3.19.1 Tensor Covector Algebraic Interpretation

Tensor covector algebraic interpretation explores how covectors operate within tensor algebra, defining their role in linear functionals and dual spaces.

Tensor Covector Algebraic Interpretation is the viewpoint that treats a covector purely as an algebraic object, an element of the dual vector space V* equipped with its own vector space structure, its own basis theory, its own subspace lattice, and its own algebra of linear operations, independent of any geometric or analytic imagery. This interpretation grounds every later geometric reading of covectors in a precise algebraic foundation built from linear functionals, matrix representations, kernels, annihilators, and the functorial behavior of duality under linear maps.

The Dual Space as a Vector Space

Vector Space Structure of V*

The set V* of all linear functionals from V to the base field F is itself a vector space under pointwise addition and scalar multiplication.

(α + β) (v) = α (v) + β (v), (c α) (v) = c (α (v))

Verifying the vector space axioms for V* is immediate because they reduce to the corresponding axioms for scalars in F. This algebraic closure is what allows one to speak of linear combinations of covectors, of subspaces of V*, and of bases of V* on exactly the same footing as for V.

Dimension and Finite-Dimensional Duality

For finite-dimensional V, the dual space V* has the same dimension as V, and choosing a basis e_1, ..., e_n of V produces a dual basis e^1, ..., e^n of V* satisfying the Kronecker delta pairing relation. Consequently V and V* are isomorphic as vector spaces, though not canonically: any such isomorphism depends on the choice of basis, whereas the double dual V** is canonically isomorphic to V through the evaluation map, independent of any basis choice.

Matrix Representation

Covectors as Row Vectors

Fixing a basis of V identifies vectors with column vectors in F^n. Under this identification, a linear functional alpha is represented by a 1 by n row matrix of its coefficients, and evaluation becomes matrix multiplication.

α (v) = [α_{1} \dots α_{n}] [\begin{matrix} v^{1} \\ ⋮ \\ v^{n} \end{matrix}]

This algebraic packaging makes covector algebra computationally identical to row-vector algebra, and it clarifies that the abstract dual space, once a basis is fixed, is nothing more than the space of 1 by n matrices over F.

Composition of Linear Maps and Pullback as Matrix Transpose

If a linear map f from V to W is represented by an m by n matrix A relative to chosen bases, the pullback map f* from W* to V*, defined by f*(beta) = beta composed with f, is represented by the transpose matrix A^T. This is a purely algebraic fact: transposition of matrices is the coordinate shadow of the abstract pullback operation on dual spaces, and it explains why pullback reverses the direction of the original map, since transposition of an m by n matrix produces an n by m matrix acting in the reverse direction.

Kernels, Rank, and Annihilators

Kernel of a Covector

For a nonzero covector alpha, the kernel is a codimension-one subspace of V, since alpha, viewed as a linear map to the one-dimensional space F, has rank one by the rank-nullity theorem.

dim (ker α) = dim (V) - 1

Every codimension-one subspace of V arises as the kernel of some nonzero covector, and two nonzero covectors have the same kernel exactly when one is a nonzero scalar multiple of the other, giving a bijection between one-dimensional subspaces of V* and hyperplanes through the origin in V.

Annihilator Subspaces

For a subspace W of V, the annihilator W-naught is the subspace of V* consisting of all covectors vanishing identically on W.

W^{0} = {α \in V^{*} : α (w) = 0 for all w \in W}

The annihilator satisfies the dimension identity dim(W) + dim(W-naught) equals dim(V), and it is naturally isomorphic to the dual space of the quotient V/W, giving an algebraic bridge between subspace lattices of V and V* that underlies exact sequence arguments in linear algebra.

Algebra of the Pullback Operation

Pullback as an Algebraic Homomorphism

The assignment sending a linear map f to its pullback f* is itself algebraic in nature: it is contravariantly functorial, meaning it reverses composition order and preserves identities.

(g \circ f) * = f^{*} \circ g^{*}

for composable linear maps f from U to V and g from V to W. This contravariant functoriality is a purely algebraic law, verified directly from the definition of pullback by unwinding both sides on an arbitrary vector, and it is what makes the assignment V maps to V*, f maps to f* a well-defined contravariant functor on the category of vector spaces.

Rank Behavior Under Pullback

The rank of the pullback f* equals the rank of the original map f, since transposition preserves matrix rank. Consequently f* is injective exactly when f is surjective, and f* is surjective exactly when f is injective, a pair of algebraic equivalences that follow directly from the rank-nullity theorem applied to f and to f*, and that recur throughout duality theory as the precise algebraic dictionary translating properties of a map into properties of its dual.