Quotient maps are closed under direct sum

algebraic-topologygeneral-topology

This result seems wrong but I can't find where went wrong in this argument.

Let $q:X \rightarrow Y$ be a quotient map of top spaces. $g:A \rightarrow Y$ a continuous map. Then I claim the map
$(q,g):X \oplus A \rightarrow Y$
is a quotient map.


(Flawed?) Proof: We use the universal property of quotient map. $(q,g)$ is surjective as $q$ is. $(q,g)$ is quotient if and only if for any map $h:Y \rightarrow Z$, $h(q,g)$ is continuous if and only if $h$ is. One direction follows by continuity of $(q,g)$.

Suppose $h(q,g)$ is continuous. Then $hq:X \rightarrow Z$ is continuous. But by universal property with respect to $q:X \rightarrow Y$, $h$ is continuous. So $(q,g)$ is indeed a quotient map.

Best Answer

Your map $(q,g)$ is a dual to the diagonal map into products (if we have maps $f_i : X \to Y_i$ with a common domain, $\nabla_i f_i: X \to \prod_i Y_i$ defined by $\pi_i \circ \nabla_i f_i = f_i$ for all $i$ is also continuous), where here we have a common codomain $Y$ and we define $(q,g)$ by $(q,g) \circ j_X = q$ and $(q,g) \circ j_A = g$ and the universal property for sums implies that $(q,g)$ is continuous (it has the final topology wrt the embeddings $j_X: X \to X \oplus A$ and $j_A: A \to X \oplus A$).

That $(q,g)$ is quotient can also be seen by the definitions: indeed it is surjective as $q$ is (just as $\nabla_i f_i$ is injective whenever just one $f_i$ is) and if $(q,g)^{-1}[A] = q^{-1}[A] \oplus g^{-1}[A]$ is open in $X \oplus A$ this means that $q^{-1}[A]$ is open in $X$ and $g^{-1}[A]$ is open in $A$. The first part already implies that $A$ is open in $Y$ as $q$ is quotient, and we are done.

Related Question