7
$\begingroup$

Let $X$ be a spectrum. Is there a canonical construction/functor that would associate to this spectrum, an inverse spectrum $X'$, in the sense that $$\pi_*(X)\cong \pi_{-*}(X')?$$

To be more precise, such a spectrum $X'$ can always be constructed by attaching cells to produce the right homotopy groups, but is there a more conceptual way of creating it?

$\endgroup$

1 Answer 1

15
$\begingroup$

Not exactly what you asked for, but there is the Brown-Comenetz dual of $X$, $I_{\mathbb{Q}/\mathbb{Z}}X$, which has the property that its homotopy groups are Pontryagin dual to the homotopy groups of $X$ (in the negative degree):

$\pi_{-\ast}(I_{\mathbb{Q}/\mathbb{Z}}X) \simeq Hom(\pi_\ast(X),\mathbb{Q}/\mathbb{Z})$

So, if the homotopy groups of $X$ are all finite, then the Pontragin dual groups will be (non-canonically) isomorphic and this will have the required property.

See here for more details

$\endgroup$
3
  • 1
    $\begingroup$ There's a question that always bothered me. Maybe you know the answer. How canonical is the assignment $X\mapsto I_{\mathbb Q/\mathbb Z}X$? More precisely, is the assignment $X\mapsto I_{\mathbb Q/\mathbb Z}X$ a functor from the $\infty$-groupoid of spectra to itself? It's obviously a functor at the level of the homotopy category. But what about at the $\infty$-groupoid level? $\endgroup$ Dec 23, 2018 at 21:42
  • 1
    $\begingroup$ @AndréHenriques I've probably misunderstood your question, but does this answer it? If $I_\mathbf{Q/Z} := I_\mathbf{Q/Z}(S^0)$ denotes the Brown-Comenetz dualizing spectrum, then $I_\mathbf{Q/Z}(X) = F(X, I_\mathbf{Q/Z})$, and this is functorial in $X$. $\endgroup$
    – skd
    Dec 24, 2018 at 5:47
  • 2
    $\begingroup$ @skd. Yes, of course... that does answers my question. But I now realise that I haden't asked the right question. Here's the right question: what is $I_{\mathbb Q/\mathbb Z}$? More precisely, is there a way to define it up to contractible space of choices? $\endgroup$ Dec 24, 2018 at 16:18

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.