Throughout this paragraph, K will denote a commutative ring and A a unital associative K-algebra. Every A-module will be assumed to be given the K-module structure obtained by restricting the scalars to K.
1. Norms and traces relative to a module
Definition 1
Let M be an A-module admitting a finite basis as a K-module. For all a∈A, let aM be the endomorphism x↦ax of the K-module M. The trace, determinant and characteristic polynomial of aM are called respectively the trace, norm and characteristic polynomial of z relative to M.
The trace and norm of a are therefore elements of K, denoted respectively by TrM/K(a) and NM/K(a); the characteristic polynomial of a is an element of K[X], denoted by PcM/K(a;X). We omit K in the above notation when there is no risk of confusion.
From the properties of the trace and determinant of an endomorphism (II, § 4, no. 3 and § 8, no. 1) we obtain the relations
Let (ei)1≤i≤n be a basis of the K-module M and (mij(a)) the matrix of the endomorphism aM with respect to this basis. The functions mij are linear forms on the K-module A and
It follows from the method of calculating a determinant (§ 8, no. 11, formula (50)) that
(7) PcM(a;X)=Xn+c1Xn−1+⋯+cn
where
(8) c1=−TrM(a),cn=(−1)nNM(a).
For λ∈K,
(9) TrM(λ)=n.λ,NM(λ)=λn,PcM(λ;X)=(X−λ)n.
Let K′ be a commutative K-algebra. Let M′=K′⊗KM and A′=K′⊗KA, so that M′ has an A′-module structure (§ 4, Example 2). As a K′-module M′ admits the basis consisting of the 1⊗ei (1≤i≤n) and the matrix of aM with respect to (ei) is equal to the matrix of (1⊗a)M′ with respect to (ei′). Then
for if f is an isomorphism of M onto M′, the matrix of aM with respect to a basis B of M over K is the same as the matrix of aM′ with respect to the basis f(B) of M′.
Proposition 1
Let M=M0⊃M1⊃…⊃Mr={0} be a decreasing sequence of submodules of an A-module M such that each of the K-modules Pi=Mi−1/Mi (1≤i≤r) admits a finite basis. Then the K-module M admits a finite basis and
TrM(a)=i=1∑rTrPi(a),NM(a)=i=1∏rNPi(a)
(15)
PCM(a;X)=i=1∏rPCPi(a;X).
Let Bi′ be a basis of Pi over K; then a system of representatives Bi of Bi′ (mod. Mi) is a basis of a supplementary submodule of the K-module Mi in the K-module Mi−1 (II, § 1, no. 11, Proposition 21). The union B of the Bi (1≤i≤r) is a basis of M over K. Let Xtt be the matrix of the endomorphism aPi with respect to the basis Bi′. It is immediate that the matrix of aM with respect to B is of the form
and the proposition follows from formulae (4), (5) and (6) of no. 1 and formula (31) of § 8, no. 6.
Proposition 2
Let A,A′ be two K-algebras, M an A-module and M′ an A′-module. Suppose that M and M′ are free K-modules of respective dimensions n and n′ and consider M⊗KM′ as an (A⊗KA′)-module (§4, no. 3, Example 2). Then, for a∈A and a′∈A′,
TrM⊗M′(a⊗a′)=TrM(a)TrM′(a′)
(16)
NM⊗M′(a⊗a′)=(NM(a))n′(NM′(a′))n.
(17)
Formula (16) follows from II, § 4, no. 4, formula (26) and formula (17) from § 8, no. 6, formula (33).
3. Norm and trace in an algebra
Definition 2
Let A be a K-algebra which is a finite-dimensional free K-module. For every element a∈A, the trace (resp. norm,\footnote{† This notion should not be confused with the notion for norm in an algebra over a valued field (\emph{General Topology}), IX, § 3, no. 7.} resp. characteristic polynomial) of a relative to A and K is the trace (resp. determinant, resp. characteristic polynomial) of the endomorphism x↦ax of the K-module A.
The trace, norm and characteristic polynomial of a∈A relative to A and K are denoted by TrA/K(a), NA/K(a) and PcA/K(a;X); we omit K and even A from this notation when there is no risk of confusion. Note that the trace (resp. norm, characteristic polynomial) of a∈A is just the trace (resp. norm, characteristic polynomial) of a relative to the A-module As.
