To make use of the Lie algebra action of $\mathsf{gl}_2(\mathbb{C})$ to establish a isomorphism in modular representation theory, I would like an answer to this question:

Let $K$ be a field of prime characteristic. When is there a subring $R$ of the complex numbers and a maximal ideal $M$ of $R$ such that $R/M \cong K$?

Clearly no such ring $R$ exists if $K$ has strictly more than $|\mathbb{C}|$ independent transcendental elements. Is this the only obstruction? Is there a reasonably explicit way to construct a suitable $R$ when $K$ is the algebraic closure of $\mathbb{F}_p$?

As a follow-up (which at first I thought I needed, but I now see I can get around by working with $\mathrm{GL}_2(\mathbb{C})$ rather than $\mathrm{SL}_2(\mathbb{C})$), note that if $R/M \cong K$ then the induced map $\mathrm{GL}_d(R) \rightarrow \mathrm{GL}_d(K)$, defined on a $d \times d$ matrix with entries in $R$ by applying $R \twoheadrightarrow R/M \cong K$ to each entry, is a surjective group homomorphism.

Is is true in general that the restriction of the group homomorphism $\mathrm{GL}_d(R) \rightarrow \mathrm{GL}_d(K)$ to $\mathrm{SL}_d(R)$ is surjective onto $\mathrm{SL}_d(K)$, or are there further obstructions?

  • 8
    $\begingroup$ If $X$ is a transcendence base for $\mathbb C$ over $\mathbb Q$, then the subring it generates is a free commutative ring over $X$. That subring has a surjection onto any ring of cardinality at most $|X|=|\mathbb C|$. $\endgroup$ – Keith Kearnes Feb 26 '20 at 15:15
  • 2
    $\begingroup$ Follow-up question: yes (exercise). Hint: elementary matrices. $\endgroup$ – YCor Feb 26 '20 at 16:26
  • 3
    $\begingroup$ Clearly this is very easy for some people. But it genuinely comes up in my research and I think the down-voters are being rather harsh. $\endgroup$ – Mark Wildon Feb 26 '20 at 16:29
  • 1
    $\begingroup$ @MarkWildon no, it's that it generates $SL_d(K)$ for $K$ a field, and the arrow is in the right direction. The map $SL_d(R)\to SL_d(K)$ is already surjective in restriction to the elementary subgroup $E_d(R)$. $\endgroup$ – YCor Feb 26 '20 at 18:40
  • 1
    $\begingroup$ An algebraic closure of $\mathbb{F}_p$ is given by $\bar{\mathbb{F}}_p = R/M$ where $R$ is the ring of algebraic integers and $M$ any maximal ideal containing $p$. $\endgroup$ – Todd Leason Feb 28 '20 at 5:04

The complex numbers have transcendence degree the continuum over $\mathbb Q$ so contain a copy of the field of rational functions in continuum many variables over $\mathbb Q$. This in turn contains the ring of polynomials in continuum many variables over $\mathbb Z$, which surjects onto any ring of cardinality at most the continuum.

  • $\begingroup$ Thank you. This answers my first question. Maybe it's too much to ask for a more explicit construction, but I would be interested to know when the subring can be taken inside the ring of algebraic integers of $\mathbb{C}$. $\endgroup$ – Mark Wildon Feb 26 '20 at 16:22
  • $\begingroup$ @MarkWildon I think the construction of a transcendence basis of $\mathbb{C}$ over $\mathbb{Q}$ (or one of $\mathbb{R}$ over $\mathbb{Q}$) is not easy (up to my untutored knowledge, only 3 explicit algebraically independent elements are known !). $\endgroup$ – Duchamp Gérard H. E. Feb 26 '20 at 17:43
  • $\begingroup$ The subring can be taken inside the algebraic integers iff $K$ is an algebraic extension of $\mathbb{F}_p$. For, let $R$ be the ring of algebraic integers in $\mathbb{C}$. Since any subring $S$ of $R$ is an algebraic extension of $\mathbb{Z}$, the quotient of $S$ by a maximal ideal $m$ is an algebraic extension of $\mathbb{F}_p$ where $(p)=\mathbb{Z}\cap m$. ... $\endgroup$ – Todd Leason Feb 28 '20 at 5:29
  • $\begingroup$ ... Conversely, let $K$ be an algebraic extension of $\mathbb{F}_p$. Let $M$ be a maimal ideal of $R$ containing $p$. Then $R/M$ is an algebraic closure of $\mathbb{F}_p$. Hence we can assume $K \le R/M$. Let $\pi: R \to R/M$ and let $S = \pi^{-1}(K)$. Then $L=S/(M \cap S)$. qed $\endgroup$ – Todd Leason Feb 28 '20 at 5:36

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.