MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

This is a somewhat naive question to which I don't know the answer. There is a map of spectra $S \to S$, defined up to homotopy, given by multiplication by $-1$, and it satisfies the relation $(-1)^2 \simeq 1$. Can this be refined to a strict $\mathbb{Z}/2$-action on the sphere $S$ (and thus on all spectra)?

My feeling is that the answer is probably not, but I haven't been able to prove it. I do know that when one inverts $2$, there is a $\mathbb{Z}/2$-action: if I understand correctly, the obstructions to rigidifying a homotopy $\mathbb{Z}/2$-action to an actual $\mathbb{Z}/2$-action on a spectrum $X$ live in the $\mathbb{Z}/2$-cohomology groups of $B Aut(X)$ (where $Aut(X)$ is the monoid of homotopy self-equivalences of $X$), which should be zero when $2$ acts invertibly on $\pi_*X $.

share|cite|improve this question
up vote 7 down vote accepted

There's often no self-map of $S$ which induces multiplication by $(-1)$, because it's usually not cofibrant-fibrant. There's no homotopical obstruction to it existing, because the map $E \mapsto E \wedge S^1$ is an autoequivalence of the stable homotopy category, and $\mathbb{Z}/2$ acts on $S^1$. Or, if you prefer, we a map of based $\mathbb{Z}/2$-spaces $$ \mathbb{Z}/2_+ \to S^0 $$ and taking the homotopy fiber of the induced map of suspension spectra, we get a homotopy fiber sequence of spectra with $\mathbb{Z}/2$-action $$ S \to S \wedge \mathbb{Z}/2_+ \to S $$ with the first term having a negation action.

One can see visibly that in, e.g., the category of symmetric spectra, the sphere has no $-1$ self-map, but the equivalent free spectrum $F_1(S^1)$ does (for appropriate definitions of "S^1"). In the framework of Adams' "Stable homotopy and generalized homology", because maps are only defined "eventually", the analogue of $F_1(S^1)$ and $S$ are isomorphic objects, so you do have this map.

So if you want it on-the-nose, it depends on your framework. If you want it homotopically, it always exists (roughly because the map $BGL_1(S) \to B\mathbb{Z}/2$ splits).

share|cite|improve this answer
Fair enough. You're absolutely right that there is a $\mathbb{Z}/2$-action on the sphere $S^n$ (considered as a topological space), and I'm embarrassed that I convinced myself that this wasn't true. Thanks! – Akhil Mathew Sep 29 '12 at 21:09
@Akhil: I've often managed to convince myself of ridiculous things using fancy language. – Tyler Lawson Sep 29 '12 at 21:52
And, to be fair, this is very close to facts which are false. There is not a $\mathbb{Z}/2$-action by $(-1)$ on the sphere which is really "central" in the sense that you can coherently move it across $\wedge$ as you like. – Tyler Lawson Sep 29 '12 at 21:55
That's true. Somehow I imagined that having a $\mathbb{Z}/2$-action would make the stable category more "additive" than it is (before passage to homotopy), but I never thought about it on spaces. – Akhil Mathew Oct 2 '12 at 12:00

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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