Bourbaki

Theorem 2 (Approximation Theorem for Absolute Values)

Tag 04W0, permanent. Cite it as [Bourbaki, Tag 04W0].

Commutative Algebra, chapter VI, § 7. Approximation theorem

Let fif_i (1≤i≤n1 \leq i \leq n) be absolute values on the same field KK which are not improper and no two of which are equivalent. Let ai (1≤i≤n)a_i \ (1 \leq i \leq n) be elements of KK and ε\varepsilon a real number >0> 0. Then there exists x∈Kx \in K such that fi(x−ai)≤εf_i(x - a_i) \leq \varepsilon for all ii.

Let KiK_i denote the field K with the topology defined by fif_i. The result to be proved is equivalent to the following: in the product P=K1×⋯×KnP = K_1 \times \cdots \times K_n, the closure D‾\overline{D} of the diagonal D is equal to P. This is obvious for n=1n = 1. Suppose that this point has been established in the case of k absolute values for k<nk < n.

We show first that there exists, for 2≤h≤n2 \leq h \leq n, an element xhx_h of K such that f1(xh)<1,f2(xh)>1f_1(x_h) < 1, f_2(x_h) > 1 and fi(xh)≠1f_i(x_h) \neq 1 for 3≤i≤h3 \leq i \leq h. We argue by induction on hh. If h=2h = 2, this follows from the fact that f1f_1 and f2f_2 are not equivalent. We therefore suppose that the existence of xh−1x_{h-1} has been shown and prove that of xhx_h. If fh(xh−1)≠1f_h(x_{h-1}) \neq 1, we may take xh=xh−1x_h = x_{h-1}; if fh(xh−1)=1f_h(x_{h-1}) = 1, we choose z∈K∗z \in K^* such that fh(z)≠1f_h(z) \neq 1 and xh=z(xh−1)sx_h = z(x_{h-1})^s solves the problem for s sufficiently large. We have thus proved the existence of xhx_h.

As the integer qq tends to infinity, f1(xnq)f_1(x_n^q) tends to 0, f2(xnq)f_2(x_n^q) tends to +∞+\infty and fi(xnq)f_i(x_n^q) tends to 0 or +∞+\infty for i≥3i \geq 3. Writing yq=xnq(1+xnq)−1y_q = x_n^q(1 + x_n^q)^{-1},

1−yq=(1+xnq)−1; 1 - y_q = (1 + x_n^q)^{-1};
hence the sequence (yq)(y_q) tends to 0 in K1K_1, to 1 in K2K_2 and to 0 or 1 in KiK_i for i≥3i \geq 3. By changing the numbering of the KiK_i, it may therefore be assumed that there exists an integer rr (1≤r<n1 \leq r < n) such that D‾\overline{D} contains the point (e1,…,en)(e_1, \ldots, e_n) where ei=1e_i = 1 for 1≤i≤r1 \leq i \leq r and ei=0e_i = 0 for r+1≤i≤nr + 1 \leq i \leq n. Now, D‾\overline{D} is a vector sub-K-space of P. Hence D‾\overline{D} contains the diagonals D' and D'' of
P′=K1×⋯×Kr, P' = K_1 \times \cdots \times K_r,
and P′′=Kr+1×⋯×KnP'' = K_{r+1} \times \cdots \times K_n. By the induction hypothesis, P′=D‾′P' = \overline{D}' and P′′=D‾′′P'' = \overline{D}''. Hence D‾=P\overline{D} = P.

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

Cited by

Cites

Nothing.

Transcribed from the printed volume. Not translated.