Bourbaki

Remarks

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

Algebra, chapter IX, § 8. Types de formes quadratiques

1) Every non-degenerate quadratic form QQ of finite dimension whose type is zero (i.e. such that θ(Q)=T0\theta(Q) = T_0 with the above notation) is a neutral form. There indeed exist neutral forms N,N′N, N' such that QτNQ \tau N is equivalent to N′N'. This shows that QQ is of even dimension, and hence that there exists a neutral form N1N_1 having the same dimension as QQ. Since QQ and N1N_1 have the same type, it follows from prop. 1 that they are equivalent, and hence that QQ is neutral.

2) For every finite-dimensional quadratic form QQ over AA, let δ(Q)\delta(Q) denote the class modulo 2 of the dimension of QQ. We have

δ(QτQ′)=δ(Q)+δ(Q′). \delta(Q \tau Q') = \delta(Q) + \delta(Q').
Since every neutral form NN is of even dimension, we have δ(N)=0\delta(N) = 0; the relation Q∼Q′Q \sim Q' therefore entails δ(Q)=δ(Q′)\delta(Q) = \delta(Q'). Thus the restriction of δ\delta to the group W\mathfrak{W} of types of quadratic forms over AA is a homomorphism of W\mathfrak{W} into the group Z/(2)\mathbf{Z}/(2). This homomorphism is surjective when AA is of characteristic ≠2\neq 2, but is not so if AA is of characteristic 2, since a quadratic form of odd dimension is then degenerate, its associated bilinear form being alternating (cf. § 5).

3) Let aa be an element ≠0\neq 0 of AA. If NN is a neutral form, then so is aNaN. It follows that the relation Q∼Q′Q \sim Q' entails aQ∼aQ′aQ \sim aQ'. For every element TT of the group W\mathfrak{W}, we shall put

a.T=θ(aT). a.T = \theta(aT).
We thus obtain an external composition law between the group A∗A^* of the non-zero elements of AA and the group W\mathfrak{W}. The following formulas result immediately from the definition:
a.(T+T′)=a.T+a.T′,ab.T=a.(b.T) a.(T + T') = a.T + a.T', \quad ab.T = a.(b.T)
(a,b in A∗,T,T′ in W). (a, b \text{ in } A^*, \quad T, T' \text{ in } \mathfrak{W}).

On the other hand, if a,ba, b and a+ba + b are in A∗A^*, one does not in general have (a+b).T=a.T+b.T(a + b).T = a.T + b.T (T∈WT \in \mathfrak{W}).

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

Cited by

Nothing.

Cites

Nothing.

Translated with machine help, not yet reviewed by a person.