-1

First of all let's we give the following definition.

Definition

A space $X$ is weakly locally connected if for each neighborhood $N_x$ of any $x\in X$ there exist a connected subspace $F_x$ and a neighborhood $U_x$ such that $$ U_x\subseteq F_x\subseteq N_x $$

So I am trying to understand if $f$ is a surjective, closed and continuous map from a space $X$ to a space $Y$ then this last is weakly locally connected whether the first is it.

So by continuity $f^{-1}[V]$ is open if $V$ is an open neighborhood of $y\in Y$ and thus by surjectivity for any $x\in f^{-1}[V]$ there exists a connected set $F_x$ and an open set $U_x$ such that $$ x\in U_x\subseteq F_x\subseteq f^{-1}[V] $$ So putting $$ F:=\bigcup_{x\in f^{-1}[y]}F_x $$ we observe that $$ U:=\bigcup_{x\in f^{-1}[y]}U_x $$ is an open set such that $$ f^{-1}[y]\subseteq U\subseteq F\subseteq f^{-1}[V] $$ so that by closedness $Y\setminus f[X\setminus U]$ is an open set such that $$ \tag{1}\label{1}y\in Y\setminus f[X\setminus U]\subseteq f[U]\subseteq f[F]\subseteq V $$ Now we observe that $y$ is in $f[F_x]$ for any $x\in f^{-1}[y]$ so that the intersection $$ \bigcap_{x\in f^{-1}[y]} f[F_x] $$ is not empty and thus the union $$ \bigcup_{x\in f^{-1}[y]}f[F_x] $$ is connected because by continuity $f[F_x]$ is connected for any $x\in f^{-1}[y]$. So by $\eqref{1}$ we infer that $Y$ is weakly locally connected.

So I ask if what I proved is true and thus if it is then I ask if I well proved it whereas I ask to disprove it with a counterexample. So could someone help me, please?

Paul Frost
  • 87,968
  • Are neigborhoods required to be open? – Paul Frost Dec 02 '22 at 18:06
  • 1
    @PaulFrost The definition I gave is from Munkres topology text: it seems to me that Munkres does not requires this. What do you think about? However any neighborhood contain an open neighborhood so that it seem to me irrelevant if the neighborhood int the definition are or are not open: am I wrong? Could you help me, please? Forgive my confusion. – Antonio Maria Di Mauro Dec 02 '22 at 18:08
  • This is equivalent to local connectedness – Jakobian Dec 02 '22 at 18:20
  • @Jakobian Well I say that $X$ is locally connected if any $x\in X$ has a local connected base, right? However now $U_x$ could be not equal to $F_x$ so that why this equivalent to local connectedness? could you explain please? Forgive my confusion. – Antonio Maria Di Mauro Dec 02 '22 at 18:30
  • 2
    @AnneBauval not for points, but if a space is connected im kleinen at every point, that's equivalent to local connectedness – Jakobian Dec 02 '22 at 18:46
  • 1
    So yes, what I said is true, you're wrong – Jakobian Dec 02 '22 at 18:46
  • @Jakobian Thank you. Should I delete my comment, or leave it because yours may then be useful to others? – Anne Bauval Dec 02 '22 at 18:49
  • @Jakobian So could you prove the equivalence? – Antonio Maria Di Mauro Dec 02 '22 at 18:49
  • 2
    @Jakobian Do not worry: here there is a proof. – Antonio Maria Di Mauro Dec 02 '22 at 18:52
  • 1
    Okay, so the question became to prove or disprove if $f$ is a surjective, closed and continuous map from $X$ to $Y$ the $Y$ is locally connected if $X$ is it, right? However we know that weakly locally connected (at every points) is equivalent to locally connected so that is is also possibile to prove or to disprove that $Y$ is weakly locally connected at every point, right? So is the statement true? – Antonio Maria Di Mauro Dec 02 '22 at 18:55
  • Your proof is correct, nicely done. – Jakobian Dec 02 '22 at 18:59
  • @Jakobian Okay, thanks for the check!!! ;-) – Antonio Maria Di Mauro Dec 02 '22 at 19:00
  • 1
    @AnneBauval You should not delete your comment. It is useful, even if the claim was wrong. – Paul Frost Dec 02 '22 at 20:32
  • 1
    @AntonioMariaDiMauro You are right, it is irrelevant whether neigborhoods are required to be open. However, Munkres requires openness. See p 96 of his "Topology". – Paul Frost Dec 02 '22 at 20:41
  • @PaulFrost Oh, perfect! I did not know Munkres did that: thanks for the information!!! :-) – Antonio Maria Di Mauro Dec 02 '22 at 20:47
  • 1
    @AntonioMariaDiMauro One suggestion: It is correct to say "... thus by surjectivity for any $x∈f^{−1}[V]$ ...", but it may be misleading because later you only consider $x∈f^{−1}[V]$. So I would say "thus by surjectivity for any $x∈f^{−1}[y]$". And perhaps you should write an answer to your own question including the fact that weakly locally connected is the same as locally connected (if $X$ is weakly locally connected at all points). If you know that, you get a more general theorem: See my answer to https://math.stackexchange.com/q/4567646. – Paul Frost Dec 03 '22 at 17:44
  • @PaulFrost First of all thanks for the edit and even more for your very nice reference!!! Anyway, as you suggested, I wrote an answer where I proved that weakly locally connected is just equivalent to local connectedness and thus where I showed that it is possible prove the stament with your nice reference: thanks very much for the assistance! I hope I gave a good answer...^-^ – Antonio Maria Di Mauro Dec 04 '22 at 10:28

