Bourbaki

§ 1. Hyperplanes, chambers and facets

In this section, E denotes a real affine space of finite dimension d and T the space of translations of E (cf. Algebra, Chap. II, § 9). If a and b are two points of E, [ab] (resp. ]ab[, resp. ]ab]) will denote the closed segment (resp. open segment, resp. segment open at a and closed at b) with extremities a, b. The space T is provided with its unique separated topological vector space topology, cf. Topological Vector Spaces, Chap. I, § 2, no. 3; it is isomorphic to Rd\mathbf{R}^d. The space E is provided with the unique topology such that, for all e∈Ee \in E, the map t↦e+tt \mapsto e + t from T to E is a homeomorphism.

We denote by H a locally finite set of hyperplanes of E (General Topology, Chap. I, § 1, no. 5).

1. Notations

Let H be a hyperplane of E. Recall that E – H has two connected components, called the open half-spaces bounded by H. Their closures are called the closed half-spaces bounded by H. Let x,y∈Ex, y \in E. Then x and y are said to be strictly on the same side of H if they are contained in the same open half-space bounded by H, or equivalently, if the closed segment with extremities x and y does not meet H; x and y are said to be on opposite sides of H if x belongs to one of the open half-spaces bounded by H and y to the other. If x∈Ex \in E and t∈Tt \in T, then x and t are said to be strictly on the same side of H if this is so for x and h+th + t for all h∈Hh \in H.

Let A be a non-empty connected subset of E. For any hyperplane H of E that does not meet A, DH(A)D_H(A) denotes the unique open half-space bounded by H that contains A. If N\mathfrak{N} is a set of hyperplanes of E, none of which meet A, put

DN(A)=⋂H∈NDH(A). D_{\mathfrak{N}}(A) = \bigcap_{H \in \mathfrak{N}} D_H(A).
If A consists of a single point a, we write DH(a)D_H(a) and DN(a)D_{\mathfrak{N}}(a) instead of DH({a})D_H(\{a\}) and DN({a})D_{\mathfrak{N}}(\{a\}).

2. Facets

The set of points of E that do not belong to any hyperplane H of the set H\mathscr{H} is open since H\mathscr{H} is locally finite. More precisely, we have the following result:

Proposition 1

Let a be a point of E. There exists a connected open neighbourhood of a that does not meet any hyperplane H that belongs to H\mathscr{H} and does not pass through a. Moreover, there exist only finitely many hyperplanes that belong to H\mathscr{H} and pass through a.

The set N\mathscr{N} of hyperplanes H such that H∈HH \in \mathscr{H} and a∉Ha \notin H is locally finite since it is contained in H\mathscr{H}. Hence, the set U of points of E that do not belong to any of the hyperplanes of the set N\mathscr{N} is open. Since a∈Ua \in U, there is a connected open neighbourhood of a contained in U. The remainder of the proposition is clear.

Given two points x and y of E, denote by R\{x, y\} the relation

" For any hyperplane H∈HH \in \mathscr{H}, either x∈Hx \in H and y∈Hy \in H or xx and yy are strictly on the same side of H."

Clearly, R is an equivalence relation on E.

Definition 1

A facet of E relative to H\mathscr{H} is an equivalence class of the equivalence relation defined above.

Proposition 2

The set of facets is locally finite.

This is clear since H\mathscr{H} is locally finite.

Let F be a facet and a a point of F. A hyperplane H∈HH \in \mathscr{H} contains F if and only if a∈Ha \in H; the set F\mathscr{F} of these hyperplanes is thus finite; their intersection is an affine subspace L of E, which we shall call the affine support of F; the dimension of L will be called the dimension of F.

If N\mathscr{N} is the set of hyperplanes H∈HH \in \mathscr{H} not containing F, then

F=L∩⋂H∈NDH(a). F = L \cap \bigcap_{H \in \mathscr{N}} D_H(a).

We shall prove that the closure of F is given by

F‾=L∩⋂H∈NDH(a)‾. \overline{F} = L \cap \bigcap_{H \in \mathscr{N}} \overline{D_H(a)}.

