Does there exist a non-trivial commutative monoid that does not have non-trival idempotents yet has a trivial Grothendieck group?

To partially explain my motivation, a cute little monoid! Consider a commutative monoid $M = \{0, k, a\},\ a + a = a + k = k + k = k$. It is very trivial, but at the same time interesting, because its Grothendieck group is trivial, yet $a$ is not an idempotent (which illustrates how idempotents can 'drag down' other elements to the kernel of the canonical map $x \mapsto x - 0$), and $a$ is also not a sum of an invertible and an idempotent element.

  • $\begingroup$ It was a misguided comment; sorry for the confusion. $\endgroup$ – Arturo Magidin Feb 2 '12 at 2:02

Yes. Let $\mathbb{Z}_+ = \{1,2,3,\ldots\}$, let $M = \{(0,0)\}\cup (\mathbb{Z}_+\times\mathbb{Z}_+)$, and consider the monoid structure on $M$ defined by $$ (x_1,y_1) + (x_2,y_2) \;=\; \begin{cases}(x_2,y_2) & \text{if }x_1<x_2 \\ (x_1,y_1+y_2) & \text{if }x_1 = x_2 \\(x_1,y_1) & \text{if }x_1 > x_2\end{cases} $$ Then the Grothendieck group for this monoid is trivial, since $(x,y) + (x+1,y) = (x+1,y)$ for all $(x,y)\in\mathbb{N}\times\mathbb{N}$. However, the monoid itself is non-trivial and commutative, and has no nontrivial idempotent elements.

Incidentally, this monoid can also be described in the following way: let $\mathbb{N}[x]$ be the monoid of all polynomials with natural number coefficients under addition, and let $\sim$ be the congruence relation on $\mathbb{N}[x]$ defined by $p(x)\sim q(x)$ if and only if $p(x)$ and $q(x)$ have the same leading term. Then the quotient $\mathbb{N}[x]/{\sim}$ is isomorphic to the monoid $M$ defined above.

Though I do not have a proof, I would guess that any example must be infinite. The example above isn't even finitely-generated.

  • 1
    $\begingroup$ Suppose $M$ is a finite monoid without nontrivial idempotents. If $x\in M$ is not $1$, then the set $\{x, x^2,x^3,\dots\}$ is finite and a little work shows it contains an idempotent. The idempotent must be $1$, by hypothesis, so $x$ is in fact invertible in $M$. We thus see that $M$ is a group. $\endgroup$ – Mariano Suárez-Álvarez Feb 3 '12 at 7:31

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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