Bourbaki

Lemma 1

Tag 0CTW, permanent. Cite it as [Bourbaki, Tag 0CTW].

Commutative Algebra, chapter IX, § 1. Vecteurs de Witt

If x and y are two elements of A congruent modulo Jm\mathbf{J}_m, one has

xpn≡ypnmod  Jm+n. x^{p^n} \equiv y^{p^n} \mod \mathbf{J}_{m+n} .
By induction on n, one reduces to the case where n=1n = 1. Let P denote the polynomial ∑i=0p−1XiYp−1−i\sum_{i=0}^{p-1} X^i Y^{p-1-i} of Z[X,Y]\mathbf{Z}[X, Y]. In view of the assumption made on x and y, one has P(x,y)≡P(x,x)≡p.xp−1mod  JmP(x, y) \equiv P(x, x) \equiv p.x^{p-1} \mod \mathbf{J}_m. Now one has Jm+p.A⊂J1\mathbf{J}_m + p.\mathbf{A} \subset \mathbf{J}_1, whence P(x,y)∈J1P(x, y) \in \mathbf{J}_1. Finally, xp−yp=(x−y)P(x,y)x^p - y^p = (x - y) P(x, y) belongs to JmJ1⊂Jm+1\mathbf{J}_m \mathbf{J}_1 \subset \mathbf{J}_{m+1}.

Let us prove a) by induction on n. The case n=0n = 0 is immediate. Suppose n⩾1n \geqslant 1. Under the hypotheses of a), one has (4) aip≡bipmod  Jm+1a_i^p \equiv b_i^p \mod \mathbf{J}_{m+1} for 0⩽i⩽n−10 \leqslant i \leqslant n-1 according to Lemma 1 , (5) Φn−1(a0p,...,an−1p)≡Φn−1(b0p,...,bn−1p)mod  Jm+n\Phi_{n-1}(a_0^p, ..., a_{n-1}^p) \equiv \Phi_{n-1}(b_0^p, ..., b_{n-1}^p) \mod \mathbf{J}_{m+n} according to the induction hypothesis applied to the elements a0p,...,an−1p,b0p,...,bn−1pa_0^p, ..., a_{n-1}^p, b_0^p, ..., b_{n-1}^p of A, and (6) Φn(a0,...,an)−pn.an≡Φn(b0,...,bn)−pn.bnmod  Jm+n\Phi_n(a_0, ..., a_n) - p^n.a_n \equiv \Phi_n(b_0, ..., b_n) - p^n.b_n \mod \mathbf{J}_{m+n} according to formulas (2) and (5). Since an−bna_n - b_n belongs to Jm\mathbf{J}_m, the element pn.an−pn.bnp^n.a_n - p^n.b_n belongs to Jm+n\mathbf{J}_{m+n} and one deduces from (6) the congruence

Φn(a0,...,an)≡Φn(b0,...,bn)mod  Jm+n, \Phi_n(a_0, ..., a_n) \equiv \Phi_n(b_0, ..., b_n) \mod \mathbf{J}_{m+n} ,
whence a).

Let us prove b) by induction on n. The case n=0n = 0 is immediate. Suppose n⩾1n \geqslant 1. Under the assumptions of b), we have ai≡bimod  Jma_i \equiv b_i \mod \mathbf{J}_m for 0⩽i⩽n−10 \leqslant i \leqslant n-1 by the induction hypothesis, and we deduce from this, as previously, the congruences (4), (5) and (6). But by assumption Φn(a0,...,an)\Phi_n(a_0, ..., a_n) and Φn(b0,...,bn)\Phi_n(b_0, ..., b_n) are congruent mod. Jm+n\mathbf{J}_{m+n}, and hence we have pn.(an−bn)∈Jm+np^n.(a_n - b_n) \in \mathbf{J}_{m+n}. Since the relation p.x∈Jk+1p.x \in \mathbf{J}_{k+1} implies x∈Jkx \in \mathbf{J}_k for every x∈Ax \in \mathbf{A} and every k⩾1k \geqslant 1, we have an−bn∈Jma_n - b_n \in \mathbf{J}_m, which completes the proof.

Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.

Cited by

Nothing.

Cites

Nothing.

Translated with machine help, not yet reviewed by a person.