It is clear that the right-hand side contains the left-hand side. Conversely, let x∈L∩⋂H∈NDH(a)‾x \in L \cap \bigcap_{H \in \mathscr{N}} \overline{D_H(a)}. The open segment with extremities a and x is contained in L and in each of the DH(a)D_H(a) for H∈NH \in \mathscr{N}, and hence in F. It follows that x is in the closure of F, hence the formula.

Proposition 3

Let F be a facet and L its affine support.

(i) The set F is a convex open subset of the affine subspace L of E. (ii) The closure of F is the union of F and facets of dimension strictly smaller than that of F. (iii) In the topological space L, the set F is the interior of its closure.

Since every open half-space and every hyperplane are convex subsets of E, formula (2) shows that F is the intersection of a family of convex subsets, and hence is convex. On the other hand, let a be in F, and let U be a convex open neighbourhood of a in E that does not meet any of the hyperplanes in the set N\mathfrak{N} of H∈HH \in \mathfrak{H} such that a∉Ha \notin H. For any H∈NH \in \mathfrak{N}, we thus have U⊂DH(a)U \subset D_H(a), hence L∩U⊂FL \cap U \subset F, so that F is open in the topological space L.

Let b be a point of F‾−F\overline{F} - F, belonging to a facet F′F', and let N′\mathfrak{N}' be the set of hyperplanes H∈NH \in \mathfrak{N} passing through b; put N′′=N−N′\mathfrak{N}'' = \mathfrak{N} - \mathfrak{N}'. For any H in N′′\mathfrak{N}'', we have b∉Hb \notin H and b∈DH(a)b \in D_H(a), hence b∈DH(a)b \in D_H(a) and DH(b)=DH(a)D_H(b) = D_H(a); by the definition of a facet, we thus have

F′=L∩⋂H∈N′H∩⋂H∈N′′DH(a), F' = L \cap \bigcap_{H \in \mathfrak{N}'} H \cap \bigcap_{H \in \mathfrak{N}''} D_H(a),

whereas (3) implies that

F‾=L∩⋂H∈N′DH(a)‾∩⋂H∈N′′DH(a)‾, \overline{F} = L \cap \bigcap_{H \in \mathfrak{N}'} \overline{D_H(a)} \cap \bigcap_{H \in \mathfrak{N}''} \overline{D_H(a)},

hence F′⊂F‾F' \subset \overline{F}. We cannot have N′=∅\mathfrak{N}' = \varnothing, for this would imply that F=F′F = F' by (2) and (4), contrary to the hypothesis that b∉Fb \notin F and b∈F′b \in F'. The support of F′F' is the set L′=L∩⋂H∈N′HL' = L \cap \bigcap_{H \in \mathfrak{N}'} H; we have a∈La \in L, but a∉Ha \notin H for HH in N′\mathfrak{N}', so L′≠LL' \neq L and finally dim⁡L′<dim⁡L\dim L' < \dim L, that is dim⁡F′<dim⁡F\dim F' < \dim F. This proves (ii).

Let H be in N′\mathfrak{N}' and let D be the open half-space bounded by H and distinct from DH(a)D_H(a); we have b∈H∩Lb \in H \cap L, and it is immediate that D∩LD \cap L is a half-space of L bounded by the hyperplane H∩LH \cap L of L; consequently, every neighbourhood of b in L meets D∩LD \cap L, and since D∩LD \cap L is disjoint from F‾\overline{F} by (3), we see that the point b of F‾−F\overline{F} - F cannot be in the interior of F‾\overline{F} in the topological space L. Since F is open in L, we have (iii). Q.E.D.

Corollary

Let F and F' be two facets. If F‾=F‾′\overline{F} = \overline{F}', the facets F and F' are equal.

This follows from (iii).

Proposition 4

Let F be a facet, and let L be an affine subspace of E that is the intersection of hyperplanes belonging to H\mathfrak{H}; denote by N\mathfrak{N} the set of hyperplanes H∈HH \in \mathfrak{H} that do not contain L. The following conditions are equivalent:

(i) There exists a facet F' with support L that meets F‾\overline{F}. (ii) There exists a facet F' with support L contained in F‾\overline{F}. (iii) There exists a point x in L∩F‾L \cap \overline{F} that does not belong to any of the hyperplanes of N\mathfrak{N}.

