MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Given $A\in Mat_{n\times n}(R)$ where $R$ is a non-commutative associative ring are there exist any (non-zero) matrices $X, Y\in Mat_{n\times n}(R)$ such that $XAY=diag(a_1, \ldots , a_n)$ for some $a_i$? The interesting answer for me is if $A=(x_{i,j})$ and $R=\mathbb Z [x_{i,j}]$ (free associative non-commutative algebra on $x_{i,j}$ over $\mathbb Z$).

For example if $R$ is commutative then we put $X=Id$, $Y=(det(A_{i,j}))$ and get $XAY=det A\cdot Id$. What about non-commutative polynomials?

Upd1: I want to have a non-commutative polynomial equality. Also I want $X$ and $Y$ to be in general invertible.

Upd2: Ok, I've understood that Update1 wasn't correct. I'm interested in having such matrices over $R=Mat_{m\times m}(A)$ where $A$ is a commutative ring with $1$.

share|cite|improve this question
If the ring has a unity, then it is easy: just choose $a_{ij} \neq 0$ then $$ e_{1i}Ae_{j1}=a_{ij}e_{11} $$ where $e_{kl}$ is the matrix with $1$ in the $(k,l)$ entry and zero elsewhere. If the ring does have $1$, the statement is not generally true, since you can define $xy=0$ for all $x,y \in R$. – Keivan Karai Nov 30 '10 at 17:11
You want $X$ and $Y$ invertible or at least non-zero divisors. Otherwise the answer is no for obvious reasons. – Andreas Thom Nov 30 '10 at 19:12
@Andreas: I want to have a generalization of the equality: $A A^{V}=det A\cdot Id$. – zroslav Nov 30 '10 at 19:31

Take $X=Y=E_{1,1}$ (the matrix unit). Then $XAY=x_{1,1}E_{1,1}$, a diagonal matrix. If $R$ does not have 1, take $X=Y=aE_{1,1}$ for any $a\ne 0\in R$. Then $XAY=ax_{1,1}aE_{1,1}$ (it may be a zero matrix, but zero matrix is diagonal).

Update. Since you now want to find invertible $X,Y$, I would recommend starting with $2\times 2$-matrices and reading the book by Cohn, "Free rings and their relations", especially Chapter 2, Section 2.6.

share|cite|improve this answer
up vote -2 down vote accepted

I've understood that if $R=Mat_n(K)$ then for every $A\in Mat_m(R)$ exist $B\in Mat_m(R)$, s.t. $AB=\lambda Id$. $A=(a_{ij,kl})$ is a $mn\times mn$-matrix over $K$ and $B=A^{V}$ is a $m\times m$-matrix over $R$.

share|cite|improve this answer

Your Answer


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.