Let $X$ be Hausdorff space and let $x \in X$ such that there is a closed neighborhood $U$ such that $\{x\}$ is a strong deformation retract of $U$. If $Y$ is another Hausdorff space, I would like to show the inclusion maps induce isomorphisms

$$\tilde{H}_p(X) \oplus \tilde{H}_p(Y) \to \tilde{H}_p(X \vee Y)$$

where $X \vee Y = X \times \{y\} \cup \{x\} \times Y$ for some $y \in Y$. Since $H$ could be any homology theory, I need to derive this directly from the Eilenberg - Steenrod axioms. I probably need to use the additivity axiom which states that for a space $X = \coprod X_i$, the inclusion maps induce an isomorphism

$$\bigoplus H_n(X_i) \to H_n(X)$$

However, I'm not sure how to relate $X \coprod Y$ with $X \vee Y$. I would to solve this problem without using Mayer - Vietoris. Any help would be appreciated.

  • $\begingroup$ Be careful, I think that -exactly- the point is not to use additivity. The additivity axiom (for a reduced theory) says exactly what you're trying to prove. The point here is that additivity for finite wedge is included in the Eilenberg-Steenrod axioms (not so for infinite wedges). $\endgroup$ – Bruno Stonek Jan 13 '16 at 10:01

Consider a pair $(X \lor Y, X)$. Let $Z = X - U$. By excision axiom, inclusion $(X \lor Y - Z, X - Z) \subset (X \lor Y, X)$ induces isomorphism in homology. As we have strong deformation retraction of $U$ on $x_0$, and since $(X \lor Y - Z, X - Z) = (U \lor Y, U)$, we get a deformation retraction of a pair $(U \lor Y, U) \to (Y, x_0)$, and the homotopy axiom says that this induces isomorphism in homology. The inverse is given by the map given by inclusion. Overall, we have that inclusion $(Y, x_0) \to (X \lor Y, X)$ induces an isomorphism. This map has an obvious left inverse $(X \lor Y, X) \to (Y, x_0)$ that contracts $X$ to a point. By easy categorical considerations, this map also must induce an isomorphism in homology.

Now consider the long exact sequence for pair $(X \lor Y, X)$:

$ \cdots \to \tilde{H}_n(X) \to \tilde{H}_n(X \lor Y) \to \tilde{H}_n(X \lor Y, X) \to \tilde{H}_{n-1}(X) \to \cdots$

The map $\tilde{H}_n(X) \to \tilde{H}_n(X \lor Y)$ is induced by inclusion $(X, x_0) \to (X \lor Y, x_0)$. There's an obvious left inverse to this inclusion, a map $(X \lor Y, x_0) \to (X, x_0)$ that contracts $Y$ in $X \lor Y$ to a point. As this map has a left inverse, the map induced in homology also has a left inverse, so it's a monomorphism, and we have an exact sequence $0 \to \tilde{H}_n(X) \to \tilde{H}_n(X \lor Y)$

Similarly, consider $\tilde{H}_n(X \lor Y) \to \tilde{H}_n(X \lor Y, X)$. As this map is also induced by inclusion $(X \lor Y, x_0) \to (X \lor Y, X)$, and from previous consideration we know that the contraction $(X \lor Y, X) \to (Y, x_0)$ induces an isomorphism, we consider a composition of these two. This map has a right inverse given by inclusion $(Y, x_0) \to (X \lor Y, x_0)$, which induces a left inverse for a map $\tilde{H}_n(X \lor Y) \to \tilde{H}_n(Y)$, which necessarily is an epimorphism. Now compose it with the inverse isomorphism $\tilde{H}_n(Y) \to \tilde{H}_n(X \lor Y, X)$ to obtain that $\tilde{H}_n(X \lor Y) \to \tilde{H}_n(X \lor Y, X)$ is also epimorphism, and thus we get an exact sequence $\tilde{H}_n(X \lor Y) \to \tilde{H}_n(X \lor Y, X) \to 0$. Remember that we composed it with the isomorphism inverse to the one $\tilde{H}_n(X \lor Y, X) \to \tilde{H}_n(Y)$ we used earlier, so here in this exact sequence, the map $\tilde{H}_n(X \lor Y) \to \tilde{H}_n(X \lor Y, X)$ is the same as in long exact sequence.

Finally, we connect these two exact sequences to get a short exact sequence:

$0 \to \tilde{H}_n(X) \to \tilde{H}_n(X \lor Y) \to \tilde{H}_n(X \lor Y, X) \to 0$

As the first map is induced by inclusion and it has a right inverse we mentioned earlier, we obtain that this s.e.s. split, so $\tilde{H}_n(X \lor Y) \simeq \tilde{H}_n(X) \oplus \tilde{H}_n(X \lor Y, X) \simeq \tilde{H}_n(X) \oplus \tilde{H}_n(Y)$.

Note that additivity axiom was not used.

| cite | improve this answer | |

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.