If these conditions are satisfied, L∩DN(F)L \cap D_{\mathfrak{N}}(F) is the unique facet with support LL contained in F‾\overline{F}.

(i) ⟹\Longrightarrow (ii): Since F‾\overline{F} is a union of facets (Prop. 3, (ii)), every facet that meets F‾\overline{F} meets a facet contained in F‾\overline{F}, and so is equal to it.

(ii) ⟹\Longrightarrow (iii): Every point xx of F′F' satisfies (iii) since every hyperplane of H\mathfrak{H} containing xx contains F′F', and hence LL.

(iii) ⟹\Longrightarrow (i): Let xx be a point satisfying (iii) and let F′F' be the facet containing xx; it is clear that F′F' meets F‾\overline{F}. Let H∈HH \in \mathfrak{H}; then x∉Hx \notin H if H∈NH \in \mathfrak{N} and clearly x∈Hx \in H if H∉NH \notin \mathfrak{N}; consequently the support of F′F' is the intersection of the hyperplanes of H−N\mathfrak{H} - \mathfrak{N}, and is equal to LL.

Finally, let F′F' be a facet with support LL contained in F‾\overline{F}, and let xx be a point of F′F'; since no hyperplane of N⊂H\mathfrak{N} \subset \mathfrak{H} passes through xx, Prop. 1 shows that there exists a convex open neighbourhood UU of xx that does not meet any hyperplane of N\mathfrak{N}. Since xx is in the closure of FF, U∩F≠∅U \cap F \neq \varnothing; now N\mathfrak{N} is the set of hyperplanes H∈HH \in \mathfrak{H} that do not contain F′F', and for all HH in N\mathfrak{N} we have DH(x)=DH(U)=DH(U∩F)=DH(F)D_H(x) = D_H(U) = D_H(U \cap F) = D_H(F) and formula (2) implies that

F′=L∩DN(F). F' = L \cap D_{\mathfrak{N}}(F).

Q.E.D.

3. Chambers

Definition 2

A chamber of EE relative to H\mathfrak{H} (or simply a chamber if there is no ambiguity regarding H\mathfrak{H}) is a facet of EE relative to H\mathfrak{H} that is not contained in any hyperplane belonging to H\mathfrak{H}.

Let UU be the open set in EE consisting of the points that do not belong to any hyperplane of H\mathfrak{H}; since a hyperplane of H\mathfrak{H} must contain any facet that it meets, the chambers are the facets contained in UU; every chamber is a convex (hence connected) open subset of EE by Prop. 3, (i); since the chambers form a partition of UU, they are exactly the connected components of UU. Every convex subset AA of UU is connected, and thus contained in a chamber, which is unique if AA is non-empty. It is clear that the chambers are the facets with support EE, and Prop. 3, (iii) shows that every chamber is the interior of its closure. Finally, let CC be a chamber and AA a non-empty subset of CC; formulas (2) and (3) imply that

C=⋂H∈HDH(A)=DH(A),C‾=⋂H∈HDH(A)‾ C = \bigcap_{H \in \mathfrak{H}} D_H(A) = D_{\mathfrak{H}}(A), \quad \overline{C} = \bigcap_{H \in \mathfrak{H}} \overline{D_H(A)}

since DH(A)=DH(a)D_H(A) = D_H(a) for all a∈Aa \in A.

Proposition 5

Let CC be a non-empty subset of EE. Assume that there exists a subset H′\mathfrak{H}' of H\mathfrak{H} with the following properties:

a) For any H∈H′H \in \mathfrak{H}', there exists an open half-space DHD_H bounded by HH such that C=⋂H∈H′DHC = \bigcap_{H \in \mathfrak{H}'} D_H.

b) *The set C does not meet any hyperplane belonging to H−H′\mathfrak{H} - \mathfrak{H}'. Under these conditions, C is a chamber defined by H\mathfrak{H} in E, and DH=DH(C)D_H = D_H(C) for all H∈HH \in \mathfrak{H}.

Properties a) and b) show that C is a convex subset of U; hence there is a chamber C′C' with C⊂C′C \subset C'. Since C⊂DHC \subset D_H, we have DH=DH(C)D_H = D_H(C) for all H in H′\mathfrak{H}', hence C=DH′(C)⊃DH(C)C = D_{\mathfrak{H}'}(C) \supset D_{\mathfrak{H}}(C) since H′⊂H\mathfrak{H}' \subset \mathfrak{H}; we have DH(C)=C′D_{\mathfrak{H}}(C) = C' by (6), hence C⊃C′C \supset C'. Finally therefore, C=C′C = C'.

Proposition 6

Every point of E is in the closure of at least one chamber.

If E reduces to a single point, this is clear. Otherwise, let a∈Ea \in E and let H1,…,HmH_1, \ldots, H_m be the hyperplanes of H\mathfrak{H} containing aa. Since H\mathfrak{H} is locally finite, there exists a neighbourhood V of aa that does not meet any hyperplane of H\mathfrak{H} other than H1,…,HmH_1, \ldots, H_m. Let D be a straight line passing through aa and not contained in any of the HiH_i; if x∈D,x≠ax \in D, x \neq a, and xx is sufficiently close to aa, the open segment ]ax[]ax[ is contained in V and does not meet any of the HiH_i. Then ]ax[⊂U]ax[ \subset U; since ]ax[]ax[ is connected, it is contained in a chamber C, hence a∈C‾a \in \overline{C}.

Proposition 7

Let L be an affine subspace of E and Ω\Omega a non-empty open subset of L.

(i) There exists a point a in Ω\Omega that does not belong to any of the hyperplanes of H\mathfrak{H} that do not contain L.

(ii) If L is a hyperplane and L∉HL \notin \mathfrak{H}, there exists a chamber that meets Ω\Omega.

(iii) If L is a hyperplane and L∈HL \in \mathfrak{H}, there exists a point a in Ω\Omega that does not belong to any hyperplane H≠LH \neq L of H\mathfrak{H}.

Denote by N\mathfrak{N} the set of hyperplanes H with H∈HH \in \mathfrak{H} and L⊄HL \not\subset H, and by L\mathscr{L} the set of hyperplanes of the affine space L of the form L∩HL \cap H with H∈NH \in \mathfrak{N}. It is clear that L\mathscr{L} is a locally finite set of hyperplanes in L, and Prop. 6 shows that Ω\Omega meets a chamber Γ\Gamma defined by L\mathscr{L} in L. If aa is a point of Γ∩Ω\Gamma \cap \Omega, then a∉Ha \notin H for all H∈NH \in \mathfrak{N}, hence (i).

Assume now that L is a hyperplane; any hyperplane containing L is then equal to it, so we may distinguish two cases:

a) L∉HL \notin \mathfrak{H}: then N=H\mathfrak{N} = \mathfrak{H}, and we have a∉Ha \notin H for all H∈HH \in \mathfrak{H}; thus aa belongs to a chamber defined by H\mathfrak{H} in E, hence (ii).

b) L∈HL \in \mathfrak{H}: then N=H−{L}\mathfrak{N} = \mathfrak{H} - \{L\}, hence (iii).

4. Walls and faces

Definition 3

Let C be a chamber of E. A face of C is a facet contained in the closure of C whose support is a hyperplane. A wall of C is a hyperplane that is the support of a face of C.

Every wall of C belongs to H\mathfrak{H}. Prop. 4 shows that a hyperplane L∈HL \in \mathfrak{H} is a wall of C if and only if C≠DH−{L}(C)C \neq D_{\mathfrak{H} - \{L\}}(C). Moreover, every wall of C is the support of a single face of C.

Proposition 8

Every hyperplane* HH belonging to H\mathfrak{H} is the wall of at least one chamber.

By Prop. 7, (iii), there exists a point aa of HH that does not belong to any hyperplane H′≠HH' \neq H of H\mathfrak{H}; by Prop. 6, there exists a chamber CC such that a∈C‾a \in \overline{C}; Prop. 4 then shows that HH is a wall of CC.

