Let r and s be two integers ⩾0. If Hr,s=∑b∈Bλbeb, where λb∈Q, is the decomposition of H with respect to the basis (eb)b∈B, then
vp(λb)⩾−(r+s−1)θfor all b∈B. The ring
AZ(p)(I) is identified with the sub-
Z(p)-module of
AQ(I) generated by the words
w∈Mo(I). As
LZ(p)(I) is a direct factor of
AZ(p)(I),
LZ(p)(I)=AZ(p)(I)∩LQ(I).
Let f be the integer such that f≤(r+s−1)θ<f+1. Relation (4) is equivalent to vp(λb)≥−f for all b∈B, that is Hr,s∈p−fLZ(p). But this is equivalent also by (5) to Hr,s∈p−fAZ(p)(I).
By formula (11) of § 6, no. 4, it suffices to show that, for every integer m≥1 and all integers r1,…,rm,s1,…,sm such that
r1+⋯+rm=r,s1+⋯+sm=s, ri+si≥1for 1≤i≤m, we have
vp(m!r!…rm!s1!…sm!)≤f. But by Lemma 2,
vp(ri!si!)≤(ri+si−1)θ and
vp(m!)≤vp(m!)≤(m−1)θ; the left hand side of (7) is therefore bounded above by
θ(m−1+i=1∑m(ri+si−1))=θ(r+s−1); as it is an integer, it is
≤f, which completes the proof.
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.