Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Let $k$ be a field and $R=k[x_1,\cdots,x_i]$. Suppose we have two unital $R$-algebras $A$ and $B$ which are isomorphic to $R^{\oplus j}$ as modules, $Z(A)\cong Z(B)\cong R$, and there exists an injective map $f:A\to B$ which maps $1_A$ to $1_B$ (in particular, it maps $Z(A)=R\cdot 1_A$ isomorphically onto $Z(B)=R\cdot 1_B$). Is it always true that such a map $f$ is an isomorphism? If not, what extra conditions are necessary to guarantee that it will be an isomorphism?

share|improve this question
    
After posting my answer it occurred to me that you didn't specify, whether you want $f$ to be an isomorphism of $R$-modules or $R$-algebras? –  Jyrki Lahtonen Apr 22 '12 at 9:59
    
$f$ ought to be an $R$-algebra isomorphism. –  KReiser Apr 22 '12 at 10:00

1 Answer 1

up vote 3 down vote accepted

No, such a mapping $f$ need not respect multiplication. For example, we could have $A=B={\mathcal M}_{2\times2}(R)$, when $Z(A)=Z(B)$ consists of the scalar matrices, and define $f$ by $$ f:\pmatrix{a&b\cr c&d\cr}\mapsto\pmatrix {a&-b\cr 3c&d\cr}.$$

This is an isomorphism of $R$-modules, its restriction to the scalar matrices is the identity mapping. But it does not respect multiplication. For example, the nilpotent matrix $$ A=\pmatrix{1&-1\cr1&-1\cr}$$ gets mapped to an invertible matrix.


Adding an example, where $f$ is a homomorphism of $R$-algebras, but is not onto.

Let us introduce an element $y$ with the property $y^2=x_1$. Let us introduce another element $u$ subject to relations $u^2=1$, $ur=ru$ for all $r\in R$ and $uy=-yu$. Let $B$ be the free $R$-module of rank four with a basis $\{1,y,u,uy\}$. We can faithfully represent $B$ in the ring of $2\times2$ matrices with entries in the ring $R[y]$ as follows: $$ u\mapsto \pmatrix{0&1\cr1&0\cr},\qquad y\mapsto \pmatrix{y&0\cr0&-y\cr}, $$ and extending this $R$-linearly, so $$a=r_1+r_2y+r_3u+r_4uy\mapsto \pmatrix{r_1+r_2y&r_3-r_4y\cr r_3+r_4y&r_1-r_2y\cr}. $$ We easily check that $B$ is closed under multiplication, $R\subseteq Z(B)$, and hence $B$ is an $R$-algebra.

Claim. $Z(B)=R$.

Proof. Assume that $a=r_1+r_2y+r_3u+r_4uy\in Z(B)$. Then $au=r_3-r_4y+r_1u-r_2uy$ and $ua=r_3+r_4y+r_1u+r_2uy$, so we must have $r_2=r_4=0$. But also $ay=r_1y+r_3uy$ and $ya=r_1y-r_3uy$, so we must also have $r_3=0$. Q.E.D.

The submodule $A$ with basis $\{1,y,x_1u,x_1uy\}$ is also isomorphic to $R^4$ as an $R$-module. Furthermore, it is closed under multiplication, so it is a subalgebra of $B$. The above argument shows that $Z(A)=R$. So as a counterexample of an $R$-algebra homomorphism that meets all the criteria without being an isomorphism of algebras I offer the inclusion mapping $f=i:A\rightarrow B$.

share|improve this answer
    
IOW, in the latter example $B$ is an $R$-order in a cyclic algebra, and $A$ is a smaller $R$-order. –  Jyrki Lahtonen Apr 22 '12 at 11:08

Your Answer

 
discard

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.