Bourbaki

Remark

Tag 0C0B, permanent. Cite it as [Bourbaki, Tag 0C0B].

General Topology, chapter VIII, § 2. Angular measure, trigonometric functions

The isomorphism of the groups C∗C^* and R×T\mathbf{R} \times T implies the existence of roots of every "binomial equation" zn=az^n = a in the field CC. Using this fact and the local compatibility of CC, we can obtain another proof of the theorem of d'Alembert-Gauss (Exercise 2).

There are only two distinct isomorphisms of the group T\mathbf{T} onto the group U\mathbf{U}; for if g,g′g, g' are two isomorphisms of T\mathbf{T} onto U\mathbf{U}, and if h′h' is the inverse of g′g', then h′∘gh' \circ g is an automorphism of T\mathbf{T}, and therefore (Chapter VII, , no. 4, Proposition 6) we have identically either g′(x)=g(x)g'(x) = g(x) or g′(x)=g(−x)g'(x) = g(-x). We may always assume that gg is such that ii is the image under gg of the class mod 1 of the point 14\frac{1}{4}; then, if φ\varphi denotes the canonical homomorphism of R\mathbf{R} onto T\mathbf{T}, every strict morphism of the additive group R\mathbf{R} onto the multiplicative group U\mathbf{U} is of the form x→g(φ(x/a))x \to g(\varphi(x/a)), where aa is a real number ≠0\neq 0 (Chapter VII, § 2, no. 3, Proposition 4); note that the interval ]−12∣a∣,12∣a∣[]-\frac{1}{2}|a|, \frac{1}{2}|a|[ is the largest symmetric open interval of R\mathbf{R} which is mapped one-to-one onto its image by this strict morphism, and that we have g(φ(14))=ig(\varphi(\frac{1}{4})) = i. We shall denote the homomorphism x→g(φ(x))x \to g(\varphi(x)) by x→e(x)x \to e(x); every strict morphism of R\mathbf{R} onto U\mathbf{U} is therefore of the form x→e(x/a)x \to e(x/a), where a≠0a \neq 0. The function e(x)e(x) is continuous on R\mathbf{R}, complex valued, and satisfies the identities

(1)

∣e(x)∣=1, |e(x)| = 1,
(2)
e(x+y)=e(x)e(y), e(x + y) = e(x)e(y),

together with the relations

(3)

e(0)=1,e(14)=i,e(12)=−1,e(34)=−i,e(1)=1. e(0) = 1,\quad e(\frac{1}{4}) = i,\quad e(\frac{1}{2}) = -1,\quad e(\frac{3}{4}) = -i,\quad e(1) = 1.

From (1) and (2) it follows that

(4)

e(−x)=1e(x)=e(x)‾, e(-x) = \frac{1}{e(x)} = \overline{e(x)},

and from (2) and (3) that

e(x+14)=ie(x),e(x+12)=−e(x), e(x + \frac{1}{4}) = ie(x),\quad e(x + \frac{1}{2}) = -e(x),
e(x+34)=−ie(x),e(x+1)=e(x). e(x + \frac{3}{4}) = -ie(x),\quad e(x + 1) = e(x).

Thus the function e(x)e(x) is periodic and has 1 as a principal period.

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

Cited by

Nothing.

Cites

Transcribed from the printed volume. Not translated.