Suppose that A is the product A1×A2×⋯×Am of a finite number finite-dimensional algebras over K which are free K-modules. Using the above remark and Proposition 1 of no. 2, we have, for every element
Similarly, Proposition 2 of no. 2 shows that if A and A′ are two algebras, which are free K-modules, of finite dimensions n,n′ respectively over K, then, for a∈A,a′∈A′.
TrA⊗A′(a⊗a′)=TrA(a)TrA′(a′)
(19)
NA⊗A′(a⊗a′)=(NA(a))n′(NA′(a′))n.
(20)
Finally let A be a finite-dimensional algebra over K which is a free K-module, h a homomorphism of K into a commutative ring K′ and A′=A(K′) the K′-algebra derived from A by extending the scalars by means of h. It follows from formula (12) of no. 1 that, for all a∈A,
where hˉ is the homomorphism K[X]→K′[X] derived from h. In particular, for K′=K[X], we obtain, writing A[X]=A⊗KK[X],
PcA/K(a;X)=NA[X]/K[X](X−a).
(22)
More generally, if K′ is a commutative K-algebra and A′=A⊗KK′, then, for all x∈A′,
PcA/K(a;x)=NA′/K′(x−a).
Example
(1) Let A be a quadratic algebra over K of type (α,β) and (e1,e2) a basis of type (α,β) (§ 2, no. 3). For x=ξe1+ηe2, TrA/K(x)=2ξ+βη and NA/K(x)=ξ2+βξη−αη2; these functions are therefore identical with the Cayley trace and norm of x (§ 2, no. 24).
(2) Let A be a quaternion algebra over K. A direct calculation allows us to verify that TrA/K(x)=2T(x) and NA/K(x)=(N(x))2, where T and N are the Cayley trace and norm (§ 2, no. 4).
(3) Let A=Mn(K) and let the canonical basis (Eij) of A (II, § 10, no. 3) be arranged in lexicographic order. Then it is immediately seen that for every matrix X=∑i,jξijEij the matrix (of order n2) of the endomorphism Y↦XY is of the form
X0…00X…0…………00…X
whence TrA/K(X)=n⋅Tr(X) and NA/K(X)=(det(X))n.
4. Properties of norms and traces in an algebra
Proposition 3
Let A be a K-algebra admitting a finite basis. For an element a∈A to be invertible, it is necessary and sufficient that its NA/K(a) be invertible in K.
If a admits an inverse a′ in A, then
NA/K(a)NA/K(a′)=NA/K(aa′)=NA/K(1)=1
by formula (3) of no. 1. Conversely, if NA/K(a) is invertible, the endomorphism h:x↦ax is bijective (§ 8, no. 2, Theorem 1). Then there exists a′∈A such that aa′=1; then h(a′a−1)=aa′a−a=(aa′−1)a=0, whence aa′=1 since h is injective. Hence a′ is the inverse of a.
Proposition 4
Let A be a K-algebra admitting a finite basis. For all a∈A, PcA/K(a;a)=0.
This follows immediately from the Cayley-Hamilton theorem (§ 8, no. 11, Proposition 20).
Proposition 5
Let A be a K-algebra and m a two-sided ideal of A. Suppose that A0=A/m is a free K-module of finite dimension n, that there exists an integer r>0 such that mr={0} and that mi−1/mi is a free A0-module of finite dimension si for 1≤i≤r. Let s=s1+⋯+sr and for all a∈A let a0 denote the class of a mod. m. Then A is a free K-module of dimension ns and, for all a∈A,
TrA(a)=s⋅TrA0(a0),NA(a)=(NA0(a0))s
PcA(a;X)=(PcA0(a0;X))s.
By virtue of II, § 1, no. 13, Proposition 25, mi−1/mi is a free K-module of dimension nsi. Hence Proposition 1 of no. 2 can be applied with Pi=mi−1/mi;
this shows in the first place that A is a free K-module of dimension n(s1+⋯+sr)=ns. Moreover, the hypothesis implies that the A-module Pi is isomorphic to a direct sum of si submodules isomorphic to the A-module A0; by Proposition 1 of no. 2, therefore NPi(a)=NA0(a)si; finally therefore
NA(a)=NA0(a)s.