Proposition 9

Let* CC be a chamber and M\mathscr{M} the set of walls of CC. Then C=DM(C)C = D_{\mathscr{M}}(C) and every subset L\mathscr{L} of H\mathfrak{H} such that C=DL(C)C = D_{\mathscr{L}}(C) contains M\mathscr{M}. A subset FF of C‾\overline{C} is a facet if and only if it is a facet of EE relative to the family M\mathscr{M}.

a) Let L\mathscr{L} be a subset of H\mathfrak{H} such that C=DL(C)C = D_{\mathscr{L}}(C). Consider a hyperplane LL belonging to H\mathfrak{H} but not to L\mathscr{L}; let N\mathfrak{N} be the set of hyperplanes H≠LH \neq L belonging to H\mathfrak{H}. Then L⊂N\mathscr{L} \subset \mathfrak{N}, hence C=DN(C)C = D_{\mathfrak{N}}(C), and LL does not meet DN(C)D_{\mathfrak{N}}(C). By the implication (i) ⟹\Longrightarrow (iii) in Prop. 4, the hyperplane LL is not a wall of CC. Consequently, every wall of CC belongs to L\mathscr{L}.

b) We assume that C=DL(C)C = D_{\mathscr{L}}(C). Let HH be a hyperplane belonging to L\mathscr{L} that is not a wall of CC, and put L′=L−{H}\mathscr{L}' = \mathscr{L} - \{ H \}. By the implication (iii) ⟹\Longrightarrow (i) in Prop. 4, the convex set DL′(C)D_{\mathscr{L}'}(C) does not meet HH, so DL′(C)⊂DH(C)D_{\mathscr{L}'}(C) \subset D_H(C) and C=DL′(C)C = D_{\mathscr{L}'}(C). If F\mathfrak{F} is a finite subset of L\mathscr{L} that does not contain any wall of CC, we conclude by induction on the cardinal of F\mathfrak{F} that C=DL−F(C)C = D_{\mathscr{L} - \mathfrak{F}}(C).

c) Let aa be a point of CC; clearly, C⊂DM(a)C \subset D_{\mathfrak{M}}(a). Let a′a' be a point of DM(a)D_{\mathfrak{M}}(a); since the closed segment [aa′][aa'] is compact, the set F\mathfrak{F} of hyperplanes H∈HH \in \mathfrak{H} that meet [aa′][aa'] is finite. Since aa and a′a' are strictly on the same side of every wall of CC, no wall of CC belongs to F\mathfrak{F}; by b), we have C=DH−F(C)C = D_{\mathfrak{H} - \mathfrak{F}}(C). Since a′∈DH−F(a)a' \in D_{\mathfrak{H} - \mathfrak{F}}(a), we have a′∈Ca' \in C. We have therefore proved that DM(a)⊂CD_{\mathfrak{M}}(a) \subset C, which establishes the first part of the proposition.

d) To prove the last assertion of the proposition, it clearly suffices to show that a subset FF of C‾\overline{C} that is a facet of EE relative to M\mathscr{M} is a facet of EE relative to H\mathfrak{H}, or that every hyperplane H∈HH \in \mathfrak{H} that meets FF contains FF. So let HH be a hyperplane that meets FF but does not contain it. Since FF is open in its affine support, it is not completely on one side of HH. It follows that C‾\overline{C} is not completely on one side of HH and hence that the hyperplane HH does not belong to H\mathfrak{H}, which completes the proof.

Remark 1

It follows from formula (6) and Prop. 9 that the closure of a chamber CC is the intersection of the closed half-spaces that are bounded by a wall of CC and contain CC.

Remark 2

Let FF be a facet whose support is a hyperplane LL; we shall show that FF is a face of two chambers. Let N\mathfrak{N} be the set of hyperplanes H≠LH \neq L belonging to H\mathfrak{H}; put A=DN(F)A = D_{\mathfrak{N}}(F) and denote by D+D^+ and D−D^- the two open half-spaces bounded by LL. The set AA is open and contains F⊂LF \subset L, and since every point of LL is in the closure of D+D^+ and D−D^-, the sets C+=A∩D+C^+ = A \cap D^+ and C−=A∩D−C^- = A \cap D^- are non-empty; these are chambers. Moreover, the hyperplane LL meets DN(F)=DN(C+)D_{\mathfrak{N}}(F) = D_{\mathfrak{N}}(C^+); Prop. 4 shows that LL is a wall of C+C^+ and that

