3.7 Tensor Vector Covector Pairing Operation

Tensor Vector Covector Pairing Operation is the natural bilinear pairing between a vector space V and its dual V* considered as a single symmetric operation, V x V* -> F, that can be read either as a covector evaluating a vector or, through the canonical double-dual identification, as a vector evaluating a covector. It generalizes the plain evaluation f(v) by treating both orderings, <v, f> and <f, v>, as expressions of the same underlying pairing, and it is the operation that makes the relationship between V and V* fully symmetric rather than one-directional.

The Pairing as a Bilinear Form

Definition

The vector-covector pairing is the map

defined by <v, f> = f(v), so that reading the pairing with the vector listed first produces exactly the same scalar as evaluating the covector on the vector in the usual order. The pairing is bilinear: linear in v for fixed f, and linear in f for fixed v, following directly from the vector space structure of V and V*.

Symmetry of Notation, Not of Roles

Although the notation <v, f> and <f, v> are used interchangeably and always agree in value, the two arguments still play structurally different roles: f is the object doing the evaluating, since it is literally a function, while v is the object being evaluated. The symmetry is one of notation and numerical result, not of the underlying construction.

The Double Dual and the Canonical Embedding

Vectors as Functionals on Covectors

For a finite-dimensional vector space V, there is a canonical linear map ι : V -> V** sending each vector v to the functional ι(v) on V* defined by

for every f in V*. This map takes a fixed vector v and reinterprets it as a rule that consumes covectors, using exactly the vector-covector pairing to define its action.

The Map Is an Isomorphism in Finite Dimensions

When V is finite-dimensional, ι is a linear isomorphism, meaning every functional on V* arises this way from some vector v in V, and distinct vectors give distinct functionals. This canonical identification, requiring no choice of basis, is what justifies treating V and V** as literally the same space, and it is precisely what allows the pairing to be read symmetrically as <v, f> or <f, v>.

Component Form

Coordinate Expression

Relative to a basis e_1, ..., e_n of V and its dual basis e^1, ..., e^n, with v = v^i e_i and f = f_i e^i, the pairing has the same summation form regardless of the order in which it is written:

Because ordinary multiplication of numbers commutes, v^i f_i = f_i v^i, the coordinate formula gives no computational reason to prefer one ordering over the other; the choice of writing the vector or covector first is purely a matter of convention.

Role in General Tensor Contraction

Building Block for Higher Contractions

The vector-covector pairing is the elementary contraction step used repeatedly when reducing a higher-rank tensor. For a (p, q) tensor, contracting one of its p upper slots against one of its q lower slots is, at that single pair of indices, exactly this pairing operation applied locally, holding all other indices fixed.

Consistency Across Multiple Contractions

When several such pairings are applied in sequence to fully contract a tensor down to a scalar, the order in which the individual upper-lower pairs are contracted does not affect the final result, since each pairing operates independently on its own pair of indices and addition of the field is associative and commutative.

Diagrammatic Summary

The diagram shows the pairing operating symmetrically in both directions between V and V*, with both notational orderings collapsing to the identical scalar value in the field F.