1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Topology proofs

  1. Aug 2, 2008 #1
    i've texed up three proofs in from elementary topology. can someone please check them?

    actually i'll just retype them here for convenience


    Let [itex] f: X_{\tau} \rightarrow Y_{\nu} [/itex] be continuous and injective. Also let [itex] Y_{\nu} [/itex] be Hausdorff.

    Prove : [itex] X_{\tau}[/itex] is Hausdorff.

    Proof : Pick [itex]y_1[/itex] and [itex] y_2 [/itex] in [itex] f(X) [/itex]. By injectivity of f there exist [itex]x_1 =[/itex]
    [itex]f^{-1}(y_1)[/itex] and [itex] x_2 = f^{-1}(y_2) [/itex] such that they are both unique. [itex] f(X) [/itex] is Hausdorff so by Theorem 1 there exist disjoint open neighborhoods [itex] U [/itex] and [itex] V [/itex] of [itex]y_1[/itex] and [itex] y_2 [/itex] respectively. Then [itex] f^{-1}(U \bigcap V) = f^{-1}(U) \bigcap f^{-1}(V) = \emptyset [/itex] and by continuity [itex] f^{-1}(U) [/itex] and [itex] f^{-1}(V) [/itex] are open. Finally by definition of [itex] f^{-1} : x_1 \in f^{-1}(U) [/itex] and [itex] x_2 \in f^{-1}(V) [/itex] which, as stated previously, are two open disjoint sets in [itex] X_{\tau} [/itex]. Hence [itex] X_{\tau} [/itex] is Hausdorff.\\


    Injectivity is necessary. Take for example [itex] X = \{a,b\} [/itex] and [itex] Y = \{a\} [/itex] and to be [itex] f: X_{\mathcal{I}} \rightarrow Y_{\mathcal{I}} [/itex]. Explicitly [itex] f(\{a,b\}) =\{a}\}[/itex]. f is continuous, [itex] Y_{\mathcal{I}} [/itex] is obviously Hausdorff and [itex] X_{\mathcal{I}} [/itex] is obviously not.\\

    This does not prove that Hausdorff is a strong topological property because we have proven a stronger converse. To prove that Hausdorff is a strong topological property we would have to have proven that [itex] f: X_{\tau} \rightarrow Y_{\nu} [/itex] continuous, not necessarily injective, and [itex] X_{\tau}[/itex] Hausdorff implies [itex] f(X) [/itex] Hausdorff.


    Let [itex] X_{\tau}[/itex] be [itex] T_1 [/itex] and [itex] A \subseteq T [/itex] and [itex] x \in A' [/itex].\\

    Prove : Any neighborhood of x intersects \textit{A} in infinitely many points.

    Proof : Assume that there exists a neighborhood of x, in [itex] X_{\tau}[/itex] that intersects \textit{A} in only finitely many points to derive a contradiction. Let [itex] N_x [/itex] be such a neighborhood. [itex] N_x [/itex] is [itex] T_1 [/itex] by Theorem 1. Hence we can seperate x from all points in [itex]N_x [/itex] by other neighborhoods. Since there are finitely many points in [itex] N_x [/itex] there are finitely many such neighborhoods. Let \textit{N} be the intersection of those neighborhoods. [itex] N -\{x\}[/itex] therefore is itself a non-trivial open neighborhood of \textit{x} which does not intersect \textit{A}. This contradicts that x is a limit point of \textit{A} and therefore any neighborhood of x intersects \textit{A} in infinitely many points.


    Let [itex] A \subseteq X_{\tau}[/itex] and [itex]X_{\tau}[/itex] be regular.\\

    Prove : A is regular.

    Proof : Pick [itex] A_1 \subseteq A [/itex], [itex]A_1[/itex] closed in the subspace topology, and [itex] x \in A - A_1 [/itex]. Then [itex]A_1 = B[/itex] for some [itex]B \in \tau[/itex] and we can find two open sets [itex]N_B[/itex] and [itex]N_x[/itex] by the the regularity of [itex]X_{\tau}[/itex] which are disjoint. [itex]N_B \bigcap A[/itex] and [itex]N_x \bigcap A[/itex] are two disjoint sets in A which contain [itex]A_1[/itex] and x respectively. Therefore A is regular.

    Theorem 1 : A subspace of a [itex]T_i[/itex] space for [itex]i \leq 2[/itex] is [itex]T_i[/itex].

    Attached Files:

    • ch8.pdf
      File size:
      44.8 KB
  2. jcsd
  3. Aug 3, 2008 #2
    I'm not an expert at all... but in 8.2.5 shouldn't you say "pick [tex]y_{1} \neq y_{2}[/tex] in [tex]f(X)[/tex]?
  4. Aug 3, 2008 #3
    yea you're right
  5. Aug 3, 2008 #4
    When you want to prove that X is hausdorff I think it is nicer to say: Pick [itex] x_1 \neq x_2[/itex] in X, and then do as you do. It is because when it is X you want to prove is hausdorf I think you should start there, because you have to prove it for any two x'es in X.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook