Let A be a commutative ring and M an A-module. For every integer n⩾0, the A-module the tensor product of n modules equal to M (also called the n-th tensor power of M) is denoted by ⨂nM, or M⊗n, or Tn(M), or TAn(M), or Tensn(M); then T1(M)=M; also we write T0(M)=A. The A-module the direct sum⨁n⩾0Tn(M) is denoted by T(M) or Tens(M). We shall define a graded A-algebra structure of type N on T(M), by defining for every ordered pair of integers p⩾0,q⩾0, an A-linear mapping
mpq:Tp(M)⊗ATq(M)→Tp+q(M)
(§3, no. 1, Remark). For p>0 and q>0, mpq is the associativity isomorphism (II, §3, no. 9) and, when p=0 (resp. q=0), m0,q is the canonical isomorphism of A⊗ATq(M) onto Tq(M) (resp. mp,0 is the canonical isomorphism of Tp(M)⊗AA onto Tp(M) (II, §3, no. 4, Proposition 4)). Then, for xi∈M, α∈A,
It is immediate that the multiplication thus defined on T(M) is associative and admits as unit element the unit element 1 of A=T0(M).
Definition 1
For every module M over a commutative ring A, the tensor algebra of M, denoted by T(M), or Tens(M), or TA(M), is the algebra ⨁n≥0Tn(M) with the multiplication defined in (1). The canonical injection ϕ:T1(M)→T(M) (II, § 1, no. 12) (also denoted by ϕM) is called the canonical injection of M into T(M).
Proposition 1
Let E be a (unital) A-algebra and f:M→E an A-linear mapping. There exists one and only one A-algebra homomorphism g:T(M)→E such that f=g∘ϕ.
In other words, (T(M),ϕ) is a solution of the universal mapping problem (Set Theory, IV, § 3, no. 1), where Σ is the species of A-algebra structure, the α-mappings being the A-linear mappings from the module M to an A-algebra. Observe that here there is no question of a graduation on T(M).
For every finite family (xi)1≤i≤n of n elements of M, by definition of the product in T(M), x1⊗x2⊗⋯⊗xn=ϕ(x1)ϕ(x2)…ϕ(xn); then necessarily g(x1⊗x2⊗⋯⊗xn)=f(x1)f(x2)…f(xn) for n≥1 and g(α)=αe (if e is the unit element of E) for α∈A, which proves the uniqueness of g. Conversely, note that, for all n>0, the mapping
(x1,…,xn)↦f(x1)f(x2)…f(xn)
of Mn into E is A-multilinear; hence there corresponds to it an A-linear mapping gn:Tn(M)→E such that
g(x1⊗x2⊗⋯⊗xn)=f(x1)f(x2)…f(xn).
Also we define the mapping g0:T0(M)→E as equal to ηE (§1, no. 3), in other words g0(α)=αe for α∈A. Let g be the unique A-linear mapping of T(M) into E whose restriction to Tn(M) is gn (n≥0); it is immediate that g∘ϕ=g1=f and it remains to verify that g is an A-algebra homomorphism. By construction g(1)=e and it suffices by linearity to show that g(uv)=g(u)g(v) for u∈Tp(M) and v∈Tq(M) (p>0,q>0); now it follows from formulae (1) and (2) that this relation is true when u=x1⊗x2⊗⋯⊗xp and v∈xp+1⊗⋯⊗xp+q (where the xi belong to E). It is therefore true for u∈Tp(M) and v∈Tq(M) by linearity.
Remark
Suppose that E is a graded A-algebra of type Z, with graduation (En), and suppose also that
f(M)⊂E1.
Then it follows from (2) that g(Tp(M))⊂Ep for all p≥0 and hence g is a graded algebra homomorphism.
2. Functorial properties of the tensor algebra
Proposition 2
Let A be a commutative ring, M and N two A-modules and
u:M→N
an A-linear mapping. There exists one and only one A-algebra homomorphism
u′:T(M)→T(N)
such that the diagram
M↓ϕMT(M)uu′N↓ϕNT(N)
is commutative. Further, u′ is a graded algebra homomorphism.
