcal{N}$ is a nontrivial convex
subgroup of $mathcal{M}$, the following are equivalent:
- $M$ admits an $mathcal{M}$-metric
- $N$ admits an $mathcal{M}$-metric
- $N$ admits an $mathcal{N}$-metric
- $M$ admits an $mathcal{N}$-metric
That 4. implies 1. is trivial and 1. implies 2. because the restriction of a metric to a subspace induces the subspace topology.
The implication $2. implies 3.$ can be obtained by taking a positive $b inmathcal{N}$ and defining $sigma(x, y) = min{rho(x,y), b}$ as usual. Since $N$ is open in $M$, the quotient $M/N$ is discrete and
$M$ is homeomorphic to $(M/N) times N$. Hence, given an
$mathcal{N}$-metric on $N$, we can make $M$ locally isometric to $N$ to
arrive at 4.
Giving a useful necessary and sufficient condition is probably very hard,
but there is a sufficient condition that is not too difficult. If $mathcal{M}$ has a nontrivial countable convex subgroup, $M$ admits an
$mathcal{M}$-metric. By the previous result, we need only consider the
case $|M| = aleph_0$. Let $B$ be the set of real numbers between 0 and 1
whose binary expansions have finitely many ones. This is also a countable
dense linearly ordered set with no minimum or maximum, so it is order-isomorphic to $M$. It will suffice then to find an $mathcal{M}$-metric
that induces the half-open interval topology on $B$.
Since $M$ is countable, there is a strictly decreasing
sequence ${a_n}$ of positive elements with $inf_n a_n = 0$. Define
$rho(x, y) = 0$ when $x=y$ and otherwise $rho(x, y) = a_m$, where $m$ is the largest integer such that the binary expansions of $x$ and $y$ are
equal to $m$ places. Symmetry and positivity of $rho$ are obvious and it
is easily seen to satisfy the ultrametric triangle inequality
$rho(x, z) le max{rho(x, y),rho(y,z)}$.
To see that the topology induced by $rho$ is finer that the half-open interval topology, note that in the latter a point $x$ has a base of
neighbourhoods of the form $[x, y)$ where $y>x$. There is an $n$ such that
the binary expansions of $x$ and $y$ contain no ones after the $n$th place.
For such $n$, the $rho$-ball of radius $a_{n+1}$ about $x$ is contained in
$[x, y)$. Similarly, the topology induced by $rho$ is coarser than the
half-open interval topology, since every $rho$-ball about $x$ contains a
half-open interval starting at $x$. Thus $rho$ induces the half-open
interval topology on $B$.
Note that under the same hypothesis, $M$ also admits a real-valued metric,
by Urysohn's metrization theorem.
Side note: For topological purposes, it usually makes no difference
whether a ordered abelian group is divisible or just densely ordered. On one hand, in a densely ordered group there is for every $b > 0$ and every
positive integer $n$ an $a > 0$ such that $na <= b$. On the other hand,
in a divisible group there may be no $a > 0$ such that
$lim_{ntoinfty} a/n = 0$.
$endgroup$
$begingroup$
This is great. Thank you. So for example the right half-open interval topology on $(mathbb{Q},<)$ is $mathbb{Q}$-metrizable.
$endgroup$
– Anguepa
Apr 15 '18 at 19:28
$begingroup$
To prove the observation on densely ordered groups in the side note in Niels' answer let $0<a<b$ and note that either $2aleq b$ or $2(b-a)< b$.
$endgroup$
– Anguepa
Dec 18 '18 at 19:21
add a comment |
$begingroup$
To manage expectations, let me start with a simple observation. $M$ does not admit a translation-invariant $mathcal{M}$-metric. If $rho$ is a translation-invariant $mathcal{M}$-metric, then $rho(x, 0) = rho(-x, 0)$
for all $x$, which means that any open ball about $0$ is symmetric. Hence
the open set $[0, rightarrow)$ cannot contain such a ball. This suggests
that even if a metric exists, it may not be very useful.
