π-base: Difficulty 6

Difficulty: 6

For proofs that are somewhat challenging. Maybe it takes a while even after a hint.

10

(Extremally disconnected ∧ Locally Hausdorff)     \implies Sequentially discrete

Added:

Mar 14, 2026

Difficulty:

Contrapositively, suppose (xn)(x_n) is a sequence with xnpx_n \to p yet has infinitely many terms. If needed, take an injective subsequence. If VV is a nbd of pp that is T2T_2, then there’s a n0n_0 with xnVx_n \in V for nn0n \ge n_0. So by another subsequence, we can assume (xn)(x_n) is an injective converging sequence in a T2T_2 space.

For each nn, it’s possible to construct a neighborhood VnV_n which only contains xnx_n and no other xmx_m, nor pp: Let xnVx_n \in V and pUp \in U with UV=U \cap V = \emptyset. xmUx_m \in U for all mn0m \ge n_0, so apply the Hausdorff condition finitely many times to terms xm<n0x_m < n_0 to construct such VnVV_n \subseteq V.

Now note that x2npx_{2n} \to p and x2n+1px_{2n+1} \to p, so if E=V2nE = \bigcup V_{2n} and O=V2n+1O = \bigcup V_{2n+1}, then any nbd of pp must intersect EE and OO, so pEOp \in \cl{E} \cap \cl{O}. Yet by definition, EO=E \cap O = \emptyset, so the space is not extremally disconnected.

59

Sequential     \implies k2k_2-space

Added:

Mar 23, 2026

Difficulty:

Let AXA \subseteq X such that, for any KK compact T2T_2 and f:KXf : K \to X continuous, then f1(A)f^{-1}(A) is closed. We wish to show AA is closed. To do that, we take a sequence (xn)(x_n) in A such that xnxx_n \to x and show xAx \in A (using that XX is sequential).

We use a nice trick: ω+1\omega + 1 is compact and T2T_2 under the order topology. If we define f:ω+1Xf : \omega + 1 \to X as f(n)=xnf(n) = x_n and f(ω)=xf(\omega) = x, then ff is continuous: Note that ωω+1\omega \subseteq \omega + 1 is discrete, so clearly it is continuous at each n<ωn < \omega, and for any nbd UU of xx, there is some n0n_0 such that xnUx_n \in U for n>n0n > n_0, so f((n0,ω])Uf((n_0, \omega]) \subseteq U and ff is continuous at ω\omega. This proves f1(A)f^{-1}(A) is closed. xAx \notin A would imply f1(A)=ωf^{-1}(A) = \omega, which is not a closed set (ωω\omega \in \cl{\omega}).

194

CGWH     \implies KC

Added:

Mar 26, 2026

Difficulty:

Let LL be compact. Using the k2k_2-space property, we wish to show that for any KK compact T2T_2 and any f:KXf : K \to X continuous, that f1(L)f^{-1}(L) is closed. Our objectives are to show (1) XX is T1T_1 (2) f(K)f(K) is compact T2T_2 (3) prove f1(L)f^{-1}(L) is closed.

(1) The inclusion map id:{x}X\text{id} : \{x\} \to X is continuous and {x}\{x\} is clearly compact T2T_2, so apply Weak Hausdorff.

(2) f(K)f(K) is compact since ff is continuous. Note that T2T_2 spaces separate compact sets, and since every closed set in KK is compact, KK is T4T_4. Take xyf(K)x \ne y \in f(K). {x}\{x\} and {y}\{y\} are closed in XX by (1), so A=f1({x})A = f^{-1}(\{x\}) and B=f1({y})B = f^{-1}(\{y\}) are closed and can be separated by open nbds AUA \subseteq U and BVB \subseteq V. Note that KUK \setminus U is compact, so f(KU)f(K \setminus U) is closed and U=f(K)f(KU)U' = f(K) \setminus f(K \setminus U) is a nbd of xx. Similarly, V=f(K)f(KV)V' = f(K) \setminus f(K \setminus V) is a nbd of yy and UV=U' \cap V' = \emptyset.

(3) L=f(K)LL' = f(K) \cap L is compact. Since f(K)f(K) is T2T_2, LL' is closed in f(K)f(K) But f(K)f(K) is closed in XX by Weak Hausdorff, so LL' is closed. Then clearly f1(L)=f1(L)f^{-1}(L) = f^{-1}(L') is closed.

321

kω,1k_{\omega,1}-space     \implies Hemicompact

Added:

Apr 2, 2026

Difficulty:

By definition, let XX be a countable union of increasing compact sets {Kn}\{K_n\}. Now define the sequence of sets

Ln={xX : {x}Kn}L_n = \{ x \in X \ : \ \cl{\{x\}} \cap K_n \ne \emptyset \}

Clearly by construction, KnLnK_n \subseteq L_n for each nn, {Ln}\{L_n\} is increasing, and X=n<ωLnX = \bigcup_{n < \omega} L_n. Now let U\mathcal{U} be an open cover of LnL_n. Since it also covers KnK_n, let {U1,,Um}\{U_1, \dots, U_m\} be a countable subcover of KnK_n. For any xLnx \in L_n, there is some y{x}Kny \in \cl{\{x\}} \cap K_n. Thus, yUjy \in U_j for some 1jm1 \le j \le m, but that implies xUjx \in U_j. So this is also a finite subcover of LnL_n, proving each LnL_n is compact.

By contraposition, suppose KK is not contained in any LnL_n. So choose xnXLnx_n \in X \setminus L_n for each nn, so that {xn}Kn=\cl{\{x_n\}} \cap K_n = \emptyset. Then define Vn=Xk=n{xk}V_n = X \setminus \bigcup_{k=n}^\infty \cl{\{x_k\}}. We see that {xk}\cl{\{x_k\}} will only intersect KmK_m for k<mk < m, so that VnKm=Kmk=nm1{xk}V_n \cap K_m = K_m \setminus \bigcup_{k=n}^{m-1} \cl{\{x_k\}}, which is open in KmK_m. By the kω,1k_{\omega,1}-space property, {Vn}\{V_n\} is an open cover of KK: for any xx, xKnx \in K_n for some nn and so x{xn}    xVnx \notin \cl{\{x_n\}} \implies x \in V_n. Note that it is also increasing, which means if it had a finite subcover, we’d have KVnK \subseteq V_n for some nn, which would imply xnKx_n \notin K. So KK is not compact.

Source: M W on stack exchange, awesome trick