MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

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

If $R$ is an algebra without a unit, then the standard unitisation $R^\sharp$ can have maximal one-sided ideals other than $R$. Thus, it is natural to ask about the following. Let $R$ be an algebra without a unit (over a field with char 0 if it does matter).

Is there a unital algebra $A$ such that $R$ is the unique maximal right ideal in $A$?

EDIT: Assume $R$ has a faithful representation $\pi$ on a vector space $V$ (over $K$), what if we consider $A=\pi(R)+KI$?

share|cite|improve this question

The answer to this question is no in general. Let $K$ be a field, and let $R=F\{s,t\ :\ st=s+t \}$, the non-unital algebra generated by the non-commuting variables $s,t$ subject to the single relation $st=s+t$.

Let $A$ be any unital $K$-algebra containing $R$. In $A$ we have $(1-s)(1-t)=1-s-t+st=1$. However, $(1-t)(1-s)=1-t-s+ts\neq 1$ since $ts\neq s+t$. Thus $(1-s)$ is only right invertible, and hence $A\setminus R$ does not consist only of units.

Edited to add: An easier example, which has the added benefit of being commutative is taking $R$ to be the non-unital ring generated by commuting variables $x,y$ subject only to the additional relation $x^2=x$. The element $1-x\in A\setminus R$ is not unit, since it is a non-trivial idempotent.

share|cite|improve this answer

The nice examples given by Pace show that the answer is negative in general. There is also a natural condition on $R$ which gives a positive solution. It involves the Jacobson radical of a non-unital ring. (I'm following the first few exercises in section 4 of Lam's A First Course in Noncommutative Rings.)

The binary operation $a \circ b = a + b -ab$ on $R$ is associative and has unit element $0$. An element of $R$ is called left (right) quasi-regular if it has a left (right) inverse in the monoid $(R, \circ, 0)$ and is called quasi-regular if it has both a left and a right inverse, which are necessarily equal.

The Jacobson radical $J(R)$ is the sum of all left ideals of $R$ consisting of left quasi-regular elements. One can show that this is an ideal of $R$, and all elements in it are quasi-regular. So $J(R)$ is the largest left (right) ideal consisting of left (right) quasi-regular elements. One can show that this coincides with the usual Jacobson radical in case $R$ has an identity.

Claim: For a (not necessarily unital) ring $R$, the following are equivalent:

  1. $R$ is the unique maximal right (or left) ideal of its standard unitization $R^\sharp$;
  2. $R = J(R)$.

Proof: Assume 1 holds, so that $R^\sharp$ is local and $R = J(R^\sharp)$. Then for any $x \in R$, the element $1 - x = 1 \oplus (-x) \in K \oplus R = R^\sharp$ lies in $R^\sharp \setminus R$ and therefore is invertible. Its inverse is of the form $\alpha \oplus y \in R^\sharp$. Writing out the equation for $(1-x)(\alpha \oplus y) = 1 \oplus 0$ yields that $\alpha = 1$ and $x + y - xy = 0$, and similarly we obtain $y + x -yx = 0$. So every element of $R$ is quasi-regular and 2 holds.

Conversely, assume 2 holds. We know that $R$ is a maximal right (and maximal left) ideal of $R^\sharp$ simply because of the ring isomorphism $R/R^\sharp \cong K$; thus $J(R^\sharp) \subseteq R$. If we can show that $R$ is contained in $J(R^\sharp)$ then we will be done, for then $R \subseteq J(R^\sharp) \subseteq R$ yields that $R^\sharp/J(R^\sharp) = R^\sharp/R \cong K$ making $R$ local with unique maximal right ideal $R$.

So let $x \in R$. To show $0 \oplus x \in J(R^\sharp)$, let $\alpha \oplus r \in R$; we need to find a left inverse to $$1 \oplus 0 - (\alpha \oplus r)(0 \oplus x) = 1 \oplus-(\alpha x + rx) = 1 \oplus (-x')$$ where we set $x' = \alpha x + rx \in R$. To this end, fix a left quasi-inverse $y$ for $x$ in $R$. Then from $y+x'-yx' = 0$ we obtain $$(1 \oplus -y)(1 \oplus -x') = 1 \oplus 0$$ as desired. QED

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.