F, which meets L∩Dγl(F)L \cap D_{\gamma_l}(F), is a face of C+C^+; similarly, F is a face of C−C^-. Finally, let C be a chamber of which F is a face, and suppose for example that D+=DL(C)D^+ = D_L(C); by Prop. 4, the set Dγ(C)D_{\gamma}(C) meets F, and hence is equal to Dγ(F)D_{\gamma}(F), and we have

C=Dδ(C)=DL(C)∩Dγ(C)=D+∩Dγ(F)=C+. C = D_{\delta}(C) = D_L(C) \cap D_{\gamma}(C) = D^+ \cap D_{\gamma}(F) = C^+.

5. Intersecting hyperplanes

Recall (Algebra, Chap. II, § 9, no. 3) that two affine subspaces P and P′P' of E are said to be parallel if there exists a vector t in T such that P′=t+PP' = t + P. It is clear that the relation "P and P′P' are parallel" is an equivalence relation on the set of affine subspaces of E.

Lemma 1

Any two non-parallel hyperplanes have non-empty intersection.

Let H and H′H' be two non-parallel hyperplanes, a∈Ha \in H and a′∈H′a' \in H'; there exist two hyperplanes M and M′M' of the vector space T such that H=M+aH = M + a and H′=M′+a′H' = M' + a'; since H and H′H' are not parallel, we have M≠M′M \neq M' and hence T=M+M′T = M + M'; hence there exist u∈Mu \in M and u′∈M′u' \in M' such that a′−a=u−u′a' - a = u - u', and the point u+a=u′+a′u + a = u' + a' belongs to H∩H′H \cap H'.

Lemma 2

Let H and H′H' be two distinct hyperplanes of E, and f, f′f' two affine functions on E such that H (resp. H′H') consists of the points a in E such that f(a)=0f(a) = 0 (resp. f′(a)=0f'(a) = 0). Finally, let L be a hyperplane of E. Assume that one of the following hypotheses is satisfied: a) The hyperplanes H, H′H' and L are parallel. b) The hyperplanes H and H′H' are not parallel, and H∩H′⊂LH \cap H' \subset L. Then there exist real numbers λ,λ′\lambda, \lambda', not both zero, such that L consists of those points a∈Ea \in E at which the affine function g=λ.f+λ′.f′g = \lambda . f + \lambda' . f' vanishes.

The lemma being trivial when L=HL = H, we assume that there exists a point aa in L with a∉Ha \notin H. Put λ=f′(a),λ′=−f(a)\lambda = f'(a), \lambda' = -f(a) and

g=λ.f+λ′.f′; g = \lambda . f + \lambda' . f';

then λ′≠0\lambda' \neq 0 since a∉Ha \notin H; moreover, since H≠H′H \neq H', there exists b∈Hb \in H such that b∉H′b \notin H', so that f(b)=0,f′(b)≠0f(b) = 0, f'(b) \neq 0, and thus g(b)=−f(a).f′(b)g(b) = -f(a).f'(b) is non-zero. The set L1L_1 of points where the affine function g≠0g \neq 0 vanishes is then a hyperplane of E; we have g(a)=0g(a) = 0, so a∈L1a \in L_1.

a) Assume that H and H′H' are parallel: g and f both vanish at every point of L1∩HL_1 \cap H, hence so does f′f' since λ′≠0\lambda' \neq 0; thus every point of L1∩HL_1 \cap H belongs to H′H'; but since H and H′H' are parallel and distinct, they are disjoint, so L1∩H=∅L_1 \cap H = \varnothing, and Lemma 1 shows that L1L_1 is parallel to H. Since a∈La \in L and a∈L1a \in L_1, we thus have L=L1L = L_1.

b) Assume that H and H′H' are not parallel: by Lemma 1, there is a point c in H∩H′H \cap H'; we give E the vector space structure obtained by taking c as the origin. Then H∩H′H \cap H' is a vector subspace of EE of codimension 2, and since a∉Ha \notin H, the vector subspace MM of EE generated by H∩H′H \cap H' and aa is a hyperplane; since H∩H′⊂L∩L1H \cap H' \subset L \cap L_1 and a∈L∩L1a \in L \cap L_1, we have M⊂L∩L1M \subset L \cap L_1, hence M=L=L1M = L = L_1. Q.E.D.

