The Casorati–Weierstrass Converse: What Kind of Singularity Am I?
Suppose is analytic on the punctured disk and satisfies
for some constant and all in the punctured disk.
Question: What can you conclude about the nature of the singularity of at ? Is a removable singularity, a pole, or an essential singularity? If it is a pole, what is its order?
Answer: Growth Bound Forces Singularity Type
Key Idea / Intuition
The growth rate of near a singularity is the fingerprint of what kind of singularity it is. A pole of order behaves exactly like near . An essential singularity has no definite growth rate (it oscillates wildly). A removable singularity stays bounded. Here, tells us blows up no faster than — but is not an integer! Poles have integer orders. So the singularity cannot be a pole of any integer order. But it also cannot be essential (essential singularities are not bounded by any power of ). The key resolution: multiplying by makes the product bounded near , so the singularity of is removable — and working backward pins down exactly what must look like.
Formal Proof / Solution
Step 1: Rule out essential singularity.
Near an essential singularity, by Casorati–Weierstrass, takes values dense in ; in particular it cannot satisfy for small (a controlled growth bound). So is not an essential singularity.
Step 2: Consider .
By hypothesis,
So as . Since is analytic on and bounded near , by Riemann's removable singularity theorem, extends to an analytic function on the full disk with .
Step 3: Determine .
Since and is analytic, we can write for some integer and analytic with .
Therefore
- If : then extends to an analytic function at (removable singularity).
- If : then , a pole of order 1 (simple pole).
Step 4: Check consistency with the growth bound.
- A pole of order satisfies near .
- The bound is satisfied by poles of order , i.e., (since must be a non-negative integer).
So has either:
- A removable singularity at , or
- A pole of order 1 (simple pole) at .
Conclusion:
The bound is not tight enough to force a pole, but it rules out poles of order and rules out essential singularities entirely. The precise answer: the order of the singularity is at most 1.
Elegant takeaway: The non-integer exponent is the surprise — it forces the singularity to be "between" a simple pole and a removable singularity, and the argument via cleanly resolves this.
Source: Mathematical folklore / standard complex analysis curriculum