In this formula NA0(a) is defined by considering A0 as a left A-module and it is equal to the determinant of the K-linear mapping x↦ax of A0 into itself; but, as ax=a0x for x∈A0, NA0(a)=NA0(a0), which completes the proof of formula (23) for the norm. The two others are shown analogously.
Proposition 6
Let A be a commutative K-algebra admitting a finite basis over K and V an A-module admitting a finite basis over A. Then V admits a finite basis over K and for every A-endomorphism u of V, if uK is the mapping u considered as a K-endomorphism of V,*
Let (ai)1≤i≤m be a basis of A over K and (ej)1≤j≤n a basis of V over A; then (aiej) is a basis of V over K (II, § 1, no. 13, Proposition 25). On the other hand the third of formulae (24) can be deduced from the second applied to the endomorphism X−uˉ of the A[X]-module A[X]⊗AV (§ 8, no. 10). It will therefore suffice to show the first two formulae in (24). We shall first establish the following lemma:
*Lemma 1. Let Xij (1≤i≤n,1≤j≤n) be n2 indeterminates, X the square matrix (Xij) of order n and D(X11,…,Xnn)∈Z[X11,…,Xnn] the determinant det(X). On the other hand let A be a commutative ring, Mij (1≤i≤n,1≤j≤n) n2 matrices of order m over A, which are pairwise permutable, and M the square matrix of order mn over A which can be expressed as a square matrix of matrices (II, § 10, no. 7)
Then the determinant of M is equal to the determinant of the square matrix
D(M11,…,Mnn)
of order m.
We proceed by induction on n, the cases n=0 and n=1 being trivial.
Let Z be a new indeterminate and Nij the matrix Mij+δijZIm (δij the Kronecker index). If Dij(X11,…,Xnn) is the cofactor of Xij in the matrix X (§ 8, no. 6), then
(25)XjiDki(X11,…,Xnn)=δjkD(X11,…,Xnn)
(§ 8, no. 6, formula (28)). We write Nij′=Dij(N11,…,Nnn), which is a square matrix of order m over A[Z] and consider the product N.U, where
Performing this product in blocks (II, § 10, no. 5) and using formulae (25), we obtain
N.U=P0⋯0N12N22⋯Nn2⋯⋯⋯⋯N1nN2n⋯Nnn
where we have written P=D(N11,…,Nnn). Let
Q=N22⋯Nn2⋯⋯⋯N2n⋯Nnn
which is a matrix of order m(n−1); then (§ 8, no. 6, formula (31)) (det N) (det U) = (det P)(det Q) and det U=detN11′. But by the induction hypothesis, det Q=det(D11(N11,…,Nnn))=detN11′ and by virtue of the definition of the Nij, clearly det Q is a polynomial in A[Z] of degree m(n−1), whose term in Zm(n−1) has coefficient 1 ; it follows immediately that det Q is not a divisor of zero in the graded algebra A[Z]. We therefore conclude that det N=det(D(N11,…,Nnn)) in A[Z]; if we substitute 0 for Z in these polynomials, then det M=det(D(M11,…,Mnn)).
Having shown this lemma, the K-module V is the direct sum of the K-modules Aej (1≤j≤n); we write u(ej)=∑k=1ncjkek. For every element xej∈Aej, with x∈A, the component of u(xej) in Aek is xcjkek; it follows that the matrix of uk with respect to the basis (aiej) of the K-module V can be expressed in the form of a square matrix of matrices (Mjk), where Mjk is the matrix of the K-linear mapping xej↦xcjkek of Aej into Aek with respect to the bases (aiej)1≤i≤m and (aiek)1≤i≤m of these two K-modules (II, § 10, no. 5). If for all t∈A, M(t) denotes the matrix, with respect to the basis (ai)1≤i≤m of A over K, of the endomorphism x↦xt of A, then Mjk=M(cjk); as t↦M(t) is a ring homomorphism the matrices Mjk are permutable with one another. Then
detuK=det(D(M11,…,Mnn))
by Lemma 1. But as t↦M(t) is a ring homomorphism, D(M11,…,Mnn) is the matrix of the K-endomorphism x↦x.det(cjk) of A with respect to basis (ai); by definition its determinant is therefore NA/K(det(u)), which proves the second formula of (24). On the other hand,
Let A be a commutative K-algebra admitting a finite basis over K and B an A-algebra admitting a finite basis over A. Then B admits a finite basis over K, and, for all b∈B ("transitivity formulae")*
(26) This follows immediately from Proposition 6, setting V=B and u(x)=bx.
Remark
Suppose that the homomorphism λ↦λ.1 of K into A is injective and let K be identified with its image in A; suppose that A admits a finite basis (ei)1≤i≤n as a K-module. Let s be an automorphism of A such that s(K)=K. Let a be an element of A; then, by transporting the structure
with vji=D(λji)+∑k=1n(μjkλki−λjkμki). As ∑i,k(μikλki−λikμki)=0, therefore TrA/K(D(a))=∑i=1nD(λii), in other words
TrA/K(D(a))=D(TrA/K(a)).∗
5. Discriminant of an algebra
Definition 3
Let A be a K-algebra admitting a finite basis of n elements. The discriminant of a sequence (x1,…,xn) of n elements of A, with respect to K, denoted by DA/K(x1,…,xn), is the discriminant of the square matrix*
(TrA/K(xixj))1≤i≤n,1≤j≤n.
Consider first a basis (ei)1≤i≤n of A over K and write
eiej=k=1∑ncijkekwithcijk∈K.
Then TrA/K(ei)=∑s=1nciss, whence TrA/K(eiej)=∑k,scijkckss and therefore
Now let (xi)1≤i≤n,(xi′)1≤i≤n be two sequences of n elements of A and suppose that there exists a square matrix of order n, M=(mij), with coefficients in K, such that xi=∑j=1nmijxj′ for 1≤i≤n. We write
Then TrA/K(xixj)=∑p,qmipmjqTrA/K(xp′xq′), whence T=M⋅T′⋅tM; the rule of multiplication of determinants therefore gives
detT=detM⋅detT′⋅dettM=(detM)2detT′
whence finally
DA/K(x1,…,xn)=(detM)2DA/K(x1′,…,xn′).
The above formula shows in particular that the discriminants of two bases of A over K differ by the square of an invertible element of K and therefore generate the same (principal) ideal of K. This ideal ΔA/K is called the discriminant ideal of A over K; by formula (32) the discriminant of every sequence of n elements of A which differ only in the order of the terms have the same discriminant, for the determinant of a permutation matrix is equal to ±1.
Example
(1) If A is a quadratic algebra of type (α,β) over K, then (in the notation of § 2, no. 3) Tr(e1)=2,Tr(e2)=β,
Tr(e22)=αTr(e1)+βTr(e2)=2α+β2,
whence DA/K(e1,e2)=β2+4α.
(2) Let A=K[X]/K[X]P, where P(X)=X3+pX+q, so that if x is the image of X in A, 1, x, x2 form a basis of A over K and x3=−px−q. It is seen immediately that Tr(1)=3,Tr(x)=0,Tr(x2)=−2p, taking account of the relation x3=−px−q,Tr(x3)=−3q and Tr(x4)=2p2, whence easily DA/K(1,x,x2)=−4p3−27q2.
(3) Let A be a quaternion algebra of type (α,β,γ) over K and (1,i,j,k) a basis of A of type (α,β,γ); taking account of § 3, no. 5, formula (30), it is easily found that Tr(1)=4,Tr(i)=2β,Tr(j)=Tr(k)=0, then
DA/K(1,i,j,k)=−16γ2(β2+4α)2.
(4) Let A=Mn(K) and consider the canonical basis (Eij)1≤i≤n,1≤j≤n of A over K (II, § 10, no. 3). It is immediate that TrA/K(Eij)=0 if j=i and TrA/K(Eii)=n for all i; it follows without difficulty that the matrix (Tr(EijEhk)) of order n2 is of the form n.P, where P is a permutation matrix, whence DA/K((Eij))=±nn2.