3.18.5 Tensor Pairing Bracket Notation

Tensor Pairing Bracket Notation denotes bilinear pairings between tensors, structuring interactions in algebraic contexts.

Tensor Pairing Bracket Notation is a symbolic convention used to denote the evaluation of a covector (or, more generally, a tensor) on a vector, written as an angle-bracket pairing such as the expression that sends a covector and a vector to the scalar obtained by applying the covector to the vector. It formalizes the canonical bilinear pairing between a vector space and its dual space, providing a compact and coordinate-free way to express duality relationships that would otherwise require explicit function-application syntax.

Definition and Basic Form

The Canonical Pairing

For a vector space V over a field F and its dual space V* consisting of all linear functionals on V, the pairing bracket notation denotes the bilinear map that takes a covector and a vector and returns a scalar.

⟨ α, v ⟩ = α (v) \in F

Here alpha belongs to V* and v belongs to V. The bracket is not merely typographic sugar; it emphasizes that the pairing is symmetric in role between vectors and covectors, since a vector v can equally be regarded as a linear functional on V* by the natural evaluation v(alpha) = alpha(v). This dual viewpoint is central to finite-dimensional duality theory, where V is naturally isomorphic to its double dual V**.

Bilinearity

The pairing bracket is bilinear in both arguments, meaning it is linear in the covector slot for fixed vector, and linear in the vector slot for fixed covector.

⟨ a α + b β, v ⟩ = a ⟨ α, v ⟩ + b ⟨ β, v ⟩

and

⟨ α, c u + d v ⟩ = c ⟨ α, u ⟩ + d ⟨ α, v ⟩

for scalars a, b, c, d and vectors u, v and covectors alpha, beta. This bilinearity is what allows the pairing to be treated as a tensor of type (1,1) on V, namely an element of V* tensor V acting as a bilinear form on V* times V.

Relation to Coordinates and Bases

Dual Basis Characterization

Given a basis e_1, ..., e_n of V, the dual basis e^1, ..., e^n of V* is defined precisely through the pairing bracket by the Kronecker delta relation.

⟨ e^{i}, e_{j} ⟩ = δ_{j}^{i}

This identity is the defining property of the dual basis, and the bracket notation makes it immediately legible as a statement about pairings rather than about function composition or matrix multiplication.

Coordinate Expansion

If a covector alpha has components alpha_i in the dual basis and a vector v has components v^j in the primal basis, the pairing reduces to the familiar contraction sum.

⟨ α, v ⟩ = \sum_{i = 1}^{n} α_{i} v^{i}

Under the Einstein summation convention this is written simply as alpha_i v^i, and the bracket notation is often reserved for the basis-independent statement of this same contraction, distinguishing the invariant pairing from its coordinate representative.

Distinction from Related Notations

Bracket Notation versus Inner Product Notation

The pairing bracket must be distinguished from an inner product notation on an inner-product space, even though both may use angle brackets. An inner product pairs two vectors from the same space V and requires a chosen bilinear or sesquilinear form, whereas the tensor pairing bracket pairs a vector with a covector and requires no auxiliary structure beyond the vector space and its algebraic dual. When V carries an inner product, there is a canonical isomorphism from V to V* sending a vector to the functional it induces, and composing this isomorphism with the pairing bracket recovers the inner product; but the pairing bracket itself is more primitive and exists for any vector space, including infinite-dimensional ones where no canonical inner product is given.

Bracket Notation versus Function Application

Writing alpha(v) and writing the bracket pairing of alpha and v denote the same scalar, but the notations carry different emphases. Function-application notation stresses that alpha is an operator acting on v, while bracket notation stresses the symmetric, tensorial character of the pairing as a map from V* times V to the base field. This symmetry becomes essential when discussing the pairing between higher tensor spaces, such as the pairing of a rank-k covariant tensor with k vectors, or the pairing of mixed tensors with appropriate combinations of vectors and covectors.

Extension to Tensor Pairings

Pairing of Multilinear Tensors

The bracket notation generalizes naturally to pair a covariant tensor of rank k with an ordered k-tuple of vectors, or a contravariant tensor of rank k with an ordered k-tuple of covectors.

⟨ T, v_{1}, \dots, v_{k} ⟩ = T (v_{1}, \dots, v_{k})

Here T is an element of the tensor product of k copies of V*, viewed as a multilinear functional on V times cdots times V. This extension is what allows the pairing bracket to serve as the universal notation linking tensor algebra to multilinear algebra: any covariant tensor is fully characterized by its pairings against tuples of vectors from a basis.

Partial Pairing and Contraction

The bracket also accommodates partial pairing, where only some slots of a tensor are paired against vectors or covectors, leaving a tensor of lower rank as the result. This partial pairing is the notational precursor to tensor contraction, since pairing a single covector index against a single vector argument and summing is precisely the contraction operation. In this sense, the pairing bracket notation is the atomic building block from which the more general index-contraction formalism of tensor algebra is assembled, and it remains the preferred notation whenever basis-independence needs to be made visually explicit.

Role in Pullback and Naturality Statements

Naturality of the Pairing

The pairing bracket is the natural transformation witnessing duality: for a linear map f from V to W, the pairing satisfies the adjunction identity relating the pullback f* on W* to the original map f on V.

⟨ f^{*} β, v ⟩ = ⟨ β, f (v) ⟩

for every covector beta in W* and vector v in V. This identity is, in fact, the defining equation of the pullback map f*, and it shows why bracket notation is indispensable: the pullback of a covector is defined precisely so that this pairing identity holds for all vectors, making the bracket notation the language in which duality, pullback, and adjunction are most naturally expressed and verified.