0

This is from an example in one of my topology lectures. I am not sure if I am misunderstanding the statement.

Suppose $X = \prod_{i \in I} X_i$, where for each $X_i$ we have some topological space. Suppose also that $\forall i \in I (A_i \subset X_i)$. We endow each $A_i$ with the subspace topology with respect to $X_i$. Finally, we define $A := \prod_{i \in I} A_i)$. Is it true that the product topology of $A$ coincides with the subspace topology $A \subset X$, regardless of whether $X$ is endowed with the box or product topology? In our lecture, I think we said it's true, but I can't prove it.

My attempt:

First I note for the product topology on $A$, it is generated by the basis $$\beta_A = \big\{\prod_{i=1}^n Z_i \times \prod_{i >n} A_i: \forall i \leq n(Z_i \text{ is open in } A_i)\big\}$$

Consider the case when $X$ is endowed with the product topology. I note that for any open $U$ in $X$, we have as an element of the subspace topology $A \subset X$:

$$\begin{align} U \cap A & = \big(\prod_{i=1}^n U_i \times \prod_{i>n} X_i\big) \bigcap \prod_{i \in I} A_i \\&= \big(\prod_{i=1}^n U_i \bigcap A_i\big) \times \big(\prod_{i>n} A_i \bigcap X_i\big) \\ &= \big(\prod_{i=1}^n U_i \bigcap A_i\big) \times \prod_{i>n} A_i\end{align}$$

Where in the first line, $U_i$ is an open set in $X_i$.

In this case, I can see that the 'set-structure' of $\beta_A$ and an element of the subspace topology are the same. That is, since each $A_i$ is a subspace of $X_i$, the open set $Z_i$ must be of the form $U_i \cap A_i$. At least at as a cursory argument, this makes some sense to me that the topologies will coincide).

I run into a problem when $X$ is endowed with the box topology. I note that in the box topology, basis elements can be of the form $\prod_{i \in I} U_i$ in which $U_i$ is open in every corresponding $X_i$, and there can be infinitely many $U_i \neq X_i$. So with similar reasoning to above, elements of the subspace topology $A \subset X$ look like:

$$\begin{align} U \cap A &= \prod_{i \in I} U_i \bigcap \prod_{i \in I} A_i \\ &= \prod_{i \in I} \big(U_i \bigcap A_i \big) \\ &= \prod_{i\in I} H_i \end{align}$$

The last equality occurring in the case that the family $U_i \subset A_i$ for all $U_i$ open in $X_i$. This shows that in general, the subspace topology of $A \subset X$ may contain elements that are products of infinitely many open sets not equal to $A_i$, which obviously cannot occur in the product topology.

Question: Is the bolded statement in the first paragraph true? Does the bolded statement in the previous paragraph disprove it?

[Note: I know my proof for $X$ having the product topology is at best incomplete, since I am comparing basis elements to topology elements, rather than comparing the actual topologies. I have been stuck on how to proceed unfortunately. Nonetheless, I think my reasoning might work as a disproof in the case that $X$ has the box topology, but I am not sure.]

masiewpao
  • 2,217

2 Answers2

2

This is definitely not true. If we simply take $A_i = X_i$, then this statement would imply that the box and product topologies are the same.

Doctor Who
  • 3,461
1

In this answer I prove some general results that imply that for the product topology it is true that the product of subspaces has the subspace topology wrt the product topology. This is essentially because all these topologies are initial topologies wrt certain functions. The box topology is not initial and is not as nicely behaved in many respects.

Henno Brandsma
  • 242,131