The existence and uniqueness of u′ follow from no. 1, Proposition 1, applied to the algebra T(N) and the linear mapping ϕN∘u:M→T(N); as
u(M)⊂T1(N)=N,
the fact that u′ is a graded algebra homomorphism follows from the Remark of no. 1.
The homomorphism u′ of Proposition 2 will henceforth be denoted by T(u). If P is an A-module and v:N→P an A-linear mapping, then
T(v∘u)=T(v)∘T(u)
for T(v)∘T(u) is an algebra homomorphism rendering commutative the diagram
M↓ϕMT(M)v∘uT(v)∘T(u)P↓ϕPT(P)
T(u) is sometimes called the canonical extension of u to T(M) (which contains M=T1(M)). Note that the restriction Tn(u):Tn(M)→Tn(N) is just the linear mapping u⊗n=u⊗u⊗⋯⊗u (n times), for
Tn(u)(x1⊗⋯⊗xn)=u(x1)⊗⋯⊗u(xn)
since T(u) is an algebra homomorphism and T1(u)=u; the restriction T0(u) to A is the identity mapping. Tn(u) is called the n-th tensor power of u.
Proposition 3
If u:M→N is a surjective A-linear mapping, the homomorphism T(u):T(M)→T(N) is surjective and its kernel is the two-sided ideal of T(M) generated by the kernel P⊂M⊂T(M) of u.
T0(u):T0(M)→T0(N) is bijective and for every integer n>0,
Tn(u):Tn(M)→Tn(N)
is surjective, as is seen by induction on n using II, § 3, no. 6, Proposition 6; the latter proposition also shows, by induction on n, that the kernel In of Tn(u) is the submodule of Tn(M) generated by the products
x1⊗x2⊗⋯⊗xn
where at least one of the xi belongs to P. This shows that the kernel I=⨁n≥1In of T(u) is the two-sided ideal generated by P in T(M).
If u:M→N is an injective linear mapping, it is not always true that T(u) is an injective mapping (Exercise 1). However, this is true when u is an injection such that u(M) is a direct factor of N, for then there exists a linear mapping v:N→M such that v∘u is the identity mapping on M and therefore
T(v∘u)=T(v)∘T(u)
is the identity mapping on T(M), hence T(u) is injective and its image (isomorphic to T(M)) is a direct factor of T(N) (II, § 1, no. 9, Proposition 15). More precisely:
Proposition 4
Let N and P be two submodules of an A-module M such that their sum N+P is a direct factor in M and their intersection N∩P is a direct factor in N and in P. Then the homomorphisms T(N)→T(M), T(P)→T(M) and*
T(N∩P)→T(M),
canonical extensions of the canonical injections, are injective; if T(N), T(P) and T(N∩P) are identified with subalgebras of T(M) by means of these homomorphisms, then (5)
T(N∩P)=T(N)∩T(P).
By hypothesis, there exist submodules N′⊂N and P′⊂P such that N=N′⊕(N∩P), P=P′⊕(N∩P); then
N+P=N′⊕P′⊕(N∩P)
and there exists by hypothesis a submodule M′ of M such that
M=M′⊕(N+P)=M′⊕N′⊕P′⊕(N∩P)=M′⊕P′⊕N=M′⊕N′⊕P.
In particular, N+P, N, P and N∩P are direct factors in M, which implies, as has been seen above, that the canonical homomorphisms
T(N+P)→T(M),T(N)→T(M),
T(P)→T(M),T(N∩P)→T(M)
are injective. The three algebras T(N), T(P) and T(N∩P) are thus identified with subalgebras of T(N+P) and the latter with a subalgebra of T(M);
writing Q=N∩P, it remains to show that, if T(Q), T(N′⊕Q) and T(P′⊕Q) are identified with subalgebras of T(N′⊕P′⊕Q), then
T(N′⊕Q)∩T(P′⊕Q)=T(Q).
