Bourbaki

Exercise 24, § 1

Commutative Algebra, IX, p. 48

Let II be the set of negative integers. For every ring AA, every nilpotent two-sided ideal n\mathfrak{n} of AA, and every integer r⩾0r \geqslant 0, let CW(A,n,r)CW(A, \mathfrak{n}, r) be the subset of A1A^1 formed by the elements (ai)i∈I(a_i)_{i \in I} such that a−n∈na_{-n} \in \mathfrak{n} if n⩾rn \geqslant r. In other words, one has

CW(A,n,r)=A{0,−1,...,1−r}×n{−r,−r−1,...}. CW(A, \mathfrak{n}, r) = A^{\{0, -1, ..., 1-r\}} \times \mathfrak{n}^{\{-r, -r-1, ...\}}.
Endow CW(A,n,r)CW(A, \mathfrak{n}, r) with the product topology, each factor AA or n\mathfrak{n} being endowed with the discrete topology. Let CW(A)CW(A) denote the union of the CW(A,n,r)CW(A, \mathfrak{n}, r), and endow CW(A)CW(A) with the inductive limit topology. For this topology, CW(A)CW(A) is separated and CWu(A)CW^u(A) (cf. exercise 23) is dense in CW(A)CW(A). The elements of CW(A)CW(A) are called Witt covectors of A. a) Let tt be a positive integer, and let ω0,...,ωt\omega_0, ..., \omega_t be positive integers such that ω0\omega_0 be nonzero and that pt+1p^{t+1} divide ∑i=0tpiωi\sum_{i=0}^t p^i \omega_i. Then one has ∑i=0tωi⩾t(p−1)+p\sum_{i=0}^t \omega_i \geqslant t(p-1) + p. b) Let A\mathscr{A} be the ring Z[X,Y]\mathbf{Z}[X, Y] of polynomials in two families of indeterminates X=(Xi)i∈IX = (X_i)_{i \in I} and Y=(Yi)i∈IY = (Y_i)_{i \in I}. For every integer r⩾0r \geqslant 0, let nr\mathfrak{n}_r be the ideal of A\mathscr{A} generated by the X−nX_{-n} and Y−nY_{-n} for n⩾rn \geqslant r. Let rr and ss be two integers ⩾1\geqslant 1. Then one has
Sm(X−m,...,X0;Y−m,...,Y0)≡Sm+1(X−m−1,X−m,...,X0;Y−m−1,...,Y0) S_m(X_{-m}, ..., X_0; Y_{-m}, ..., Y_0) \equiv S_{m+1}(X_{-m-1}, X_{-m}, ..., X_0; Y_{-m-1}, ..., Y_0)
modulo nrs\mathfrak{n}_r^s for every integer m⩾r−1m \geqslant r-1 if s<ps < p and for every integer m⩾r−1+(s−p)/(p−1)m \geqslant r-1 + (s-p)/(p-1) if s⩾ps \geqslant p. (By weight arguments, now prove that the difference of the two members is a linear combination with integer coefficients of terms of the form X−m−1u0Y−m−1v0X−mu1Y−mv1...X0um+1Y0vm+1X_{-m-1}^{u_0} Y_{-m-1}^{v_0} X_{-m}^{u_1} Y_{-m}^{v_1} ... X_0^{u_{m+1}} Y_0^{v_{m+1}}) where, if one puts ωi=ui+vi\omega_i = u_i + v_i for 0⩽i⩽m+10 \leqslant i \leqslant m+1, one has ω0≠0\omega_0 \neq 0 and ∑i=0tpiωi=pm+1\sum_{i=0}^t p^i \omega_i = p^{m+1}. Then use a). c) Let AA be a ring, n\mathfrak{n} a nilpotent two-sided ideal of AA, rr a positive integer, and a,ba, b two elements of CW(A,n,r)CW(A, \mathfrak{n}, r). Then: (i) For every integer n⩾0n \geqslant 0, the sequence of elements
dm=Sm(a−n−m,...,a−n−1,a−n;b−n−m,...,b−n−1,b−n) d_m = S_m(a_{-n-m}, ..., a_{-n-1}, a_{-n}; b_{-n-m}, ..., b_{-n-1}, b_{-n})
of AA is stationary. (ii) For every integer n⩾0n \geqslant 0, let c−nc_{-n} be the limit of the preceding sequence. Then the element c=(ci)i∈Ic = (c_i)_{i \in I} belong to CW(A,n,r)CW(A, \mathfrak{n}, r). One will put a+b=ca + b = c. (iii) The preceding law of addition endows CW(A)CW(A) with a commutative group structure, compatible with its topology. For every nilpotent two-sided ideal n\mathfrak{n} of AA and every integer r⩾0r \geqslant 0, the subset CW(A,n,r)CW(A, \mathfrak{n}, r) of CW(A)CW(A) is a topological subgroup of it. The same holds for CWu(A)CW^u(A). d) Let AA and BB be two rings, φ\varphi a ring homomorphism from AA into BB. Let us denote by CW(φ)CW(\varphi) the mapping of CW(A)CW(A) into CW(B)CW(B) which to (ai)i∈I(a_i)_{i \in I} associates (φ(ai))i∈I(\varphi(a_i))_{i \in I}. Then CW(φ)CW(\varphi) is a group homomorphism and a continuous mapping.

Solution

No solution has been written yet.

Machine translation of the English, by gpt-5.4. Not checked by a person.