Let Φ be a sesquilinear form for J on E, (x~1,…,xn) a system of n elements of E, and (aij)(i,j=1,…,n) a family of n2 elements of A; put yi=∑i=1najixi. We then have
DΦ(y1,…,yn)=det(aij)⋅det(aij)j⋅DΦ(x1,…,xn).
Indeed, since Φ is a sesquilinear form, we have Φ(yi,yj)=∑k,maikΦ(xk,xm)ajmj. Hence, if we denote by A the matrix (aij), the matrix (Φ(yi,yj)) is equal to A.(Φ(xi,xj)).tAj. Since det(tA)=det(A), and det(Aj)=det(A)j, our assertion is proved.
In particular, if (ei) and (ei′) are two bases of E,D and D′ the discriminants of Φ with respect to these bases, and a the determinant of the passing matrix from the basis (ei) to the basis (ei′), we have
(4)D′=aajD.
It follows from prop. 1 that, if (ei) is a basis of E and (xi) an arbitrary system of n elements of E, DΦ(e1,…,en) divides DΦ(x1,…,xn). In particular the discriminants of Φ with respect to any two bases of E generate the same principal ideal of A.
Let (Ei)i∈I be a finite family of finite-dimensional free A-modules, Φi a sesquilinear form for J on Ei, and Bi a basis of Ei. If Φ denotes the direct sum of the Φi (§ 1, no 3) and B the basis of ∏i∈IEi obtained by joining the Bi, then obviously
(5)DΦ(B)=i∈I∏DΦi(Bi).
Let Φ be a sesquilinear form for J on E, h a homomorphism of A into a commutative ring A′, Φ′ the sesquilinear form on A′⊗AE obtained by extension of Φ (§ 1, no 4) and (x1,…,xn) an arbitrary system of elements of E. Since A′⊗AE is a free A′-module, DΦ′(1⊗x1,…,1⊗xn) is defined, and we have obviously
(6)DΦ′(1⊗x1,…,1⊗xn)=h(DΦ(x1,…,xn)).
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.
Nothing.
Nothing.