We follow the treatment given in Reed and Simon.
Let be a self-adjoint operator on a separable Hilbert space, with domain . Then there exists a finite measure space , a unitary operator , and a real-valued function such that
if and only if , and
For ,
The last equation should be read as a diagonalization; it says that
for almost all .
We have yet to give a careful definition of what it means for an operator to be self-adjoint, but in outline, the proof runs roughly as follows. If is self-adjoint, then
and map bijectively onto .
The maps and are closed, and therefore by the closed graph theorem, their set-theoretic inverses are (closed) bounded maps , from to .
and commute, and one is the adjoint of the other. Hence each is (bounded and) normal.
By the spectral theorem for bounded normal operators, there is a measure space , a unitary isomorphism , and a measurable, bounded, complex-valued function so that for all .
We therefore have a commutative diagram
where is the image of the operator given by multiplying by . It follows that we have a commutative diagram
whence, if we put , we obtain .
We now prove these statements.
Recall that a partially defined linear operator is closed if it has a closed graph in .
If is densely defined, then for each there is at most one for which
and the set of for which such exists defines the domain of the partially defined transpose map defined by .
If is densely defined, then is closed.
The trick is to exhibit the graph of as the orthogonal complement of some space. This suffices because orthogonal complements are always closed…