12
$\begingroup$

We can speak of rings in monoidal categories, including also the non-Cartesian case. What about fields?


Question 1: Definitions

What are some possible notions of a (skew or commutative) field in a symmetric monoidal category $\mathcal{C}$?

So far, I've found the following:

  • Approach #1: Taking the point of view of fields as rings where nonzero elements have inverses, one could start by considering groups of units, as done in the nLab page on topological fields.
  • Approach #2: Viewing fields representation-theoretically, we could define a field in $\mathcal{C}$ to be a ring object $k$ such that every $k$-module in $\mathcal{C}$ is (isomorphic to a) free one.
  • Approach #3: The notion of an ideal makes sense in any "nice" monoidal category; see Section 4.2 of Martin brandenburg's PhD thesis. One could define a field in $\mathcal{C}$ as a ring object $k$ in $\mathcal{C}$ having only $k$ and $(0)$ as ideals.

How do these approaches compare to each other? What are other possible definitions?


Question 2: Examples

Finally, what are some examples of fields in monoidal categories? In particular:

  • When $\mathcal{C}=\mathsf{Top}$, the first approach above recovers topological fields (i.e. topological rings $k$ which are fields but also whose inverse map $a\mapsto a^{-1}$ defines a continuous map $(-)^{-1}\colon k^{\times}\to k^{\times}$). Do the other approaches recover this continuity condition too?
  • For $\mathcal{C}=\mathsf{Sch}$, we have a well-studied notion of a ring scheme, of which a very important example is the ring scheme $\mathbb{W}$ of Witt vectors (for an introduction, see Eric's translation of Grothendieck's Groupes de Barsotti–Tate et Cristaux de Dieudonné). What are some examples of field schemes?
  • For $\mathcal{C}=\mathsf{CCoAlg}_{R}$, rings in $\mathcal{C}$ give the notion of a Hopf ring. What are examples of Hopf fields?
$\endgroup$
4
  • 6
    $\begingroup$ I don't believe that there will be a singular "the" correct notion. $\endgroup$ Commented Jul 31, 2021 at 22:37
  • 2
    $\begingroup$ I like division algebra meaning A-mod is semisimple and A is simple as an A-mod, so that you get an Artin-Wedderbern theorem for semisimple algebra objects. You could say it's a field if it's endowed with a lift to the center making it commutative, but I'm not sure how natural or useful that is. Part of me wants to say it's a mistake to try to talk about commutativity unless you're in a braided category. $\endgroup$ Commented Jul 31, 2021 at 23:01
  • 2
    $\begingroup$ There are already some non-equivalent definitions of internal fields in topoi, which is just a special case of the Cartesian case. So I agree with Theo's comment. $\endgroup$ Commented Aug 1, 2021 at 12:04
  • $\begingroup$ There is an essential difference between rings and fields that does not seem to have been recognised in the question or answers: there is an exclusive disjunction between zero and invertible elements. To express this you need a category that is extensive (has stable disjoint coproducts) as well as having a monoidal structure. $\endgroup$ Commented Sep 3, 2022 at 19:55

2 Answers 2

8
$\begingroup$

In the stable homotopy category, it is standard to define a field spectrum to be a ring spectrum $F$ such that every nonzero homogeneous element of the graded homotopy ring $\pi_*(F)$ is invertible. This ensures that every $F$-module spectrum is isomorphic to $\bigoplus_{i\in I}\Sigma^{d_i}F$ for some indexing set $I$ and some system of degrees $d_i\in\mathbb{Z}$.

For each prime $p$ and integer $n>0$ there is a Morava $K$-theory spectrum $K(p,n)$ with $\pi_*(K(p,n))=(\mathbb{Z}/p)[v_n,v_n^{-1}]$ with $|v_n|=2(p^n-1)$. Additionally, we have the Eilenberg-MacLane spectra $K(p,\infty)=H\mathbb{Z}/p$ and $K(0)=H\mathbb{Q}$. These are all field spectra. It is an important theorem that in a certain sense (not quite the most obvious one) this is a complete list of the prime field spectra. All of these examples are commutative except in the case $p=2$ where instead we have a rule like $ab-ba=v_nQ_n(a)Q_n(b)$ for a certain operation $Q_n$ with $Q_n^2=0$.

$\endgroup$
1
  • $\begingroup$ Thanks! (Incidentally, the case of field spectra was what made me think of this question in the first place!) $\endgroup$
    – Emily
    Commented Aug 6, 2021 at 22:33
4
$\begingroup$

Paul Balmer, Henning Krause, and Greg Stevenson have investigated the concept of 'field' in tensor-triangulated geometry (see Definition 1.1).

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .