6
$\begingroup$

We know that $\mathbb{C}P^{\infty}$ is the classifying space of line bundles. Also we know that $\mathbb{C}P^{\infty}$ is an H space that is we have $$\mu: \mathbb{C}P^{\infty} \times \mathbb{C}P^{\infty} \to \mathbb{C}P^{\infty}$$ I have read that this map is the classifying map for tensor product of two line bundles but I don't understand what this means.

For example if $\xi_1$ and $\xi_2$ are two vector bundles over the space $X$ how does one use this map $\mu$ to get $\xi_1 \otimes \xi_2$? My guess is that suppose $\xi_i$ are classified by the map $f_i: X \to \mathbb{C}P^{\infty}$ . Then we have the map $g: X \to \mathbb{C}P^{\infty} \times \mathbb{C}P^{\infty}$ given by $g(x)=(f_1(x),f_2(x))$ then $\mu \circ g$ is the classifying map for the tensor product. Is my interpretation correct?

If so how does one prove this? Any hints are welcome. Thank you.

$\endgroup$
5
  • $\begingroup$ I guess how you would prove this depends on what your definition of $\mu$ is. You could define the $H$-space structure on $\mathbb{C}P^\infty$ to be the classifying map of this tensor product of line bundles, and then your statement would be tautological. $\endgroup$
    – JHF
    Commented Apr 15, 2017 at 21:55
  • $\begingroup$ @JHF The way we get the $\mu$ for me is by observing that it is $K(Z,2)$ and $K(Z,2)=\Omega K(Z,3)$. Also is my interpretation correct? $\endgroup$
    – happymath
    Commented Apr 16, 2017 at 3:12
  • $\begingroup$ Do you know the effect on the cohomology class if you take the tensor product of line bundles? If so, you just need to see that your loop multiplication induces ordinary addition on cohomology. $\endgroup$
    – user17892
    Commented Apr 16, 2017 at 16:32
  • $\begingroup$ @JustinYoung do you mean the fact that $c_1(L \otimes K)=c_1(L) + c_1(K)$? where $c_1$ denotes the chern class. Can you please elaborate on the second comment? How does knowing that pull back of the generator is sum of the two generators help? $\endgroup$
    – happymath
    Commented Apr 16, 2017 at 16:57
  • $\begingroup$ If two line bundles have the same $c_1$ they are equivalent bundles, since $c_1$ is the classifying map. $\endgroup$
    – user17892
    Commented Apr 16, 2017 at 18:25

1 Answer 1

4
$\begingroup$

This isn't so much as an answer as assembling the various comments into something more cohesive.

First observe that it suffices to very this claim for the (external) tensor product of the universal line bundle $\gamma:L(\gamma)\rightarrow \mathbb{C}P^\infty$ with itself. This follows since the classifying map of any tensor product $L_1\hat\otimes L_2$ of any two lines bundles $L_i\rightarrow X_i$ will factor through this universal tensor product.

Now observe that that the H-space structure on $\mathbb{C}P^\infty\simeq K(\mathbb{Z},2)$ is unique since $H^2(\mathbb{C}P^\infty\wedge\mathbb{C}P^\infty)=0$. Therefore the homotopy class of any map $\mathbb{C}P^\infty\times \mathbb{C}P^\infty\rightarrow \mathbb{C}P^\infty$ that extends the fold $\nabla:\mathbb{C}P^\infty\vee \mathbb{C}P^\infty\rightarrow \mathbb{C}P^\infty$ along the wedge inclusion $\mathbb{C}P^\infty\vee\mathbb{C}P^\infty\hookrightarrow \mathbb{C}P^\infty\times \mathbb{C}P^\infty$ is the multplication on $\mathbb{C}P^\infty$.

If $\mu:\mathbb{C}P^\infty\times\mathbb{C}P^\infty\rightarrow \mathbb{C}P^\infty$ classifies $\gamma\hat\otimes\gamma$ then the two bundle maps $i_1:\gamma\cong \gamma\hat\otimes \epsilon\hookrightarrow\gamma\hat\otimes\gamma$, $i_2:\gamma\cong \epsilon\hat\otimes \gamma\hookrightarrow\gamma\hat\otimes\gamma$ cover the inclusions of $\mathbb{C}P^\infty$ into each factor in the product and this shows that $\mu$ extends the folding map. It follows that up to homotopy $\mu$ is the H-space multiplication.

Looking at this in a slightly different direction you can use the Segre map $\mu:\mathbb{C}P^\infty\times \mathbb{C}P^\infty\rightarrow \mathbb{C}P^\infty$ (given here https://mathoverflow.net/questions/11117/h-space-structure-on-infinite-projective-spaces) and easily lift this to a bundle map $\gamma\hat\otimes\gamma\rightarrow \gamma$ which is an isomorphism of fibres $\mathbb{C}\otimes\mathbb{C}\cong\mathbb{C}$. It follows that this map $\mu$ is classifying map for $\gamma\hat\otimes\gamma$ as well as the unique h-space multiplication on $\mathbb{C}P^\infty$.

Once you have classified the external tensor product you can then internalise it. If $L_1\rightarrow X$, $L_2\rightarrow X$ are a pair of line bundles over a space $X$ then the internal tensor product is defined by the pullback $L_1\otimes L_2=\Delta^*(L_1\hat\otimes L_2)$ where $\Delta:X\rightarrow X\times X$ is the diagonal. If $f_i:X\rightarrow \mathbb{C}P^\infty$, $i=1,2$, classifies $L_i$ then $L_1\hat\otimes L_2\cong (f_1^*\gamma)\hat\otimes(f_2^*\gamma)\cong (f_1\times f_2)^*\gamma\hat\otimes\gamma\cong (f_1\times f_2)^*\mu^*\gamma$ is classified by $\mu\circ(f_1\times f_2)$. Thus $L_1\times L_2$ is classified by the composition

$X\xrightarrow{\Delta}X\times X\xrightarrow{f_1\times f_2}\mathbb{C}P^\infty\times\mathbb{C}P^\infty\xrightarrow{\mu}\mathbb{C}P^\infty$

as your intuition suggested.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .