Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Suppose we have a sufficiently nice (example below) pathwise connected topological space $X$ with a double cover $\pi : \widetilde X \to X$. There is an associated free $\mathbb Z_2$-action on $\widetilde X$ interchanging the sheets of the cover; denote by $\iota : \widetilde X \to \widetilde X$ the map corresponding to the nonzero element of $\mathbb Z_2$. The question is: under which conditions is $\iota$ detectable by its action on the homology of $\widetilde X$, $\iota_* : H_*(\widetilde X) \to H_*(\widetilde X)$? In other words, when is $\iota_*$ not the identity? And when it is not, how can one compute the $\pm 1$-eigenspaces of $\iota_*$?

(The coefficients may be freely chosen to give the "best" answer possible.)


1) If $\pi$ is the trivial covering, $\widetilde X$ is the disjoint union of two copies of $X$ and therefore $H_*(\widetilde X) = H_*(X) \oplus H_*(X)$. The action of $\iota_*$ is by interchanging the summands and therefore it is detectable.

2) If $X = \widetilde X = S^1$ and $\pi$ is the unique nontrivial double covering, then $\iota$ is isotopic to the identity on $S^1$ and therefore acts trivially on $H_*(S^1)$.

3) If $X = \mathbb R P^2$, $\widetilde X = S^2$ and $\pi : \widetilde X \to X$ is the orientation covering, then $\iota$ acts as the antipodal map on $S^2$. The map $\iota_*$ acts trivially on $H_0(S^2)$ but on $H_2(S^2;\mathbb Z) = \mathbb Z$, for example, it acts as the multiplication by $-1$.

4) The real example I have in mind is as follows. Let $M$ be a smooth manifold, $L \subset M$ a smooth submanifold, and assume for simplicity both are closed and connected. Fix a point $x_0 \in L$, and let $X$ be the space of smooth maps $(D^2, \partial D^2, 1) \to (M,L,x_0)$, representing $0 \in \pi_2(M,L,x_0)$. Here $D^2 \subset \mathbb C$ is the closed unit disk. There is an isomorphism $\pi_1(X,x_0) \simeq \pi_3(M,L,x_0)$ ($x_0$ being the constant map, of course). Suppose $\widetilde X$ is the double cover corresponding to a real line bundle whose first Stiefel-Whitney class is given by a homomorphism $w : \pi_3(M,L,x_0) \to \mathbb Z_2$ (in view of the above isomorphism this is equivalent to specifying a class in $H^1(X;\mathbb Z_2)$). What can be said about this example?

share|improve this question
As a generalization of (3), homology detects the involution on the orientation double cover of any closed manifold. –  Eric Wofsey Sep 9 '13 at 0:57

2 Answers 2

Here's something you can say. Recall (Hatcher 3G.1) that if $G$ is a finite group and $\pi : X \to X/G$ a covering map associated to an action of $G$ on a space $X$, then over a field $F$ of characteristic $0$ or not dividing $|G|$ the transfer map $H^k(X/G, F) \to H^k(X, F)$ (given by sending a singular simplex in $X/G$ the sum of its $|G|$ distinct lifts in $X$) is injective and has image precisely the fixed subgroup $H^k(X, F)^G$.

In particular, letting $F = \mathbb{Q}$, if $X$ has finitely generated homology and $G$ acts trivially on the homology of $X$ then $X$ and $X/G$ have the same Euler characteristic, as in your example 2). Taking the contrapositive, if $X$ has finitely generated homology and $\chi(X) \neq \chi(X/G)$ then $G$ can't act trivially on the homology of $X$, as in your example 3). Unfortunately I have no idea if your spaces have finitely generated homology.

share|improve this answer

Like Qiaochu says, if you use a field of characteristic $\not= 2$ as coefficients then the cohomology of the base is simply isomorphic to $+1$ eigenspace of the cohomology the double cover. I don't know about the specific application you have in mind in 4), but generally you can also get some useful information about the cohomology with $\mathbb{Z}_2$ coefficients by applying the Gysin sequence for a 0-dimensional sphere bundle (unoriented, unless the double cover is trivial). This gives a long exact sequence

$$ H^0(X; \mathbb{Z}_2) \stackrel{\pi^*}{\to} H^0(\widetilde X; \mathbb{Z}_2) \stackrel{I}{\to} H^0(X; \mathbb{Z}_2) \stackrel{\cup\, w_1}{\to} H^1(X; \mathbb{Z}_2) \cdots ,$$

where $I$ is fibrewise integration, and the snake maps are cup product with the Stiefel-Whitney class of the line bundle corresponding to the cover. Note that $\pi^* \circ I$ equals $1 + \iota^*$ on $H^k(\widetilde X; \mathbb{Z}_2)$. This can be useful for instance for understanding whether the integral cohomology $H^k(\widetilde X; \mathbb{Z})$ is spanned by the $\pm 1$ eigenspaces of $\iota^*$ or whether their direct sum has some coquotient $\mathbb{Z}_2^r$.

share|improve this answer

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.