Now, consider the commutative diagram
N′⊕Q↓Q⟶⟶N′⊕P′⊕Q↓P′⊕Q
where the horizontal arrows are the canonical injections and the vertical arrows the canonical projections. We derive a commutative diagram
T(N′⊕Q)↓rT(Q)⟶u⟶vT(N′⊕P′⊕Q)↓sT(P′⊕Q)
where r and s are surjective homomorphisms (Proposition 3) and u and v injective homomorphisms. Hence, to prove (6), note that the right hand side is obviously contained in the left; it therefore suffices to verify that if
x∈T(N′⊕Q)∩T(P′⊕Q),
then x∈T(Q). Now the definition of the homomorphism s shows that its restriction to T(P′⊕Q) (identified with a subalgebra of T(N′⊕P′⊕Q)) is the identity mapping; the hypothesis on x therefore implies that s(u(x))=x. But then also v(r(x))=x, in other words x belongs to the image of T(Q) in T(P′⊕Q), which was to be proved.
Remark
Note in particular that the hypotheses of Proposition 4 always hold for arbitrary submodules N,P of M when A is a field (II, § 7, no. 3, Proposition 4). Moreover, if N⊂P and N=P, then Tn(N)=Tn(P) for all n≥1, since if R is a complement of P in N, then Tn(P)∩Tn(R)={0} by (4) and Tn(R)={0}.
Corollary
Let K be a commutative field and M a vector space over K. For every element z∈T(M), there exists a smallest vector space N of M such that z∈T(N) and N is of finite rank over K.
It is understood in this statement that for every vector subspace P of M, T(P) is canonically identified with a subalgebra of T(M). Let z∈T(M); z can be expressed as a linear combination of elements each of which is a finite product of elements of M=T1(M); all the elements of M which occur in these products generate a vector subspace Q of finite rank and z∈T(Q).
Let F be the (non-empty) set of vector subspaces P of finite rank such that z∈T(P). Every decreasing sequence of elements of F is stationary, since they are vector spaces of finite rank. Hence F has a minimal element N (Set Theory, III, § 6, no. 5). It remains to verify that every P∈F contains N; now, z∈T(P)∩T(N)=T(P∩N) (Proposition 4); in view of the definition of N, this implies N∩P=N, that is P⊃N.
The subspace N of M is said to be associated with z.
3. Extension of the ring of scalars
Let A,A′ be two commutative rings and ρ:A→A′ a ring homomorphism. Let M be an A-module, M′ an A′-module and u:M→M′ an A-homomorphism; as the canonical injection ϕM′:M′→TA′(M′) is also an A-homomorphism (by restriction of scalars), so is the composition MuM′ϕM′TA′(M′). An A-algebra homomorphism TA(M)→ρ∗(TA′(M′)) is derived (no. 2), also denoted by T(u):TA(M)→TA′(M′), which is the unique A-homomorphism rendering commutative the diagram
M↓ϕMTA(M)uT(u)M′↓ϕM′TA′(M′)
If σ:A′→A′′ is a commutative ring homomorphism, M′′ an A′′-module and v:M′→M′′ an A′-homomorphism, the above uniqueness property shows that
T(v∘u)=T(v)∘T(u).
Proposition 5
Let A,B be two commutative rings, ρ:A→B a ring homomorphism and M an A-module. The canonical extension
ψ:TB(B⊗AM)→B⊗ATA(M)
of the B-linear mappinglB⊗ϕM:B⊗AM→B⊗ATA(M)is an isomorphism of graded B-algebras.
Consider the two A-algebra homomorphisms: the canonical injection j:B=T0(B⊗AM)→T(B⊗AM) and the homomorphism
h=T(i):T(M)→T(B⊗AM)
derived (cf. formula (8)) from the canonical A-linear mapping
i:M→B⊗AM.
As T0(B⊗AM) is contained in the centre of T(B⊗AM), Proposition 5 of § 4, no. 2 can be applied and an A-algebra homomorphism
which shows immediately that ψ′ is also a B-algebra homomorphism. It suffices to prove that ψ∘ψ′ and ψ′∘ψ are the identity mappings on B⊗AT(M) and T(B⊗AM) respectively. Now, these two algebras are generated by B⊗AM and clearly ψ∘ψ′ and ψ′∘ψ coincide with the identity mapping on B⊗AM, whence the conclusion.
