As the question title asks for, how do others "visualize" Witt vectors? I just think of them as algebraic creatures. Bonus points for pictures.

  • 2
    $\begingroup$ Do you mean an element? the ring of Witt vectors? it's spectrum? And over a field, I guess... $\endgroup$ – Xarles Jul 15 '18 at 7:38
  • 2
    $\begingroup$ Here is a picture on Wikipedia: en.wikipedia.org/wiki/P-adic_number#/media/… $\endgroup$ – Jason Starr Jul 15 '18 at 9:12
  • 3
    $\begingroup$ Do you mean the $p$-typical Witt vectors of a perfect ring in characteristic $p$, or something more general? The $p$-typical Witt vectors of $\mathbb F_q$ are the unique unramified extension of $\mathbb Z_p$ with residue field $\mathbb F_q$, which I think is pretty explicit (but bigger examples become more annoying to compute). $\endgroup$ – R. van Dobben de Bruyn Jul 15 '18 at 10:10
  • $\begingroup$ I like to think of the ring of Witt vectors $W(R)$ over some commutative ring $R$ as a collection of countably many solenoids, a twisted version of the "line" $\operatorname{Spec}(R)$. Naturally, $W_n(R)$ will be a collection of $n$ such solenoids. The gluing is then given as a sort of fitting-together of worm gears. $\endgroup$ – Dat Minh Ha Aug 10 '20 at 22:40

You can view the spectrum of the ring Witt vectors, in the sense of scheme theory, pretty reasonably.

If $R$ is $p$-torsion free, then $\mathrm{Spec}(W_n(R))$ is $n+1$ copies (or $n$ if you use the traditional indexing) of $\mathrm{Spec}(R)$ glued together in a suitable way along their fibers over $p$. But there are two qualifications. First, there is some "Frobenius twisting" involved, which is impossible to visualize because the Frobenius morphism is impossible to visualize, as far as I know. Second, each component is not simply glued transversally to the previous ones, but there is some higher order gluing.

A simple example is $\mathrm{Spec}(W_n(\mathbf{Z}))$. It consists of $n+1$ copies of $\mathrm{Spec}(\mathbf{Z})$, numbered $0$ to $n$, where the $k$-th copy is glued to the $(k-1)$-st copy modulo $p^k$. So copy $1$ is glued to copy $0$ transversally. Copy $2$ is glued to copy $1$ tangentially, but only to order $1$, and so on.

There is a little picture of $W_1$ on page 5 of my paper The basic geometry of Witt vectors, I. The affine case.


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.