Bourbaki

Example 3

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

General Topology, chapter III, § 1. Topologies on groups

If G is a topological group, its topology is compatible with the structure of the group G0G^0 which is the opposite of G; G0G^0, with this topology, is said to be the topological group opposite to G.

Axioms (GTI)(\mathrm{GT}_\mathrm{I}) and (GTII)(\mathrm{GT}_\mathrm{II}) are equivalent to the following:

(GT'). The mapping (x,y)→xy−1(x, y) \to xy^{-1} of G×GG \times G into GG is continuous.

Clearly (GTI)(\mathrm{GT}_\mathrm{I}) and (GTII)(\mathrm{GT}_\mathrm{II}) together imply (GT'). Conversely, (GT') implies (GTII)(\mathrm{GT}_\mathrm{II}), for x→ex−1=x−1x \to ex^{-1} = x^{-1} is then continuous; and (GT') and (GTII)(\mathrm{GT}_\mathrm{II}) together imply (GTI)(\mathrm{GT}_\mathrm{I}), for (x,y)→x(y−1)−1=xy(x, y) \to x(y^{-1})^{-1} = xy is then continuous.

If aa is any element of GG, the left translation x→axx \to ax (resp. the right translation x→xax \to xa) is continuous, by (GTI)(\mathrm{GT}_\mathrm{I}), and is therefore a homeomorphism of GG onto GG. The mappings x→axbx \to axb, as aa and bb run through GG, thus form a group of homeomorphisms of GG; the mappings x→axa−1x \to axa^{-1} (resp. x→ax,x→xax \to ax, x \to xa), where aa runs through GG, form a subgroup of this group of homeomorphisms. Again, since the symmetry x→x−1x \to x^{-1} is an involutory permutation of GG, axiom (GTII)(\mathrm{GT}_\mathrm{II}) shows that this mapping is a homeomorphism of GG onto GG.

If AA is an open (resp. closed) subset of GG, and if xx is any point of GG, then the sets x.A,A.xx.A, A.x and A−1(∗)A^{-1} (*) are open (resp. closed), for they are the transforms of AA under one of the preceding homeomorphisms. If AA is open and BB is any subset of GG, then ABAB and BABA are open, because they are unions of open sets [axiom (OI)(\mathrm{O}_\mathrm{I})]. If VV is a neighbourhood of ee in GG, then VAVA and AVAV are neighbourhoods of AA; for if WW is an open neighbourhood of ee contained in VV, then WAWA and AWAW are open and contain AA.

On the other hand, ABAB need not be closed when AA is closed, even if BB is closed too (cf. § 4, no. 1, Corollary 1 to Proposition 1).

\* For example, in the additive group of the real line R\mathbf{R}, the subgroup Z\mathbf{Z} of the rational integers is closed, and so is the subgroup θZ\theta \mathbf{Z} consisting of all integer multiples nθn\theta of an irrational number θ\theta; but the subgroup Z+θZ\mathbf{Z} + \theta \mathbf{Z} of R\mathbf{R}, which is the set of all real numbers m+nθm + n\theta (where mm and nn take all integer values) is not closed in R\mathbf{R}, as we shall see in Chapter V, § 1.

Again, let AA be the subset of the additive group of R×R\mathbf{R} \times \mathbf{R} which consists of all (x,y)(x, y) pairs such that x⩾0x \geqslant 0 and 0⩽y⩽1−1x+10 \leqslant y \leqslant 1 - \frac{1}{x+1}; and let BB be the set of all pairs (x,0)(x, 0), as xx runs through R\mathbf{R}. AA and BB are closed, but A+BA + B is the set of all pairs (x,y)(x, y) such that 0⩽y<10 \leqslant y < 1, and is not closed in R×R\mathbf{R} \times \mathbf{R}. \*

Let XX be a topological space and let ff and gg be two mappings of XX into a topological group GG. If ff and gg are continuous at a point x0x_0

(*) We recall that if AA and BB are two subsets of a group GG, then A.BA.B or ABAB denotes the set of all products xyxy where x∈Ax \in A and y∈By \in B; A−1A^{-1} denotes the set of all elements x−1x^{-1} where x∈Ax \in A. If BB consists of a single element xx, we write x.Ax.A or xAxA (resp. A.xA.x or AxAx) in place of {x}.A\{ x \}.A (resp. A.{x}A.\{ x \}).

of X, then so are f−1f^{-1} and fgfg(*), by Theorem 2 of Chapter I, § 2, no. 1. In particular, the continuous mappings of X into G form a subgroup of the group GXG^X of all mappings of X into G.

Again, let ff and gg be two mappings of a set X, filtered by a filter F\mathfrak{F}, into a Hausdorff topological group G. If lim⁡Ff\lim_{\mathfrak{F}} f and lim⁡Fg\lim_{\mathfrak{F}} g exist, then so do lim⁡Ff−1\lim_{\mathfrak{F}} f^{-1} and lim⁡Ffg\lim_{\mathfrak{F}} fg, and we have (Chapter I, § 7, no. 4, Proposition 9, Corollary 1)

(1)

lim⁡Ff−1=(lim⁡Ff)−1, \lim_{\mathfrak{F}} f^{-1} = (\lim_{\mathfrak{F}} f)^{-1},
(2)
lim⁡Ffg=(lim⁡Ff)(lim⁡Fg). \lim_{\mathfrak{F}} fg = (\lim_{\mathfrak{F}} f) (\lim_{\mathfrak{F}} g).

When G is a commutative group, written additively, the axiom (GT') indicates that (x,y)→x−y(x, y) \to x - y is a continuous mapping. If ff and gg are mappings of a topological space X into G, continuous at a point x0x_0, then f−gf - g is continuous at this point. The formulas (1) and (2) can be transcribed similarly.

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.