1
$\begingroup$

Let $R$ be a ring. An endomorphism $\alpha:R\to R$ is said to be right central reflexive if for all $a,b\in R,$ $aRb=(0)\implies bR\alpha (a)\subset Z(R)$, where $Z(R)$ denotes the centre of the ring.

The ring $R$ is called right $\alpha$-central reflexive if $\alpha$ is a right central endomorphism for $R$.

Let us consider a statement as follows:

For all $a,b\in R$, $aR\alpha (b)=(0)\implies bRa\subset Z(R)$.....(*)

It can be proved that if $\alpha:R\to R$ is onto, $R$ is right central $\alpha$-reflexive implies (*) is true. I want to show show that surjectivity of $\alpha$ is not superfluous. But I could not find an into endomorphism $\alpha$ such that $(*)$ does not hold but $R$ is a right central $\alpha$-reflexive ring.

Please help me in this regard.

$\endgroup$
4
  • 1
    $\begingroup$ I'm confused. You say that $\alpha$-reflexive implies $(*)$ when $\alpha$ is onto. If you want to show that it's not true in general, you need to exhibit a case that is $\alpha$-reflexive but fails $(*)$. Here you are rather asking about the converse (whether $(*)$ implies $\alpha$-reflexive), but didn't claim this to be true when $\alpha$ is onto. $\endgroup$
    – YCor
    Oct 16, 2017 at 7:24
  • $\begingroup$ Sorry, I was wrong. I have changed the question. $\endgroup$
    – Anupam
    Oct 17, 2017 at 2:05
  • $\begingroup$ Different people use different conventions for the meaning of "ring". For you, rings are unital and associative, but not necessarily commutative, right? $\endgroup$
    – YCor
    Oct 17, 2017 at 4:32
  • $\begingroup$ yes, the ring we have used is associative, unital and not necessarily commutative. $\endgroup$
    – Anupam
    Oct 20, 2017 at 2:29

1 Answer 1

2
$\begingroup$

Here's a simple example.

Fix a domain $k$. Consider the (associative unital) ring $R=k\langle x,y\rangle/\langle\!xy=0\rangle\!\rangle$. Let $t$ be its endomorphism $x\mapsto 0$, $y\mapsto y$.

0) A basis of $R$ is given by the $y^nx^m$, $n,m\ge 0$.

1) I claim that $R$ right central reflexive, and even satisfies $aRb=0$ implies $bRt(a)=0$.

Indeed, suppose $aRb=0$ with $a,b$ nonzero. Write $a=P(y)+Q(y)x$ and $b=\sum_{i\ge 0}R_i(y)x^i$. Since $ayb=0$, we have $0=(P+Qx)y(\sum_{i\ge 0}R_ix^i)=\sum_iP(y)yR_i(y)$. Projecting (with respect to the basis) to $k[y]x^i$, we deduce that $P(y)yR_i(y)=0$ for all $i$. Since $b\neq 0$, there exists $i$ such that $R_i\neq 0$ and we deduce $P(y)=0$, so $a=Q(y)x$. Thus $t(a)=0$, so $bRt(a)=0$.

2) $(*)$ fails with $(a,b)=(1,x)$: indeed $1Rt(x)=0$, but $xR1=xR$ contains $x$ and is not central as $0=xy\neq yx$.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.