Let g1 and g2 be μ-measurable functions on X.
a) The partial operator mg1+mg2 is closable and its closure is the multiplication operator mg1+g2;
b) One has mg1∘mg2⊂mg1g2;
c) Suppose that g1 is bounded. Then one has mg2∘mg1=mg1g2. Moreover, the domain of mg2 is contained in the domain of mg1g2, and mg1∘mg2 is the restriction of mg1g2 to dom(mg2).
It is elementary that mg1+g2 is an extension of mg1+mg2; the latter operator is therefore closable, and mg1+mg2⊂mg1+g2.
Let f∈L2(X,μ) be such that the function h=(g1+g2)f belongs to L2(X,μ). For every integer n⩾1, let Xn denote the set of x∈X such that ∣g1(x)∣+∣g2(x)∣⩽n, and let φn be the characteristic function of Xn. We have φnf∈ dom(mg1+mg2). Since (φnf)(x) tends to f(x) for every x∈X, and since for every n∈N one has ∣φnf∣⩽∣f∣ and ∣(g1+g2)φnf∣⩽ ∣(g1+g2)f∣=∣h∣, with h∈L2(X,μ) by hypothesis, Lebesgue's theorem (INT, IV, p. 137, § 3, No.o7, Theorem 6) implies that the sequence of pairs of classes in L2(X,μ)×L2(X,μ) of (φnf,(g1+g2)φnf), which belong to the graph of mg1+mg2, tends to the pair of classes of (f,h) in L2(X,μ). The closure of mg1+mg2 is therefore indeed equal to mg1+g2.
It is elementary that mg1∘mg2⊂mg1g2. Suppose that g1 is bounded, so that mg1 is a continuous linear mapping on L2(X,μ). Then the domain of mg2∘mg1 is the set of classes of functions f∈L2(X,μ) such that g2(g1f)∈L2(X,μ), that is to say the domain of mg1g2. Hence mg2∘mg1=mg1g2.
It is likewise elementary that dom(mg1∘mg2) = dom(mg2) is contained in dom(mg1g2), and that the restriction of mg1g2 to this space is equal to mg1∘mg2.
One must beware of believing that mg1∘mg2=mg1g2 in general (exercise 10 of IV, p. 347). Nevertheless, one has the following partial result:
Bourbaki sets the proof directly after the statement, and the corpus does not separate them, so what is above is the statement and its proof together.
Nothing.
Nothing.