- The linear functional that takes a function to its Fourier transform atk:
f∈S(R)→f ̃(k)
- The linear functional that takes a function to its value atq:
f∈S(R)→f(q)
- The linear functional that takes a function to the value of its derivative
atq:
f∈S(R)→f′(q)
We would like to think of these as “generalized functions”, corresponding toTψ
given by the integral in equation 11.8, for someψwhich is a generalization of a
function.
From the formula 11.3 for the Fourier transform we have
f ̃(k) =T√ 1
2 πe−ikq
[f]
so the first of the above linear functionals corresponds to
ψ(q) =
1
√
2 π
e−ikq
which is a function, but “generalized” in the sense that it is not inS(R) (or
even inL^2 (R)). This is an eigenfunction for the operatorP, and we see that
such eigenfunctions, while not inS(R) orL^2 (R), do have a meaning as elements
ofS′(R).
The second linear functional described above can be written asTδwith the
corresponding generalized function the “δ-function”, denoted by the symbol
δ(q−q′), which is taken to have the property that
∫+∞
−∞
δ(q−q′)f(q′)dq′=f(q)
δ(q−q′) is manipulated in some ways like a function, although such a function
does not exist. It can however be made sense of as a limit of actual functions.
Consider the limit as→0 of functions
g=
1
√
2 π
e−
(q− 2 q′)^2
These satisfy ∫
+∞
−∞
g(q′)dq′= 1
for all >0 (using equation 11.6).