I apologize in advance if this question is too technical. I haven't found a reference in the literature yet, and it seems difficult enough that perhaps it has not been answered.

Let $A$, $B$, and $C$ be based topological spaces, all compactly generated and weak Hausdorff. Let $\wedge$ and $F$ denote smash product and space of based functions, respectively, modified in the usual way so as to produce CGWH spaces. Consider the "assembly" map $$ A \wedge F(B,C) \rightarrow F(B, A \wedge C) $$ It is the adjoint of the smash product of the identity of $A$ and the evaluation map of $F(B,C)$.

Is this map always a closed inclusion?

(It is not hard to show that it is injective, has closed image, and that if $K$ is a closed subset of the domain, then $K \cup \{*\}$ is a closed subset of the codomain. If you take $A$, $B$, and $C$ to be $[0,\infty)$, you appear to get a counterexample.)

EDIT: I believed earlier that this counterexample could maybe be fixed by making everything compactly generated, but that doesn't help, you can still produce a non-convergent sequence in the source that converges in the target to the basepoint, and this still produces a closed set whose image is not compactly closed. So this is actually not true! Juan, thanks for pointing me to Lewis's thesis.


Isn't this addressed in Lewis' Appendix A (on compactly generated spaces)? His Proposition 8.5 gives a criteria (for A compact), to get a closed inclusion. But your comment on adding the disjoint basepoint is important ("The adjunction od the disjoing basepoint is essential to the [...] result"). See his CounterExample 8.6 following the proposition.

| cite | improve this answer | |
  • $\begingroup$ Hi Juan, you're right, he does address it, and specifically shows that the answer to my question is no. I'm interested in the case where $A$ is the $n$-sphere, and it looks like I should expect this map to not be a closed inclusion. Too bad... I wanted to say that $\Sigma^\infty_+ LX$ is homeomorphic to the totalization $Tot(\Sigma^\infty_+ X^{\bullet + 1})$, but it looks like that's not true now. $\endgroup$ – Cary Apr 14 '15 at 13:19

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.