When I talked about my alternative proof of Helgason’s support theorem in class, I asserted that our function had compact support in if and only if its partial derivatives satisfied for any multi-index .
In other words, a function is of exponential type if its partial derivatives grow like those of an exponential function.
This is true, but the claims I stated on the board were not correct — I did not mention the fact that we needed to use extra information about , in particular the fact that is all we need to complete our proof.
It is possible to state much more general results, but to get the idea across let me state the following result in one dimension:
Claim Assume then if and only if is entire analytic and for all
for some constant .
To prove this, we note the following
Lemma Assume then if and only if is entire analytic and for all
for some constant . In this case we will say that is analytic of exponential type.
Pointer to Proof I may elaborate on these ideas in a later post, but for now I will point you to references that prove more general results.
In Rudin’s Real and Complex Analysis, chapter 19, Theorem 19.3 we see that if
- is entire
- (It is an function on the real line)
- for all
Then is the Fourier transform of a function in .
Similarly, in Complex Analysis, p. 122, Theorem 3.3, Stein and Shakarchi show that if is continuous with “moderate decrease” on , then can be extended to an entire function on the complex plane satisfying if and only if is supported in . (Their statement looks a bit different because they use a different definition for the Fourier transform, I restated the result for consistency). Thanks for pointing this out, Huy!
Both of these results directly imply the lemma.
Having let Stein and Rudin do the heavy lifting, we can prove our claim using the direct estimates I put on the board in class. Here is a recap:
Step 1: If has compact support, the derivatives of satisfy our bounds and is analytic.
Specifically, let . Then
So we have our derivative bound. The standard Paley-Wiener theorem tells us that is analytic of exponential type.
Now we need to go the other direction
Step 2: If is analytic and the derivatives of satisfy our bounds, then has compact support
This result is straightforward as well. We will prove that $latex \hat{f}$ is of exponential type as defined in our lemma, i.e. .
Proving that is of exponential type. Our lemma implies that must have compact support in .