The same reasoning applied to $mathbb{R}$ can be applied to many other cases. If $M$ admits an $mathcal{M}$-metric, then $d(M) = |M|$, i.e. $M$ does not have a dense subset of lower cardinality. Since the balls with centre and radius in a dense subset form a base for the topology, we have
$w(M) le d(M)d(M) = d(M)$. On the other hand we have $w(M) ge |M|$ for
the same reason as in $mathbb{R}$. (Note that here we do not have to distinguish between order-dense, dense in the
order topology and dense in the half-open interval topology, as they are
equivalent for a dense linear order.)
This necessary condition can be sharpened by an observation that is useful
for non-Archimedean groups. If $mathcal{N}$ is a nontrivial convex
subgroup of $mathcal{M}$, the following are equivalent:
- $M$ admits an $mathcal{M}$-metric
- $N$ admits an $mathcal{M}$-metric
- $N$ admits an $mathcal{N}$-metric
- $M$ admits an $mathcal{N}$-metric
That 4. implies 1. is trivial and 1. implies 2. because the restriction of a metric to a subspace induces the subspace topology.
The implication $2. implies 3.$ can be obtained by taking a positive $b inmathcal{N}$ and defining $sigma(x, y) = min{rho(x,y), b}$ as usual. Since $N$ is open in $M$, the quotient $M/N$ is discrete and
$M$ is homeomorphic to $(M/N) times N$. Hence, given an
$mathcal{N}$-metric on $N$, we can make $M$ locally isometric to $N$ to
arrive at 4.
Giving a useful necessary and sufficient condition is probably very hard,
but there is a sufficient condition that is not too difficult. If $mathcal{M}$ has a nontrivial countable convex subgroup, $M$ admits an
$mathcal{M}$-metric. By the previous result, we need only consider the
case $|M| = aleph_0$. Let $B$ be the set of real numbers between 0 and 1
whose binary expansions have finitely many ones. This is also a countable
dense linearly ordered set with no minimum or maximum, so it is order-isomorphic to $M$. It will suffice then to find an $mathcal{M}$-metric
that induces the half-open interval topology on $B$.
Since $M$ is countable, there is a strictly decreasing
sequence ${a_n}$ of positive elements with $inf_n a_n = 0$. Define
$rho(x, y) = 0$ when $x=y$ and otherwise $rho(x, y) = a_m$, where $m$ is the largest integer such that the binary expansions of $x$ and $y$ are
equal to $m$ places. Symmetry and positivity of $rho$ are obvious and it
is easily seen to satisfy the ultrametric triangle inequality
$rho(x, z) le max{rho(x, y),rho(y,z)}$.
To see that the topology induced by $rho$ is finer that the half-open interval topology, note that in the latter a point $x$ has a base of
neighbourhoods of the form $[x, y)$ where $y>x$. There is an $n$ such that
the binary expansions of $x$ and $y$ contain no ones after the $n$th place.
For such $n$, the $rho$-ball of radius $a_{n+1}$ about $x$ is contained in
$[x, y)$. Similarly, the topology induced by $rho$ is coarser than the
half-open interval topology, since every $rho$-ball about $x$ contains a
half-open interval starting at $x$. Thus $rho$ induces the half-open
interval topology on $B$.
Note that under the same hypothesis, $M$ also admits a real-valued metric,
by Urysohn's metrization theorem.
Side note: For topological purposes, it usually makes no difference
whether a ordered abelian group is divisible or just densely ordered. On one hand, in a densely ordered group there is for every $b > 0$ and every
positive integer $n$ an $a > 0$ such that $na <= b$. On the other hand,
in a divisible group there may be no $a > 0$ such that
$lim_{ntoinfty} a/n = 0$.
$endgroup$
$begingroup$
This is great. Thank you. So for example the right half-open interval topology on $(mathbb{Q},<)$ is $mathbb{Q}$-metrizable.
