For each finite subsetJofI, letφ(J)=supL⊂J(pL−1). Then for each subsetLofJwe have
∣pL−1−sL∣≤φ(J)σL.
This is clear if L is empty. We proceed by induction on card (L). Let L=K∪{λ}, where λ∈/K; then pL=pK(1+aλ) and sL=sK+aλ, so that pL−1−sL=(pK−1−sK)+(pK−1)aλ; hence, by the inductive hypothesis and the definition of φ(J), we have
∣pL−1−sL∣≤φ(J)σK+φ(J)∣aλ∣=φ(J)σL,
which proves the lemma.
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.