4. Direct limit of tensor algebras
Let (Aα,ϕβα) be a directed direct system of commutative rings and (Mα,fβα) a direct system of Aα-modules (II, § 6, no. 2); let A=lim→Aα and M=lim→Mα, which is an A-module. For α≤β an Aα-algebra homomorphism (no. 3, formula (8)) fβα′=T(fβα):TAα(Mα)→TAβ(Mβ) is derived canonically from the Aα-homomorphism fβα:Mα→Mβ and it follows from (9) (no. 3) that (TAα(Mα),fβα′) is a direct system of Aα-algebras. On the other hand let fα′:Mα→M be the canonical Aα-homomorphism; an Aα-algebra homomorphism fα′:TAα(Mα)→TA(M) is derived (no. 3, formula (8)) and it also follows from (9) (no. 3) that the fα′ constitute a direct system of Aα-homomorphisms.
Proposition 6
The A-homomorphism* f′=lim→fα′:lim→TAα(Mα)→TA(M)is a graded algebra isomorphism.
To simplify we write E=TA(M) and E′=lim→TAα(Mα) and let
gα:TAα(Mα)→E′
be the canonical Aα-homomorphism. Clearly the composite Aα-linear mappings MαϕMαTAα(Mα)gαE′ form a direct system and there is therefore one and only one A-linear mapping u=lim→(gα∘ϕMα):M→E′ such that
u∘fα=gα∘ϕMα
for all α. This mapping itself factorizes uniquely (no. 1, Proposition 1) into MϕME→E′, where h is an A-algebra homomorphism. It will suffice to prove that h∘f′=1E′ and f′∘h=1E.
To this end note that, for every index α, (no. 3, formula (8))
h∘fα′∘ϕMα=h∘ϕM∘fα=u∘fα=gα∘ϕMα
whence, by the uniqueness assertion of no. 1, Proposition 1,
h∘fα′=gα
for all α; it follows that (h∘f′)∘gα=gα for all α and hence h∘f′=1E′ by definition of a direct limit.
On the other hand, by virtue of no. 3, formula (8),
f′∘u∘fα=f′∘gα∘ϕMα=fα′∘ϕMα=ϕM∘fα,
whence again f′∘u=ϕM by definition of a direct limit; we conclude that f′∘h∘ϕM=ϕM and the uniqueness property of no. 1, Proposition 1 gives f′∘h=1E.
Proposition 6 can also be shown by observing that, for every integer n≥1, there is a canonical A-module isomorphism lim→TAαn(Mα)→TAn(M), as follows by induction on n from II, § 6, no. 3, Proposition 7. It is immediately verified that these isomorphisms are the restrictions of f′.
5. Tensor algebra of a direct sum. Tensor algebra of a free module. Tensor algebra of a graded module
Let A be a commutative ring and M=⨁λ∈LMλ the direct sum of a family of A-modules. It follows from II, § 3, no. 7, Proposition 7, by induction on n, that Tn(M) is the direct sum of the submodules the images of the canonical injections
Mλ1⊗Mλ2⊗⋯⊗Mλn→Tn(M)=M⊗n
relative to all the sequences (λi)∈Ln. Identifying Mλ1⊗Mλ2⊗⋯⊗Mλn with this image, it is seen that T(M) is the direct sum of all the modules
Mλ1⊗Mλ2⊗⋯⊗Mλn
where n runs through N and, for each n, (λi) runs through Ln.
We first deduce the following consequence:
Theorem 1
Let A be a commutative ring, M a free A-module and (eλ)λ∈L a basis of M. Then the elements es=eλ1⊗eλ2⊗⋯⊗eλn, where s=(λ1,…,λn) runs through the set of all finite sequences of elements of L and e∅ is used to denote the unit element of T(M), form a basis of the A-module T(M).
The elements of this basis are obviously homogeneous and the multiplication table is given by
eset=est
where st denotes the sequence of elements of L obtained by juxtaposing the sequences s and t (I, § 7, no. 2).