1 Answers1

1

The proof is correct and actually it shows that if $f$ is a closed and continuous surjection from $X$ to $Y$ then $Y$ is locally connected provide $X$ is it because weakly locally connection is just locally connection has showed to follow.

First of all we give the following definition

Definition

For a point $x$ of a topological space $X$ the component $C_x$ of $X$ is the union of all connected subspace of $X$ containing it.

So immediately we observe that singletons are trivially connected so that $C_x$ is defined for any $x\in X$. Moreover we observe that if $C_1$ and $C_2$ are two components of $X$ with not empty intersection then their union $C_1\cup C_2$ would be a component of any $x_0\in C_1\cap C_2$ but this is absurd because $C_1$ or $C_2$ would not be the union of all connected subspace containing $x_0\in X$: so we conclude that the collection of connected components is a partition. Now we give the following definition.

Definition

A space $X$ is locally connected if any $x\in X$ has a local base $\mathcal B(x)$ of connected neighborhoods.

So let's we prove the following result bearing in mind that a subset $H$ of any subspace $Y$ of a space $X$ is connected with repsect $Y$ if and only if it is connected with respect $X$.

Theorem

For any topological space $X$ the following statements are equivalent:

  1. $X$ is locally connected;
  2. the component of the open sets are open;
  3. for each neighborhood $N_x$ of any $x\in X$ there exist a connected subspace $H_x$ and an open set $A_x$ such that $$x\in A_x\subseteq H_x\subseteq N_x$$

Proof.

  1. If $X$ is locally connected then any $x\in A$ with $A$ open has a connected neighborhood $N_x$ there contained. So the component $C_x$ of $x$ in $A$ is the greatest connected subspace of $A$ containing $x$ so that if $N_x$ is connected with respect $X$ and if it is contained in $A$ then the inclusion $$ N_x\subseteq C_x $$ must holds. Now we know that if $x_0\in C_x$ then surely the equality $$ C_x=C_{x_0} $$ holds so that by the above arguments we conclude there exists a neighborhood $N_0$ of $x_0$ contained in $C_x$ and thus finally we conclude that $C_x$ is open.
  2. If $A(x)$ is an open neighborhood of any $x\in X$ then the component $C_{A(x)}$ of $x$ in $A(x)$ is connected in $X$ so that it is there open and thus $C_{A(x)}$ is an open connected neighborhood of $x$ such that $$ C_{A(x)}\subseteq A(x) $$ which proves that $3$ holds: indeed if $N(x)$ is a neighborhod of $x$ then there exists an open set $A(x)$ such that $$ x\in A(x)\subseteq N(x) $$ so that we can put $$ A_x,H_x:=C_{A(x)} $$
  3. So we know that $A_x$ is an opne set containing $x$ and contained in $H_x$ so that $H_x$ is connected neighborhood of $x$ contained in $N_x$ and thus for any $N_0\in\mathcal N(x)$ the collection $$ \mathcal C(N_0):=\{N_x\in\mathcal N(x):N_x\,\text{contained in $N_0$ and connected}\} $$ is not empty and thus finally che collection $$ \mathcal C(x):=\bigcup_{N_x\in\mathcal N(x)}\mathcal C(N_x) $$ is a local connected base for $x$.

However, we observe that if $V\in\mathcal P(Y)$ is such that $f^{-1}[V]$ is open in $X$ then its complement is closed but actually the equality $$ f^{-1}[Y\setminus V]=f^{-1}[Y]\setminus f^{-1}[V]=X\setminus f^{-1}[V] $$ holds and thus by closedness and by surjectivity $Y\setminus V$ is closed, that is $V$ is open which means that $f$ is a final map or rather that $Y$ has the final topology induced by $f$: so actually the result can be proved by this more general theorem proved by Paul Frost.