Bourbaki

Definition 1

Tag 07X0, permanent. Cite it as [Bourbaki, Tag 07X0].

Topological Vector Spaces, chapter V, § 1. Prehilbertian spaces and hilbertian spaces

Let E be a vector space over the field K. A hermitian form (on the left) on E is a map f from E × E into K satisfying the following conditions (for x1,x2,x,y1,y2,yx_1, x_2, x, y_1, y_2, y in E and λ,μ\lambda, \mu in K):

(1)

{f(x1+x2,y)=f(x1,y)+f(x2,y)f(x,y1+y2)=f(x,y1)+f(x,y2) \begin{cases} f(x_1 + x_2, y) = f(x_1, y) + f(x_2, y) \\ f(x, y_1 + y_2) = f(x, y_1) + f(x, y_2) \end{cases}

(2)

{f(λx,y)=λ‾f(x,y)f(x,μy)=μf(x,y) \begin{cases} f(\lambda x, y) = \overline{\lambda} f(x, y) \\ f(x, \mu y) = \mu f(x, y) \end{cases}

(3)

f(x,y)=f(y,x)‾. f(x, y) = \overline{f(y, x)} .

When the field K is R, the notion of hermitian form on E reduces to that of symmetric bilinear form on E × E (A, III, § 6, No. 3).

We note that the second condition (1) and the second condition (2) follow from the other three.

¹ For the reader specially interested in hilbertian spaces, we point out that only No. 7 of § 1 and No. 8 of § 4 depend on results of chapters III and IV. For this the reader can consult « Summary of some important properties of Banach spaces » which appears at the end of this volume. The only references to chapters I and II concern the definition of a convex set and of a semi-norm (II, p. 1 and p. 7), that of a topological direct sum (I, p. 4), of a total family and a topologically independent family (I, p. 12).

From (1) and (2) we deduce immediately that

f(∑jλjxj,∑kμkyk)=∑j,kλ‾jμkf(xj,yk). f(\sum_j \lambda_j x_j, \sum_k \mu_k y_k) = \sum_{j,k} \overline{\lambda}_j \mu_k f(x_j, y_k).

In particular, if E is finite dimensional, and if (ej)1≤j≤n(e_j)_{1 \leq j \leq n} is a basis of E, then for x=∑j=1nξjejx = \sum_{j=1}^n \xi_j e_j and y=∑j=1nηjejy = \sum_{j=1}^n \eta_j e_j, we have,

f(x,y)=∑j,kαjkξ‾jηk f(x, y) = \sum_{j,k} \alpha_{jk} \overline{\xi}_j \eta_k

with the notation αjk=f(ej,ek)\alpha_{jk} = f(e_j, e_k); moreover, relation (3) amounts to αjk=α‾kj\alpha_{jk} = \overline{\alpha}_{kj} for every pair of indices j,kj, k; this implies in particular that the numbers αjj\alpha_{jj} are real.

From (3), the number Q(x)=f(x,x)Q(x) = f(x, x) is real for all x∈Ex \in E. Moreover, we immediately establish the following formulas, known as polarization formulas

(5)

4f(x,y)=∑ε2=1εQ(x+εy)if K is R, 4f(x, y) = \sum_{\varepsilon^2 = 1} \varepsilon Q(x + \varepsilon y) \quad \text{if } K \text{ is } \mathbf{R},
(6)
4f(x,y)=∑ε4=1,ε∈CεQ(x+ε‾y)if K is C. 4f(x, y) = \sum_{\varepsilon^4 = 1, \varepsilon \in \mathbf{C}} \varepsilon Q(x + \overline{\varepsilon} y) \quad \text{if } K \text{ is } \mathbf{C}.

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.