1
$\begingroup$

Is it true that any map from an infinite product of Eilenberg-MacLane spaces in to an Eilenberg-MacLane space factor through, upto homotopy, a finite subproduct? Even if we take with field coefficients?

$\endgroup$
1
  • 2
    $\begingroup$ I don't think so. Let $X$ be a product of infinitely many circles, then $\pi_1(X)$ is the product of infinitely many copies of $\mathbb Z$. Know take the map $X \to K(\pi_1(X),1)$. This does not factor through any finite subset of the circles because if it did the map would factor through a finitely generated group. $\endgroup$
    – Will Sawin
    Mar 23, 2013 at 5:54

1 Answer 1

6
$\begingroup$

This is not true because it's not true in algebra. Any map $A \to B$ of abelian groups can be realized by a map $K(A,n) \to K(B,n)$ which is unique up to homotopy, and so we can just find a map $\Pi A_i \to B$ which does not factor through a finite subproduct.

If $\mathbb F$ is a field, we can construct a map $\Pi^\infty \mathbb F \to \mathbb F$ which does not factor through any finite subproduct using the Axiom of Choice. Let $e_i$ be the element in the product which is 1 in the i'th position and 0 elsewhere. These are linearly independent, and so we can extend them to a basis of $\Pi^\infty \mathbb{F}$. Let $f$ be the unique linear function that sends all the $e_i$ to 1 and sends the rest of the basis elements to 0.

Even more interesting is that we can take $\prod_{p \text{ prime}} \mathbb{Z}/p$. Any finite subproduct is a finite group, but this infinite product has nontorsion elements (e.g. $(1,1,1,1\ldots)$ is nontorsion) and so there are nontrivial maps from this infinite product to $\mathbb{Q}$.

$\endgroup$
3
  • 2
    $\begingroup$ This is certainly the "right" answer, but it is not strictly speaking correct for the question as stated. It is typically not true that an infinite product of $K(A_i,n)$s will have the homotopy type of a CW-complex, so maps from the product to $K(B,n)$ will not necessarily correspond to the group maps. In fact, for $n=0$, it is true that any map from a product of discrete spaces to another discrete space factors through a finite subproduct. $\endgroup$ Mar 23, 2013 at 16:20
  • $\begingroup$ Correction: my statement about discrete spaces is true only if the spaces are finite. $\endgroup$ Mar 23, 2013 at 17:02
  • $\begingroup$ Eric, you're correct; I was too fast, and interpreted the question as about the cohomology of the infinite product rather than the honest mapping space. $\endgroup$ Mar 24, 2013 at 3:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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