Proposition 10

Let C be a chamber, let H and H' be two walls of C, and let L be a hyperplane meeting DH(C)∩DH′(C)D_H(C) \cap D_{H'}(C). Assume that H is distinct from H' and that one of the following conditions is satisfied:

a) The hyperplanes H, H' and L are parallel.

b) The hyperplanes H and H' are not parallel, and H∩H′‾⊂L\overline{H \cap H'} \subset L.

Then, L meets C.

Let bb (resp. b′b') be a point of the face of C with support H (resp. H′H'); it is immediate that every point of the segment [bb′][bb'] distinct from bb and b′b' belongs to C.

Introduce an affine function ff that vanishes at every point of H and is such that f(x)>0f(x) > 0 for xx in DH(C)D_H(C); similarly, introduce an affine function f′f' having an analogous property with respect to H′H'. By applying Lemma 2, we can find numbers λ\lambda and λ′\lambda' and an affine function gg having the properties stated in the lemma. We have (λ,λ′)≠(0,0)(\lambda, \lambda') \neq (0, 0), and for every point xx of L∩DH(C)∩DH′(C)L \cap D_H(C) \cap D_{H'}(C), we have f(x)>0, f′(x)>0f(x) > 0,\ f'(x) > 0 and λ.f(x)+λ′.f′(x)=0\lambda.f(x) + \lambda'.f'(x) = 0, so λλ′<0\lambda \lambda' < 0. On the other hand, we have g(b)=λ′.f′(b)g(b) = \lambda'.f'(b) and g(b′)=λ.f(b′)g(b') = \lambda.f(b'), and since f(b′)>0, f′(b)>0f(b') > 0,\ f'(b) > 0, we have g(b)g(b′)<0g(b)g(b') < 0. The points bb and b′b' are thus strictly on opposite sides of the hyperplane L, hence there exists a point c of L which belongs to [bb′][bb'], and is distinct from bb and b′b', hence that belongs to C.

6. Examples: simplicial cones and simplices

a) Let aa be a point of EE, and (e1,…,ed)(e_1, \ldots, e_d) a basis of TT; every point of EE can thus be written uniquely in the form

x=a+ξ1.e1+⋯+ξd.ed x = a + \xi_1.e_1 + \cdots + \xi_d.e_d

with ξ1,…,ξd\xi_1, \ldots, \xi_d being real numbers. Denote by ei′e'_i the affine function on EE which, at any x∈Ex \in E, written in the form (7), takes the value ξi\xi_i. Denote by HiH_i the hyperplane formed by those xx such that ei′(x)=0e'_i(x) = 0, and by H\mathfrak{H} the set of hyperplanes H1,…,HdH_1, \ldots, H_d. For every subset J of the set I={1,2,…,d}I = \{1, 2, \ldots, d\}, put HJ=⋂i∈JHiH_J = \bigcap_{i \in J} H_i; for every sequence (ε1,…,εd)(\varepsilon_1, \ldots, \varepsilon_d) of numbers equal to 0, 1 or -1, denote by F(ε1,…,εd)F(\varepsilon_1, \ldots, \varepsilon_d) the set of those x∈Ex \in E such that ei′(x)e'_i(x) is of sign εi\varepsilon_i for all ii in I (General Topology, Chap. 4, § 3, no. 2). It is immediate that the facets defined by H\mathfrak{H} in EE are the sets F(ε1,…,εd)F(\varepsilon_1, \ldots, \varepsilon_d) and that these sets are pairwise distinct; the support of F(ε1,…,εd)F(\varepsilon_1, \ldots, \varepsilon_d) is equal to HJH_J if J is the set of i∈Ii \in I such that εi=0\varepsilon_i = 0; in particular, the chambers are the sets of the form F(ε1,…,εd)F(\varepsilon_1, \ldots, \varepsilon_d) where each of the numbers εi\varepsilon_i is equal to 1 or -1.