$endgroup$
– Anguepa
Apr 15 '18 at 19:28
$begingroup$
To prove the observation on densely ordered groups in the side note in Niels' answer let $0<a<b$ and note that either $2aleq b$ or $2(b-a)< b$.
$endgroup$
– Anguepa
Dec 18 '18 at 19:21
add a comment |
$begingroup$
To manage expectations, let me start with a simple observation. $M$ does not admit a translation-invariant $mathcal{M}$-metric. If $rho$ is a translation-invariant $mathcal{M}$-metric, then $rho(x, 0) = rho(-x, 0)$
for all $x$, which means that any open ball about $0$ is symmetric. Hence
the open set $[0, rightarrow)$ cannot contain such a ball. This suggests
that even if a metric exists, it may not be very useful.
The same reasoning applied to $mathbb{R}$ can be applied to many other cases. If $M$ admits an $mathcal{M}$-metric, then $d(M) = |M|$, i.e. $M$ does not have a dense subset of lower cardinality. Since the balls with centre and radius in a dense subset form a base for the topology, we have
$w(M) le d(M)d(M) = d(M)$. On the other hand we have $w(M) ge |M|$ for
the same reason as in $mathbb{R}$. (Note that here we do not have to distinguish between order-dense, dense in the
order topology and dense in the half-open interval topology, as they are
equivalent for a dense linear order.)
This necessary condition can be sharpened by an observation that is useful
for non-Archimedean groups. If $mathcal{N}$ is a nontrivial convex
subgroup of $mathcal{M}$, the following are equivalent:
- $M$ admits an $mathcal{M}$-metric
- $N$ admits an $mathcal{M}$-metric
- $N$ admits an $mathcal{N}$-metric
- $M$ admits an $mathcal{N}$-metric
That 4. implies 1. is trivial and 1. implies 2. because the restriction of a metric to a subspace induces the subspace topology.
The implication $2. implies 3.$ can be obtained by taking a positive $b inmathcal{N}$ and defining $sigma(x, y) = min{rho(x,y), b}$ as usual. Since $N$ is open in $M$, the quotient $M/N$ is discrete and
$M$ is homeomorphic to $(M/N) times N$. Hence, given an
$mathcal{N}$-metric on $N$, we can make $M$ locally isometric to $N$ to
arrive at 4.
Giving a useful necessary and sufficient condition is probably very hard,
but there is a sufficient condition that is not too difficult. If $mathcal{M}$ has a nontrivial countable convex subgroup, $M$ admits an
$mathcal{M}$-metric. By the previous result, we need only consider the
case $|M| = aleph_0$. Let $B$ be the set of real numbers between 0 and 1
whose binary expansions have finitely many ones. This is also a countable
dense linearly ordered set with no minimum or maximum, so it is order-isomorphic to $M$. It will suffice then to find an $mathcal{M}$-metric
that induces the half-open interval topology on $B$.
Since $M$ is countable, there is a strictly decreasing
sequence ${a_n}$ of positive elements with $inf_n a_n = 0$. Define
$rho(x, y) = 0$ when $x=y$ and otherwise $rho(x, y) = a_m$, where $m$ is the largest integer such that the binary expansions of $x$ and $y$ are
equal to $m$ places. Symmetry and positivity of $rho$ are obvious and it
is easily seen to satisfy the ultrametric triangle inequality
$rho(x, z) le max{rho(x, y),rho(y,z)}$.
To see that the topology induced by $rho$ is finer that the half-open interval topology, note that in the latter a point $x$ has a base of
neighbourhoods of the form $[x, y)$ where $y>x$. There is an $n$ such that
the binary expansions of $x$ and $y$ contain no ones after the $n$th place.
For such $n$, the $rho$-ball of radius $a_{n+1}$ about $x$ is contained in
$[x, y)$. Similarly, the topology induced by $rho$ is coarser than the
half-open interval topology, since every $rho$-ball about $x$ contains a
half-open interval starting at $x$. Thus $rho$ induces the half-open
interval topology on $B$.