It is seen that the basis (es) of T(M), with the multiplicative law (10), is canonically isomorphic to the free monoid of the set L (I, § 7, no. 2), the isomorphism being obtained by mapping each word s of this monoid to the element es. It follows (§ 2, no. 7) that the tensor algebra T(M) of a free module M, with a basis of indexing set L, is canonically isomorphic to the free associative algebra of L over A. In particular (§ 2, no. 7, Proposition 7), for every mapping f:L→E from L to an A-algebra E, there exists one and only one A-algebra homomorphism fˉ:T(M)→E such that fˉ(eλ)=f(λ).
Remark
The above results can equally be obtained as a consequence of the universal properties of the free associative algebra of the tensor algebra, using II, § 3, no. 7, Corollary 2 to Proposition 7.
Corollary
If M is a projective A-module, T(M) is a projective A-module.
M is a direct factor of a free A-module N (II, § 2, no. 2, Proposition 4) and hence T(M) is a direct factor of T(N) (no. 2); as T(N) is free (Theorem 1), this shows that T(M) is projective (II, § 2, no. 2).
Proposition 7
Let Δ be a commutative monoid, M a graded A-module of type Δ and (Mα)α∈Δ its graduation. For every ordered pair (α,n)∈Δ×N, let Tα,n(M) be the (direct) sum of the submodules Mα1⊗Mα2⊗⋯⊗Mαn of Tn(M) such that ∑i=1nαi=α; then (Tα,n(M))(α,n)∈Δ×N is the only graduation of type Δ×N compatible with the algebra structure on T(M) and inducing on M=T1(M) the given gradation.
It has been seen at the beginning of this no. that T(M) is the direct sum of the Tα,n(M) and the fact that this is a graduation compatible with the algebra structure follows immediately from the definitions. If (T′α,n) is another graduation of type Δ×N on T(M) compatible with the algebra structure and such that T′α,1(M)=T′α,1 for α∈Δ, it follows immediately from the definitions that, for all n≥1 and all α∈Δ, Tα,n(M)⊂T′α,n; but since T(M) is also the direct sum of the Tα,n(M), this implies that T′α,n=Tα,n(M) (II, § 1, no. 8, Remark).
6. Tensors and tensor notation
Let A be a commutative ring, M an A-module, M∗ the dual of M (II, § 2, no. 3) and I and J two disjoint finite sets; the A-module ⨂i∈I∪JEi, where Ei=M if i∈I, Ei=M∗ if i∈J, is denoted by TJI(M); the elements of TJI(M) are called tensors of type(I,J) over M. They are called contravariant if J=∅, covariant if I=∅ and mixed otherwise.
Let I′,I′′ be two subsets of I and J′,J′′ two subsets of J such that I′∪I′′=I,
I′∩I′′=∅,J′∪J′′=J,J′∩J′′=∅; then there is a canonical associativity isomorphism (II, § 3, no. 9)
TJI(M)→TJ′I′(M)⊗ATJ′′I′′(M).
Considering the tensor algebra T(M⊕M∗), it follows from no. 5 that Tn(M⊕M∗) is canonically identified with the direct sum of the TJI(M) where I runs through the set of subsets of the interval {1,n} of N and J is the complement of I in {1,n}.
When I={1,p} and J={p+1,p+q} with integers p≥0,q≥0 (where we replace I (resp. J) by ∅ when p=0 (resp. q=0)), the A-module TJI(M) is also denoted by Tqp(M); the A-modules T0n(M) and Tn0(M) are therefore by definition the A-modules Tn(M) and Tn(M∗) respectively. When I and J are arbitrary finite sets of cardinals p=Card(I) and q=Card(J), we give each of them a total ordering; then there exists an increasing bijection of I (resp. J) onto {1,p} (resp. {p+1,p+q}) and these bijections therefore define an isomorphism
TJI(M)→Tqp(M).
When M is a finitely generated projective A-module, it follows from II, § 2, no. 7, Corollary 4 to Proposition 13 and II, § 4, no. 4, Corollary 1 to Proposition 4 that there is a canonical isomorphism
(TJI(M))∗→TIJ(M).
