2

I'm trying to find a ring which has an element that has many(more than two) distinct left inverses.

I first tried the (square) matrices but then realized that if a matrix is a left unit, then it has determinant greater than 0, thus it is also a right unit and has unique inverse.

And then next I tried to do it with functions with usual addition and composition, and I found this example(in the first answer, by user23211): Kaplansky's theorem of infinitely many right inverses in monoids?

But I realized that the space of functions does not form a ring since the distribution law fails.

To make distribution law hold, we can only have homomorphisms; while the elements in the above example are not.

I also have seen that we can construct a similar examples with infinite vector spaces, but they are not rings.

Is there any ring like this?

user159234
  • 997
  • 5
  • 11
  • See https://math.stackexchange.com/a/70781/589. – lhf Apr 22 '17 at 10:54
  • 1
    I also have seen that we can construct a similar examples with infinite vector spaces, but they are not rings. Well, I know for sure the most common example (the ring of linear transformations of an infinite dimensional vector space) appears numerous times throughout the site. The problem seems to have been that it is hard to find a reference linking this with the 'infinitely-many-left-inverses' observation. – rschwieb Apr 22 '17 at 12:36

1 Answers1

3

Consider a countably infinite dimensional vector space $V$ and let $R$ be the endomorphism ring of $V$.

Suppose the vector space has a basis $\{v_k:k\in\mathbb{N}\}$; define the endomorphism $f$ by $$ f(v_k)=v_{k+1} $$ and, for every $v\in V$, the endomorphism $g_v$ by $$ g_v(v_0)=v,\qquad g_v(v_{k+1})=v_k $$ Then, for every $k\in\mathbb{N}$, $$ g_v(f(v_k))=v_k $$ so $g_v$ is a left inverse of $f$ in $R$.


It is an interesting exercise proving that, if the element $a\in R$ ($R$ any ring) has two distinct left inverses, then it has infinitely many.

egreg
  • 238,574
  • yes, this was what I saw but I somehow misunderstood that the $V$ was the space while they were saying about the endomorphism ring! thank you for stating the example so clearly. and yes as I linked I happened to know about Kplansky's theorem. it was pretty interesting! – user159234 Apr 23 '17 at 10:47