Note that under the same hypothesis, $M$ also admits a real-valued metric,
by Urysohn's metrization theorem.
Side note: For topological purposes, it usually makes no difference
whether a ordered abelian group is divisible or just densely ordered. On one hand, in a densely ordered group there is for every $b > 0$ and every
positive integer $n$ an $a > 0$ such that $na <= b$. On the other hand,
in a divisible group there may be no $a > 0$ such that
$lim_{ntoinfty} a/n = 0$.
$endgroup$
To manage expectations, let me start with a simple observation. $M$ does not admit a translation-invariant $mathcal{M}$-metric. If $rho$ is a translation-invariant $mathcal{M}$-metric, then $rho(x, 0) = rho(-x, 0)$
for all $x$, which means that any open ball about $0$ is symmetric. Hence
the open set $[0, rightarrow)$ cannot contain such a ball. This suggests
that even if a metric exists, it may not be very useful.
The same reasoning applied to $mathbb{R}$ can be applied to many other cases. If $M$ admits an $mathcal{M}$-metric, then $d(M) = |M|$, i.e. $M$ does not have a dense subset of lower cardinality. Since the balls with centre and radius in a dense subset form a base for the topology, we have
$w(M) le d(M)d(M) = d(M)$. On the other hand we have $w(M) ge |M|$ for
the same reason as in $mathbb{R}$. (Note that here we do not have to distinguish between order-dense, dense in the
order topology and dense in the half-open interval topology, as they are
equivalent for a dense linear order.)
This necessary condition can be sharpened by an observation that is useful
for non-Archimedean groups. If $mathcal{N}$ is a nontrivial convex
subgroup of $mathcal{M}$, the following are equivalent:
- $M$ admits an $mathcal{M}$-metric
- $N$ admits an $mathcal{M}$-metric
- $N$ admits an $mathcal{N}$-metric
- $M$ admits an $mathcal{N}$-metric
That 4. implies 1. is trivial and 1. implies 2. because the restriction of a metric to a subspace induces the subspace topology.
The implication $2. implies 3.$ can be obtained by taking a positive $b inmathcal{N}$ and defining $sigma(x, y) = min{rho(x,y), b}$ as usual. Since $N$ is open in $M$, the quotient $M/N$ is discrete and
$M$ is homeomorphic to $(M/N) times N$. Hence, given an
$mathcal{N}$-metric on $N$, we can make $M$ locally isometric to $N$ to
arrive at 4.
Giving a useful necessary and sufficient condition is probably very hard,
but there is a sufficient condition that is not too difficult. If $mathcal{M}$ has a nontrivial countable convex subgroup, $M$ admits an
$mathcal{M}$-metric. By the previous result, we need only consider the
case $|M| = aleph_0$. Let $B$ be the set of real numbers between 0 and 1
whose binary expansions have finitely many ones. This is also a countable
dense linearly ordered set with no minimum or maximum, so it is order-isomorphic to $M$. It will suffice then to find an $mathcal{M}$-metric
that induces the half-open interval topology on $B$.
Since $M$ is countable, there is a strictly decreasing
sequence ${a_n}$ of positive elements with $inf_n a_n = 0$. Define
$rho(x, y) = 0$ when $x=y$ and otherwise $rho(x, y) = a_m$, where $m$ is the largest integer such that the binary expansions of $x$ and $y$ are
equal to $m$ places. Symmetry and positivity of $rho$ are obvious and it
is easily seen to satisfy the ultrametric triangle inequality
$rho(x, z) le max{rho(x, y),rho(y,z)}$.
To see that the topology induced by $rho$ is finer that the half-open interval topology, note that in the latter a point $x$ has a base of
neighbourhoods of the form $[x, y)$ where $y>x$. There is an $n$ such that
the binary expansions of $x$ and $y$ contain no ones after the $n$th place.