The set C=F(1,…,1)C = F(1, \ldots, 1) formed by those xx with ei′(x)>0e'_i(x) > 0 for all i∈Ii \in I is a chamber, called the open simplicial cone with vertex a defined by the basis (e_1, \ldots, e_d). Its closure consists of the points x such that e_i'(x) \geq 0 for all i \in I when d \geq 1; otherwise, its closure is empty. For every subset J of I, let C_J be the set of points x of E such that e_i'(x) = 0 for i \in J and e_i'(x) > 0 for i \in I - J. Then C_J is a facet with support H_J, and it is an open simplicial cone with vertex a in the affine space H_J; moreover, C‾=⋃J⊂ICJ\overline{C} = \bigcup_{J \subset I} C_J. In particular, the walls of C are the hyperplanes H_i for i \in I, and the face of C contained in H_i is equal to C_{\{i\}}.

None of the sets H_i, H_J, C, C_J and F(\varepsilon_1, \ldots, \varepsilon_d) change if we replace the basis (e_1, \ldots, e_d) by a basis (\lambda_1 e_1, \ldots, \lambda_d e_d) with \lambda_i > 0 for all i.

b) Suppose now that we are given an affinely free system of points of E, say (a_0, a_1, \ldots, a_d). We know that every point of E can be written uniquely in the form x=ξ0.a0+⋯+ξd.adx = \xi_0.a_0 + \cdots + \xi_d.a_d, where \xi_0, \ldots, \xi_d are real numbers with \xi_0 + \cdots + \xi_d = 1 (Algebra, Chap. 2, § 9, no. 3). Define affine functions f0,…,fdf_0, \ldots, f_d, the function fif_i associating to the point x, written as above, the number \xi_i. Denote by H_i the hyperplane of E defined by fi(x)=0f_i(x) = 0 and by H\mathfrak{H} the set of hyperplanes H_0, H_1, \ldots, H_d; finally, put I = \{0, 1, \ldots, d\}. The open simplex with vertices a_0, \ldots, a_d is the set C of points x of E such that fi(x)>0f_i(x) > 0 for all i \in I; it is one of the chambers defined by H\mathfrak{H} in E. The closure of C is the set C‾\overline{C} of points x such that fi(x)≥0f_i(x) \geq 0 for all i \in I; it is the convex envelope of the finite set {a0,…,ad}\{a_0, \ldots, a_d\} and it is easy to see that the extreme points of C‾\overline{C} are a0,…,ada_0, \ldots, a_d.

For any subset J of I distinct from I, put HJ=⋂i∈JHiH_J = \bigcap_{i \in J} H_i and let C_J be the set of points x of E such that fi(x)=0f_i(x) = 0 for i \in J and fi(x)>0f_i(x) > 0 for i \in I - J. The set C_J is an open simplex in the affine space H_J with vertices the points a_i for i \in I - J; we have C∅=C,C‾=⋃J≠ICJC_\varnothing = C, \overline{C} = \bigcup_{J \neq I} C_J and CJ≠CJ′C_J \neq C_{J'} if J≠J′J \neq J'; moreover, C_J is a facet with support H_J. In particular, the walls of C are H_0, \ldots, H_d and C_{\{i\}} is the face contained in H_i.

For any non-empty subset K of I, let B_K be the set of points x of E such that fi(x)>0f_i(x) > 0 for i \in K and fi(x)<0f_i(x) < 0 for i \in I - K. The sets B_K are the chambers defined by H\mathfrak{H} in E and we have B_I = C. It is easy to see that C‾\overline{C} is compact; on the other hand, if K is a subset of I distinct from I, of cardinal p, the chamber B_K contains the sequence of points xnx_n defined for n \geq 2 by

fi(xn)={nfor i∈K(1−pn)/(d+1−p)for i∈I−K f_i(x_n) = \begin{cases} n & \text{for } i \in K \\ (1 - pn)/(d + 1 - p) & \text{for } i \in I - K \end{cases}

showing that B_K is not relatively compact.

Transcribed from the printed volume. Not translated.