2

Ok, so one of the formulations of the Riemann-Lebesgue Lemma says: $$ f\in L^1(\mathbb{R}) \implies \hat{f}(\omega)\to 0\;\mbox{ when } \;|\omega|\to\infty.$$

I get all the steps of the proof, except the one which says that if $f$ is an arbitrary integrable function, it may be approximated in the $L^1$ norm by a compactly supported smooth function $g\in L^1$. My question is: Why? How do you know that for all $\epsilon>0$, and for all arbitrary function $f\in L^1$ exists a compactly supported smooth function $g$ so that $||f-g||_{L^1}<\epsilon$? The "proof" can be found in many places, but this step can be found in Wikipedia https://en.wikipedia.org/wiki/Riemann%E2%80%93Lebesgue_lemma

Davide Giraudo
  • 172,925
Carlos Vázquez Monzón
  • 284
  • 1
    http://math.stackexchange.com/questions/242877/compact-support-functions-dense-in-l-1. This is actually valid for any locally compact Hausdorff topological space with a regular measure. –  Jul 05 '15 at 20:59
  • 3
    The question should be rephrased, the real question is "Why are compact supported smooth functions dense in $L^1(\mathbb{R})$?". – AD - Stop Putin - Jul 06 '15 at 11:17
  • 1
    Lebesgue Integration by Williamson, Chapter 4 , Section 4.3 Approximation to Integrable functions, gives the answer to all of your questions, including question about convolution. –  Jul 12 '15 at 09:45

2 Answers2

0

Let $f \in L^1(\mathbb{R})$. Given an $\epsilon > 0$ you can find a simple compactly supported function $\phi$ such that $\|f - \phi\|_{L^1} < \epsilon/2$. Now you just have to find a smooth and compactly supported function near $\phi$. One way to do this is by doing a convolution against a mollifier.

There are other ways to prove this. Another proof can be found on Folland's book Real Analysis: Modern Techniques and Their Applications.

Max Reinhold Jahnke
  • 553
  • Yeah ok I get it but the question now is: how do you know the convolution of $\phi$ and $f$ is a smooth compactly supported function? – Carlos Vázquez Monzón Jul 05 '15 at 22:35
  • To see that the convolution is a smooth function you just have to differentiate under the integral sign. Look at this Wikipedia article: https://en.wikipedia.org/wiki/Convolution#Differentiation

    About the support, you can show it dirrectly. Try to prove that the inclusion on the following link is true: https://en.wikipedia.org/wiki/Mollifier#Support_of_convolution

     – Max Reinhold Jahnke Jul 05 '15 at 23:21
  • @MaxReinholJahnke I looked on wikipedia but I did not knew which part said that convolution is a smooth function. – Emptymind May 27 '17 at 07:03
0

There are several ways to construct $L^1(\mathbb{R})$.

One way is to let $L^1(\mathbb{R})$ be the smallest Banach space containing $C_0^\infty(\mathbb{R})$ under the norm $\|\cdot\|=\|\cdot\|_{L^1}$.

For other definitions of $L^1$ this is also true (because it is the same space), however then this would need a proof based on the particular construction.

To clarify myself, what I mean is that:

If we define $L^1(\mathbb{R})$ as the closure of the space of compactly supported smooth functions. Then, by definition, we can always approximate the $L^1$-functions by a compactly supported smooth function.

If we define $L^1(\mathbb{R})$ in some other way then we can still approximate the $L^1$-functions by a compactly supported smooth function - but then it needs a proof depending on the particular definition.

AD - Stop Putin -
  • 10,970