Suppose now that M is a finitely generated free A-module and let (eλ)λ∈L be a basis of M (L therefore being a finite set). The basis of M∗dual to (eλ) (II, § 2, no. 6) is denoted by (eλ)λ∈L. The bases (eλ) and (eλ) of M and M∗ respectively define (no. 5) a basis of TJI(M), which we give explicitly as follows: given two mappings f:I→L and g:J→L, let efg be the element ⨂i∈I∪Jxi of TJI(M) defined by
xi=ef(i)if i∈I,xi=eg(i)if i∈J.
When (f,g) runs through the set of ordered pairs of mappings f:I→L and g:J→L, the efg form a basis of the A-module TJI(M), said to be associated with the given basis (eλ) of M. For z∈TJI(M), we can therefore write
z=(f,g)∑αgf(z)⋅efg
where the αgf are the coordinate forms relative to the basis (efg); by an abuse of language, the αgf(z) are called the coordinates of the tensor zwith respect to the basis(eλ) of the module M. The αgf constitute the dual basis of (efg), in other words they are identified with the elements of the basis of TIJ(M), associated with (eλ). When I and J are complementary subsets of an interval {1,n} of N, αgf (or αgf(z)) is denoted as follows: each element f(i) for i∈I is written as an upper index in the i-th place with a dot in the i-th place for i∈J; similarly, g(i) for i∈J is written as a lower index in the i-th place with a dot in the i-th place for i∈I. For example, for I={1,4},J={2,3}, we write αvρλ⋅⋅μ: if f(1)=λ,f(4)=μ,g(2)=ν,g(3)=ρ.
Let (e~λ)λ∈L be another basis of M and P the matrix of passing from (eλ) to (e~λ) (II, § 10, no. 8). Then the matrix of passing from (eλ) to (e~λ) (dual basis of (e~λ)) is the contragredient tP−1 of P (II, § 10, no. 8, Proposition 5). It follows (II, § 10, no. 10) that the matrix of passing from the basis (efg) of TJI(M) to the basis (e~fg) (where f (resp. g) runs through the set of mappings of I into L (resp. of J into L)) is the matrix
i∈I∪J⨂Qi,where Qi=P if i∈I,Qi=tP−1 if i∈J.
The transpose of this matrix is therefore identified with
i∈I∪J⨂Ri,where Ri=tP−1 if i∈I,Ri=P if i∈J.
Suppose now that M is an arbitrary module. Let i∈I,j∈J and write I′=I−{i},J′=J−{j}; we shall define a canonical A-linear mapping
cji:TJI(M)→TJ′I′(M),
called contraction of the index i and the index j. For this, note that the mapping of MI×(M∗)J, which associates with every family (xi)i∈I∪J, where xi∈M if i∈I and xi∈M∗ if i∈J, the element
⟨xi,xj⟩k∈(I∪J)−{i,j}xk
of TJ′I′(M), is A-multilinear; cji is the corresponding A-linear mapping.
Suppose now that M is free and finitely generated and let (eλ)λ∈L be a basis of M. Given two mappings f:I→L,g:J→L, let fi denote the restriction of f to I′=I−{i} and gj the restriction of g to J′=J−{j}; then by virtue of (12)
cji(efg)={0efigjif f(i)=g(j)if f(i)=g(j)
The expression for the coordinates of cji(z) as a function of those of z is obtained; for every mapping f′ (resp. g′) of I' into L (resp. of J' into L) and every λ∈L, let (f′,λ) (resp. (g′,λ)) denote the mapping of I into L (resp. of J into L) whose restriction to I' (resp. J') is f′ (resp. g′) and which takes the value λ at the element i (resp. j). Then, if the coordinate forms relative to the basis (ef′g′) of TJ′I′(M) are denoted by βg′f′,
βg′f′(cji(z))=λ∈L∑α(g′;λ)(f′;λ)(z).
Examples of tensors. (1) Let M be a finitely generated projective A-module. We know (II, § 4, no. 2, Corollary to Proposition 2), that there is a canonical A-module isomorphism
θM:M∗⊗AM→EndA(M)
such that θM(x∗⊗x) (for x∈M,x∗∈M∗) is the endomorphism
y↦⟨y,x∗⟩x.
