The multiplication operator mg from LKp1(X,μ) into LKp2(X,μ) is a closed operator with dense domain.
Let us first prove that the partial operator mg is closed. Let (fn,hn)n∈N be a sequence in LKp1(X,μ)×LKp2(X,μ) such that the sequence (fn,hn) of the classes of fn and hn belongs to the graph of mg and converges in LKp1(X,μ)×LKp2(X,μ) when n tends to infinity. Let (f,h) be a pair in LKp1(X,μ)×LKp2(X,μ) such that the pair (f,h) of their classes is the limit of (fn,hn).
There exists a sequence (fnk)k∈N extracted from the sequence (fn)n such that fnk(x) converges to f(x) for μ-almost every x (INT, IV, p. 131, § 3, no4, th. 3). This implies that hnk(x)=g(x)fnk(x) converges μ-almost everywhere to g(x)f(x). Moreover the sequence (hnk) converges to h in the space LKp2(X,μ). The functions h and gf are therefore equal μ-almost everywhere (loc. cit.). Thus f belong to the domain of mg and h=mg(f). This proves that mg is closed.
Let us prove that the domain of mg is dense in LKp1(X,μ). It is enough to verify that the classes of functions f∈K(X;K) belong to the closure of the domain of mg in LKp1(X,μ). Let f∈K(X;K) and let f be its class in LKp1(X,μ). For every integer n∈N, let φn denote the characteristic function of the set of elements x∈X such that ∣g(x)∣⩽n, and put fn=fφn. Then ∣gfn∣⩽n∣f∣, which belong to LKp2(X,μ), hence fn belong to the domain of mg. For every element x of X, the sequence (fn(x))n∈N converges to f(x) when n→+∞; moreover, one has ∣fn∣⩽∣f∣, which belong to LKp1(X,μ). By Lebesgue's theorem (INT, IV, p. 137, § 3, no7, th. 6), the sequence of classes of fn converges to f in LKp1(X,μ). Thus, the class of f belong to the closure of the domain of mg.
In the following proposition, it is supposed that p1=p2=2.
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.