For such $n$, the $rho$-ball of radius $a_{n+1}$ about $x$ is contained in
$[x, y)$. Similarly, the topology induced by $rho$ is coarser than the
half-open interval topology, since every $rho$-ball about $x$ contains a
half-open interval starting at $x$. Thus $rho$ induces the half-open
interval topology on $B$.
Note that under the same hypothesis, $M$ also admits a real-valued metric,
by Urysohn's metrization theorem.
Side note: For topological purposes, it usually makes no difference
whether a ordered abelian group is divisible or just densely ordered. On one hand, in a densely ordered group there is for every $b > 0$ and every
positive integer $n$ an $a > 0$ such that $na <= b$. On the other hand,
in a divisible group there may be no $a > 0$ such that
$lim_{ntoinfty} a/n = 0$.
answered Jul 31 '17 at 11:33
Niels J. DiepeveenNiels J. Diepeveen
5,8491931
5,8491931
$begingroup$
This is great. Thank you. So for example the right half-open interval topology on $(mathbb{Q},<)$ is $mathbb{Q}$-metrizable.
$endgroup$
– Anguepa
Apr 15 '18 at 19:28
$begingroup$
To prove the observation on densely ordered groups in the side note in Niels' answer let $0<a<b$ and note that either $2aleq b$ or $2(b-a)< b$.
$endgroup$
– Anguepa
Dec 18 '18 at 19:21
add a comment |
$begingroup$
This is great. Thank you. So for example the right half-open interval topology on $(mathbb{Q},<)$ is $mathbb{Q}$-metrizable.
$endgroup$
– Anguepa
Apr 15 '18 at 19:28
$begingroup$
To prove the observation on densely ordered groups in the side note in Niels' answer let $0<a<b$ and note that either $2aleq b$ or $2(b-a)< b$.
$endgroup$
– Anguepa
Dec 18 '18 at 19:21
$begingroup$
This is great. Thank you. So for example the right half-open interval topology on $(mathbb{Q},<)$ is $mathbb{Q}$-metrizable.
$endgroup$
– Anguepa
Apr 15 '18 at 19:28
$begingroup$
This is great. Thank you. So for example the right half-open interval topology on $(mathbb{Q},<)$ is $mathbb{Q}$-metrizable.
$endgroup$
– Anguepa
Apr 15 '18 at 19:28
$begingroup$
To prove the observation on densely ordered groups in the side note in Niels' answer let $0<a<b$ and note that either $2aleq b$ or $2(b-a)< b$.
$endgroup$
– Anguepa
Dec 18 '18 at 19:21
$begingroup$
To prove the observation on densely ordered groups in the side note in Niels' answer let $0<a<b$ and note that either $2aleq b$ or $2(b-a)< b$.
$endgroup$
– Anguepa
Dec 18 '18 at 19:21
add a comment |
Thanks for contributing an answer to Mathematics Stack Exchange!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
Use MathJax to format equations. MathJax reference.
To learn more, see our tips on writing great answers.
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f2331814%2fexistence-of-a-certain-near-metric-map-on-an-ordered-divisible-abelian-group%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
$begingroup$
So just to be clear: you're looking for a characterization of those ordered divisible abelian groups with the property that they are metrizable when equipped with the half-open interval topology?
$endgroup$
– Alex Kruckman
Jun 22 '17 at 14:17
$begingroup$
@AlexKruckman yes. Upfront the question is whether there are any at all.
$endgroup$
– Anguepa
Jun 22 '17 at 16:19
$begingroup$
@AlexKruckman A totally ordered divisible group $(M,0,+,<)$ is always densely ordered since, for every $x, y in M$, if $x<y$ then $x<frac{x+y}{2}<y$. Or maybe you meant that it could be not totally ordered? I was asumming that it was. I'll edit the question to make this explicit.
$endgroup$
– Anguepa
Jun 22 '17 at 19:25
$begingroup$
Oh, I missed that you were only asking about divisible groups! Sorry
$endgroup$
– Alex Kruckman
Jun 22 '17 at 19:44