From: Andrew Usher on
Frederick Williams wrote:

> We find doubly periodic functions (of one variable) in complex analysis,
> might we then find triply and quadruply periodic functions in
> "quaternionic analysis" (supposing there is such a thing)? What about
> n-tuply periodic functions (of one variable) for n = 5, 6, ...?

No. Over the quaternions, the only differentiable functions are
linear, and thus not periodic. There's a proof somewhere on the
internet. I suppose this to be true to all real algebrae other than R
and C.

Andrew Usher