Bourbaki

Example

Tag 08JP, permanent. Cite it as [Bourbaki, Tag 08JP].

Integration, chapter II, § 1. Riesz spaces and fully lattice-ordered spaces

The space RA\mathbf{R}^A of all real-valued functions defined on any set AA is a Riesz space (for the order relation ⟨x(t)≤y(t) for all t∈A⟩\langle x(t) \leq y(t) \text{ for all } t \in A \rangle); for, any two real-valued functions x,yx, y defined on AA have a supremum (resp. an infimum) equal to the mapping t↦sup⁡(x(t),y(t))t \mapsto \sup(x(t), y(t)) (resp. t↦inf⁡(x(t),y(t))t \mapsto \inf(x(t), y(t))).

One can also say that a Riesz space is a vector space EE equipped with an order structure such that, on the one hand, this structure and the additive group structure of EE define a lattice-ordered group structure on EE (A, VI, §1, No. 9), and on the other hand that the axiom (OVSII_{\mathrm{II}}) is satisfied.

Thus, all of the properties of lattice-ordered groups are applicable to Riesz spaces; we shall review here the principal ones (cf. A, VI, §1, Nos. 9 to 12), as well as indicating the consequences that flow from the axiom (OVSII_{\mathrm{II}}).

We recall first that one writes x+=sup⁡(x,0)x^+ = \sup(x, 0) (the positive part of xx), x−=(−x)+=sup⁡(−x,0)x^- = (-x)^+ = \sup(-x, 0) (the negative part of xx), ∣x∣=sup⁡(x,−x)|x| = \sup(x, -x) (the absolute value of xx); then x=x+−x−x = x^+ - x^- and ∣x∣=x++x−|x| = x^+ + x^-; here, these two relations are equivalent to

x+=12(∣x∣+x),x−=12(∣x∣−x). x^+ = \frac{1}{2}(|x| + x), \quad x^- = \frac{1}{2}(|x| - x).

The relation x≤yx \leq y is equivalent to ⟨x+≤y+ and x−≥y−⟩\langle x^+ \leq y^+ \text{ and } x^- \geq y^- \rangle. For any xx and yy, the triangle inequality holds:

(1)

∣x+y∣≤∣x∣+∣y∣. |x + y| \leq |x| + |y|.

By the invariance of order under every homothety of ratio >0> 0,

(2)

sup⁡(λx,λy)=λsup⁡(x,y)for all λ≥0. \sup(\lambda x, \lambda y) = \lambda \sup(x, y) \quad \text{for all } \lambda \geq 0.

In particular,

(3)

(λx)+=λx+,(λx)−=λx−for all λ≥0. (\lambda x)^+ = \lambda x^+, \quad (\lambda x)^- = \lambda x^- \quad \text{for all } \lambda \geq 0.

On the other hand, for λ<0\lambda < 0 we have (λx)+=(−λx)−=∣λ∣x−(\lambda x)^+ = (-\lambda x)^- = |\lambda| x^- and (λx)−=(−λx)+=∣λ∣x+(\lambda x)^- = (-\lambda x)^+ = |\lambda| x^+; it follows that, for every λ∈R\lambda \in \mathbf{R} and every x∈Ex \in E,

(4)

∣λx∣=∣λ∣⋅∣x∣. |\lambda x| = |\lambda| \cdot |x|.

The invariance of order under translation shows that for all z∈Ez \in E,

(5)

sup⁡(x+z,y+z)=z+sup⁡(x,y), \sup(x + z, y + z) = z + \sup(x, y),
whence, in particular,

(6)

sup⁡(x,y)=x+(y−x)+=12(x+y+∣x−y∣). \sup(x, y) = x + (y - x)^+ = \frac{1}{2}(x + y + |x - y|).

We have the relations

(7) inf⁡(x,y)=−sup⁡(−x,−y),\inf(x, y) = -\sup(-x, -y) ,

(8) sup⁡(x,y)+inf⁡(x,y)=x+y.\sup(x, y) + \inf(x, y) = x + y .

If x,y,zx, y, z are ≥0\geq 0 then (A, VI, §1, No. 12, Prop. 11)

(9) inf⁡(x+y,z)≤inf⁡(x,z)+inf⁡(y,z).\inf(x + y, z) \leq \inf(x, z) + \inf(y, z) .

If A and B are two subsets of E each of which has a supremum, then A + B also has a supremum and

(10) sup⁡(A+B)=sup⁡A+sup⁡B.\sup(A + B) = \sup A + \sup B .

Two elements x,yx, y of E are said to be alien3^3 (to each other) if inf⁡(∣x∣,∣y∣)=0\inf(|x|, |y|) = 0; by (8), this relation is equivalent to sup⁡(∣x∣,∣y∣)=∣x∣+∣y∣\sup(|x|, |y|) = |x| + |y|, and also, by (6), to ∣∣x∣−∣y∣∣=∣x∣+∣y∣||x| - |y|| = |x| + |y|; 0 is the only element alien to itself; for every x∈Ex \in E, x+x^+ and x−x^- are alien and may be characterized as the only alien elements y≥0,z≥0y \geq 0, z \geq 0 such that x=y−zx = y - z. If yy is alien to xx, then every z∈Ez \in E such that ∣z∣≤∣y∣|z| \leq |y| is also alien to xx. If yy and zz are alien to xx, then so is ∣y∣+∣z∣|y| + |z|, by the inequality (9); in particular, n∣y∣n|y| is alien to xx for every integer n>0n > 0, from which it follows that λy\lambda y is alien to xx for every scalar λ\lambda, since there exists an integer nn such that ∣λ∣≤n|\lambda| \leq n, whence ∣λy∣≤n∣y∣|\lambda y| \leq n|y|. If a subset A of E consists of elements alien to xx and if A has a supremum, then that supremum is also alien to xx (A, VI, §1, No. 12, Cor. of Prop. 13).

Finally, we have the decomposition lemma (A, VI, §1, No. 10, Th. 1):

If (xi)i∈I,(yj)j∈J(x_i)_{i \in I}, (y_j)_{j \in J} are two finite sequences of elements ≥0\geq 0 of E such that ∑i∈Ixi=∑j∈Jyj\sum_{i \in I} x_i = \sum_{j \in J} y_j, then there exists a finite sequence (zij)(i,j)∈I×J(z_{ij})_{(i,j) \in I \times J} of elements ≥0\geq 0 of E such that xi=∑j∈Jzijx_i = \sum_{j \in J} z_{ij} for all i∈Ii \in I, and yj=∑i∈Izijy_j = \sum_{i \in I} z_{ij} for all j∈Jj \in J.

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.