# Why is the determinant a homomorphism? [duplicate]

Possible Duplicate:
How to show $\det(AB) =\det(A)\det(B)$

Consider the map $q:\operatorname{GL}(n,R)→R^*$ given by $q(A)=\det(A)$. I know that $$q(AB)=q(A)q(B)=\det(A)\det(B)$$ does not hold if determinants of $A$ and $B$ are not unity. I would like to know why the map is homomorphic?

-

## marked as duplicate by Zev ChonolesDec 28 '11 at 15:33

the multiplicativity of the determinant holds independently of any other conditions: en.wikipedia.org/wiki/… – t.b. Dec 28 '11 at 15:03
What do you mean by $\text{det}(AB)=\text{det}(A)*\text{det}(B)$ doesn't hold? – NikolajK Dec 28 '11 at 15:04
See the answers to this question. – Pierre-Yves Gaillard Dec 28 '11 at 15:29
@Pierre: Should we close as a duplicate, then? – Zev Chonoles Dec 28 '11 at 15:32
@ZevChonoles: I think so. – Pierre-Yves Gaillard Dec 28 '11 at 15:33

It is a homomorphism because you are incorrect that $\det(AB)\neq\det(A)\det(B)$ unless $\det(A)$ and $\det(B)$ are 1; in fact the statement that $\det(AB)=\det(A)\det(B)$ is always true for matrices over any commutative ring.
This is something I've always just thought, but I had a surprising amount of difficulty coming up with a reference for it, and I couldn't think of a way of proving the general statement myself, which is off-putting. At any rate, here is one proof, which shows that $\det(AB)=\det(A)\det(B)$ holds for matrices over $\mathbb{C}$, then argues that this implies it is true for matrices over any ring $R$ because the determinant is an integer polynomial in the entries of a matrix and we can embed the polynomial ring in $n^2$ variables over $\mathbb{Z}$ into $\mathbb{C}$.