I know this is part of the definition of a topology. But the book I am using which is 'Topology without Tears' ask to prove this using mathematical induction.I hate to ask questions but proofs I have seen for this question in this site were different than mine.

Here it is...

The statement "Intersection of any two sets in $T$ belongs to $T$ " is given by the definition.

Suppose $T_1,T_2,T_3,...$ are members of $T$ and $$\bigcap_{n=1}^{k} T_n\in T \label{a}\tag{1}$$ is true. If $\,\bigcap_{n=1}^{k+1} \,T_n\in T $ is also true, the proposition can be proven by mathematical induction.

Since the element in the the left side of (\ref{a}) is a also set, let $M=\bigcap_{n=1}^{k} T_n $.

$$\bigcap_{n=1}^{k+1} T_n = \bigcap_{n=1}^{k} T_n \cap T_{k+1} $$ $$\bigcap_{n=1}^{k+1} T_n = M \cap T_{k+1} $$

Since $M\in T $ by (1) and $T_{k+1}\in T $, $\,\bigcap_{n=1}^{k+1} T_n \in T$ by definition. Thus intersection of any finite number of sets in $T$ belongs to $T$.

I want to know every mistake of this proof or ways to improve it.

  • $\begingroup$ I'm not sure what other proofs you've seen, but I think this is the standard proof to show that a finite intersection of elements in a topology $\tau$ is again in $\tau$. $\endgroup$ – Clayton Mar 14 at 19:30
  • $\begingroup$ I see no mistakes. Your proof is fine. $\endgroup$ – Mark Mar 14 at 19:30
  • 1
    $\begingroup$ The proof is perfect. The reason you don't see this exact argument elsewhere is because everyone has a different style in writing and not everyone has the same base of facts at their disposal. $\endgroup$ – Alberto Takase Mar 14 at 19:38
  • $\begingroup$ If $0$ counts as finite (which I think it should), then you should mention that the intersection of $0$ subsets of $X$ is $X$, which is also in $T$. $\endgroup$ – Andreas Blass Mar 15 at 1:42
  • $\begingroup$ “I hate to ask questions but ...” Why? Asking questions and communicating with other mathematicians is what makes math what it is! $\endgroup$ – Santana Afton Mar 15 at 1:53

The proof by induction is fine. Renaming to $M$ is not really necessary of course (I don't usually introduce "unnecessary" notation), just writing

$$\bigcap_{i=1}^{n+1} T_i = \left(\bigcap_{i=1}^{n} T_i\right) \cap T_{n+1}$$

(so using braces) is enough to make your point, I'd say.

It's actually a very common pattern: e.g. groups are closed under products of two members, but writing finite powers etc. is routinely done: having a set closed under an operation allows all finite "products' (or sums/intersections, etc. etc.) You do the proof once, and henceforth it's implicit.

It sometimes makes for slightly easier to read proofs, only having to check the $n=2$ case.

A minor nitpick: you don't mention the base case (which is $n=1$), which is almost too trivial to mention: $T_1$ in the topology gives us the $1$-ary intersection $\bigcap_{i=1}^1 T_i = T_1$ in the topology too, tautologically. The $n=2$ case could also serve as the base case, depending on the formulation of the statement to be proved. But even the $0$-ary intersection (i.e. $X$) is in the topology....


Your Answer

By clicking "Post Your Answer", you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.