Hence, by means of θM, T{1}(2)(M) (isomorphic to T11(M)) can be identified with the A-module EndA(M). Suppose that M is a free module and let (eλ)λ∈L be a basis of M; then the coordinates of a tensor z∈M∗⊗M relative to the basis (eμ⊗eλ) of this module are denoted by ζμ⋅λ. As θM(eμ⊗eλ) is the endomorphism y↦⟨y,eμ⟩eλ, the endomorphism u=θM(z)=θM(∑λ,μζμ⋅λeμ⊗eλ) maps y to ∑λ,μζμ⋅λ⟨y,eμ⟩eλ; writing y=eλ, we obtain the relation
u(eλ)=μ∈L∑ζλ⋅μeμ
in other words, the matrix of the linear mappingu=θM(z)is that whose element in the row of indexμand column of indexλisζλ⋅μ.
The definition of the trace of u (II, § 4, no. 3) shows immediately that
Tr(θM(z))=c12(z).
Therefore the element z0=∑λ∈Leλ⊗eλ (whose coordinates ζμ⋅λ are zero for λ=μ and equal to 1 for λ=μ), which is such that θM(z0)=1M, is the image of the element 1∈A=T00(M) under the mapping the transpose of the contraction
c12:T{1}(2)(M)→A.
(2) Suppose always that M is a finitely generated projective A-module; there is a canonical A-module isomorphism
μ:M∗⊗AM∗→(M⊗AM)∗
(II, § 4, no. 4, Corollary 1 to Proposition 2) and a canonical isomorphism
θ:(M⊗AM)∗⊗AM→HomA(M⊗AM,M)
(II, § 4, no. 2, Corollary to Proposition 2); also HomA(M⊗AM,M) is canonically isomorphic to the A-module L2(M,M;M) of A-*bilinear* mappings of M×M into M (II, § 3, no. 9). Composing these isomorphisms, a canonical isomorphism is obtained
χM:T{1,2}(3)(M)=M∗⊗M∗⊗M→L2(M,M;M)
such that, for x∗,y∗ in M∗, z∈M, χM(x∗⊗y∗⊗z) is the bilinear mapping
(u,v)↦⟨u,x∗⟩⟨v,y∗⟩z.
Hence, by means of χM, T{1,2}(3)(M) (isomorphic to T21(M)) can be identified with the A-module L2(M,M;M). Suppose that M is a free A-module and let (eλ)λ∈L be a basis of M; then the coordinates of a tensor z∈M∗⊗M∗⊗M relative to the basis (eλ⊗eμ⊗eν) of this module are denoted by ζλμν. The bilinear mapping χM(z) maps the ordered pair (eλ,eμ) to
ν∈L∑ζλμνeν
and therefore the ζλμν are just the constants of structure of the (in general non-associative) algebra defined on M by the bilinear mapping χM(z), with respect to the basis (eλ) (§ 1, no. 7).
Remark 2
Let (eλ)λ∈L, (eˉλ)λ∈L be two bases of M and P the matrix of passage from (eλ) to (eˉλ). On account of what was seen in Example 1, the element of P appearing in the row of index λ and the column of index μ is denoted by αμλ and the element of the contragredient tP−1 appearing in the row of index λ and the column of index μ is denoted by βλμ. Then (in the notation introduced above)
for all mappings f′:I→L and g′:J→L. The coordinates ζgf of a tensor z∈TJI(M) with respect to the basis (eλ) can therefore be expressed in terms of the coordinates ζˉg′f′ of z with respect to the basis (eˉλ) using the formulae
The matrix P−1 of passage from the basis (eˉλ) to the basis (eλ) is the transpose of tP−1, so that βλμ is the element which appears in the column of index λ and the row of index μ of P−1. The calculation of efg in terms of the eˉf′g′ and that of the ζˉg′f′ in terms of the ζgf are therefore made by replacing αμλ by βλμ and βλμ by αμλ in the above calculations and exchanging the roles of f and f′ and those of g and g′. Then
The above formulae are such that the summation is over indices which appear once as a lower index and once as an upper index. Certain authors allow themselves because of this circumstance to suppress the summation signs.