11.6. FOURIER TRANSFORMS OF JUST ABOUT ANYTHING 287

Also note that if f ∈S, then p◦ f ∈S for any polynomial p with p(0) = 0 and that

S⊆ Lp(Rn)∩L∞(Rn)

for any p ≥ 1. To see this assertion about the p( f ), it suffices to consider the case of theproduct of two elements of the Schwartz class. If f ,g ∈S, then Dα ( f g) is a finite sum ofderivatives of f times derivatives of g. Therefore, ρN ( f g)< ∞ for all N. You may wonderabout examples of things in S. Clearly any function in C∞

c (Rn) is in S. However there areother functions in S. For example e−|x|

2is in S as you can verify for yourself and so is any

function from G . Note also that the density of Cc (Rn) in Lp (Rn) shows that S is dense inLp (Rn) for every p.

Recall the Fourier transform of a function in L1 (Rn) is given by

F f (t)≡ (2π)−n/2∫Rn

e−it·x f (x)dx.

Therefore, this gives the Fourier transform for f ∈ S. The nice property which S has incommon with G is that the Fourier transform and its inverse map S one to one onto S.This means I could have presented the whole of the above theory in terms of S rather thanin terms of G . However, it is more technical.

Theorem 11.6.22 If f ∈S, then F f and F−1 f are also in S.

Proof: To begin with, let α = e j = (0,0, · · · ,1,0, · · · ,0), the 1 in the jth slot.

F−1 f (t+he j)−F−1 f (t)h

= (2π)−n/2∫Rn

eit·x f (x)(eihx j −1

h)dx. (11.18)

Consider the integrand in 11.18.∣∣∣∣eit·x f (x)(eihx j −1

h)

∣∣∣∣ = | f (x)|

∣∣∣∣∣(ei(h/2)x j − e−i(h/2)x j

h)

∣∣∣∣∣= | f (x)|

∣∣∣∣ isin((h/2)x j)

(h/2)

∣∣∣∣≤ | f (x)| ∣∣x j∣∣

and this is a function in L1(Rn) because f ∈S. Therefore by the Dominated ConvergenceTheorem,

∂F−1 f (t)∂ t j

= (2π)−n/2∫Rn

eit·xix j f (x)dx = i(2π)−n/2∫Rn

eit·xxe j f (x)dx.

Now xe j f (x) ∈ S and so one can continue in this way and take derivatives indefinitely.Thus F−1 f ∈C∞(Rn) and from the above argument,

Dα F−1 f (t) =(2π)−n/2∫Rn

eit·x(ix)α f (x)dx.

To complete showing F−1 f ∈S,

tβ Dα F−1 f (t) = (2π)−n/2∫Rn

eit·xtβ (ix)a f (x)dx

= (2π)−n/2∫Rn

i|β |eit·xDβ ((ix)a f (x))dx,

11.6. FOURIER TRANSFORMS OF JUST ABOUT ANYTHING 287Also note that if f € G, then po f € G for any polynomial p with p(0) = 0 and that6G CL?(R")NL*(R’)for any p > 1. To see this assertion about the p(f), it suffices to consider the case of theproduct of two elements of the Schwartz class. If f,g € G, then D® (fg) is a finite sum ofderivatives of f times derivatives of g. Therefore, py (fg) < °% for all N. You may wonderabout examples of things in G. Clearly any function in C> (R”) is in G. However there areother functions in G. For example eI” is in G as you can verify for yourself and so is anyfunction from ¥. Note also that the density of C, (R”) in L? (R”) shows that G is dense inL? (R") for every p.Recall the Fourier transform of a function in L! (IR") is given byF f(t) = (2m)-"? I eo F(x)de.Therefore, this gives the Fourier transform for f € G. The nice property which G has incommon with ¥ is that the Fourier transform and its inverse map G one to one onto 6.This means I could have presented the whole of the above theory in terms of G rather thanin terms of Y. However, it is more technical.Theorem 11.6.22 iff € 6, then Ff and F~'f are also in ©.Proof: To begin with, let @ =e; = (0,0,--- ,1,0,--- ,0), the 1 in the ji" slot.Fo! f(t+hej)—F'f(t inj — |f(t+ *) F(t) = (anyon? [et pexy( —)dx. (11.18)Consider the integrand in 11.18.it-x eli —] _ eilh/2)xj __ —i(h/2)x;Fy) = |¢69)|( 5_ isin ((h/2)x;) |= |f(x)| thd) < | f (x)| |x)|and this is a function in L'(IR") because f € G. Therefore by the Dominated ConvergenceTheorem,-1orf zn = (og)-n? I ix f(x)dx = (20)? [ ex" f(x)dx.nNow x*/ f(x) € G and so one can continue in this way and take derivatives indefinitely.Thus F~' f € C*(R”) and from the above argument,D*F~' f(t) =(20)-"? | ol (ix) F(x)dx.To complete showing F~'f € G,ptr f(t) = (2m) | lt 4B (ix) f(x)dx= (2m)? [ile DB ((ix)" F(x),