7
$\begingroup$

Consider a fiber sequence of spaces

$$F \overset{i}{\to} E \to B$$

The cofiber $C(i)$ of the inclusion of the fiber comes with a canonical map $C(i) \to B$. Its possible to show (using some point set topology) that the fiber of that map is homotopy equivalent to $\Omega B \ast F$ hence we obtain a new fiber sequence

$$\Omega B \ast F \to C(i) \to B$$

To demonstrate the power of this statement consider the case of the path space fibration where $B= X , E = PX \cong \ast , F = \Omega X$. The new fiber sequence we obtain would be

$$\Sigma (\Omega X \wedge \Omega X) \cong \Omega X \ast \Omega X \to \Sigma \Omega X \to X$$

This implies immediately that whenever $X$ is $n$-connective for $n\ge 1$ then the canonical map $\Sigma \Omega X \to X$ is ($2n-1$)-connective (which is sort of the Eckman Hilton dual of Freudenthal suspension theorem).

Question: Is there an Eckman-Hilton dual to this statement?

To be precise, let $X \to Y \overset{\pi}{\to} Z$ be a cofiber sequence and let $F := fib(\pi)$ be the fiber $\pi$. Again there's a canonical (upto homotopy) map $f: X \to F$ and we may take its cofiber to obtain a new cofiber sequence

$$X \to F \to C(f)$$

Is there a "closed formula" for $C(f)$ (and for the maps involved) in terms of $X,Y$ and $Z$ (and the maps) similar to the one from above?

$\endgroup$
  • $\begingroup$ IS there a reason to expect it to be a loop space?! What is the case with the cofibre sequences $X\to *\to \Sigma X$. So, I thought something like $\Omega Map_*(Z,\Sigma X)$ might be a dual object to expect to appear as the cofibre of $X\to F$ ?!? $\endgroup$ – user51223 May 4 at 14:49
8
+100
$\begingroup$

This sort of question was studied a lot by Ganea (see here and here). The first piece of bad news is that duality is not perfect: there is no formula for that cofiber that depends only on $\Sigma X$ and $Z$. A counterexample was supplied by Barratt (as Remark 3.5 in the first linked paper above).

Even worse, the homotopy type of that cofiber does not have a formula just in terms of $X$ and $Z$! A counterexample is supplied in the second linked paper as 1.4.

The best we can do is that the suspension of the cofiber is given by $\Omega Z \ast X$.

This observation leads to a nice proof of the James splitting, incidentally- with no calculation.

$\endgroup$
  • 1
    $\begingroup$ Thanks! What a bummer though... :( $\endgroup$ – Saal Hardali May 7 at 11:58

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.