Today, we developed a very general notion of differentiability in one variable. It exists on any topological ring.
I was asking the question “What is so special about certain the reals, the complexes or the p-adic numbers that you can do differentiation on it?” The most general definition of differentiability we could find surprisingly only required some basic algebra and topology. More specifically, it is perfectly possible to have differentiation on objects that have nothing at all to do with the real numbers.
The intuition is the following: Differentiating is approximating linearly. If a function is differentiable, then it is affinely linear up to corrections that vanish quadratically.
So we need to make sense of “linear” and “quadratic”. Here enter rings, which allow for addition and multiplication.
We also need to make sense of “approximating” or “vanishing”. Enter topology.
The rigorous definition now is:
Let be a topological ring. (This means it is a ring, it is a topological space and addition and multiplication are continuous.)
A map is called differentiable at with derivative if and only if for every sequences that tends to zero, there exists another sequence tending to zero, , such that the following equation holds:
Details:
- The is the affinely linear bit, the is the “quadratic” term.
- In a topological space, a sequence is called “tending to zero” iff for any neighbourhood around zero, only finitely many points of the sequence are outside the neighbourhood.
Now my questions:
If we take the reals, the complexes or the p-adics with the standard topology, do we recover the standard notion of differentiability?
What other interesting topological rings are there that we can do analysis on?