5
$\begingroup$

Note that the $2$-torus $T^2$ can be seen as a quotient space $\Bbb R^2/\Bbb Z^2$ of $\Bbb R^2$. Then any $2\times 2$ integer matrix $A=(\begin{smallmatrix} a & b\\ c & d \end{smallmatrix})$ gives a well-defined map $A:T^2\to T^2$. On the other hand, we have $H_1(T^2)=\Bbb Z^2$ and $H_2(T^2)=\Bbb Z$. What I want to show is, the map $A_*:H_2(T^2)\to H_2(T^2)$ induced by $A$, is given by $\Bbb Z\xrightarrow{\times \det(A)} \Bbb Z$, multiplication by $\det(A)$, and the map $A_*:H_1(T^2)\to H_1(T^2)$ is given by $\Bbb Z^2 \xrightarrow{A} \Bbb Z$.

Actually I want to use this result in Exercise 30 of section 2.2, in Hatcher's Algebraic Topology. (http://pi.math.cornell.edu/~hatcher/AT/AT.pdf) Parts (c) to (e) would become easy then.

The map on $H_1$ seems to be computed if we use the identification $\pi_1(T^2)=H_1(T^2)$ (which is after section 2.2, though), but I have no idea for the map on $H_2$. (Maybe a local degree argument?, but I'm not sure) Thanks in advance.

$\endgroup$

2 Answers 2

4
$\begingroup$

There's a natural map $\wedge^2(H^1(T^2, \mathbb{Z})) \to H^2(T^2, \mathbb{Z})$ given by the cup product. Show that it's an isomorphism. Then the desired result follows from the fact that $\det(A)$ is the scalar by which $A$ acts on the top exterior power, together with the universal coefficient theorem.

$\endgroup$
2
  • $\begingroup$ Thanks. Do you know a way to compute the map on $H_1$ without using the fact $\pi_1(T^2)=H_1(T^2)$? $\endgroup$
    – user302934
    Oct 27, 2020 at 5:21
  • 1
    $\begingroup$ @user302934 Compute it explicitly on generators for $H_1$ given by the segments (0,0) to (1,0) and (0,0) to (0,1). These map to the segments (0,0) to (a,c) and (0,0) to (b,d). Expand these in terms of the generators, and you're done. $\endgroup$
    – jgon
    Oct 27, 2020 at 5:50
1
$\begingroup$

If you use deRham cohomology, it's pretty clear that the area-form (which is a generator for $H^2$) gets multiplied by the determinant --- that's what the change of variables formula from calculus says.

Now use the fact that $H_2(T, \Bbb R) = hom( H^2(T, \Bbb R) ) $ to conclude that the corresponding map on $H_2$ must also be multiplication by $det$.

As an alternative, consider the 0-cocycle $\mu$ defined by "counting (signed) points". Because $T$ is connected, $\mu$ is clearly a generator. A drawing of the map $A$ shows that it takes any irrational point, like $(\pi/7, e/3)$ to $k = det(A)$ separate points, so the induced map on $H^0$ is multiplication by $k$. Poincare duality then makes the induced map on $H_n$ be multiplication by $k$ as well.

$\endgroup$

You must log in to answer this question.

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