✦ For everyone, free.

Practical knowledge for real and everyday life

Home

3.8 Tensor Natural Pairing Operation

The Tensor Natural Pairing Operation is a bilinear map connecting tensor spaces to their duals, enabling index contraction in algebra.

Tensor Natural Pairing Operation is the canonical bilinear map between a vector space and its dual that is defined without reference to any basis, coordinate system, or auxiliary structure such as an inner product, distinguishing it from pairings that require an arbitrary choice to be constructed. Because this pairing, <f, v> = f(v) for f in V* and v in V, follows directly from the definition of V* as the space of linear functionals on V, it exists automatically for every vector space, and the term natural signals precisely this basis-free, choice-free origin.


What Makes a Pairing Natural

Definition Without Auxiliary Choices

A mathematical construction is called natural, in the sense used across algebra and category theory, when it is defined purely from the given data and universal properties of the objects involved, without invoking an arbitrary selection such as a basis, an ordering, or a metric. The pairing between V and V* qualifies as natural because V* is defined as the set of linear functionals on V, and applying a functional to a vector is simply the built-in operation of that functional; no additional data needs to be supplied to make the pairing well-defined.

Contrast with Constructions Requiring a Choice

By contrast, an inner product on V, which pairs two vectors from V directly, is not natural in this sense: defining an inner product requires selecting a specific symmetric bilinear form, and different valid choices give different, inequivalent inner products on the same space. Likewise, identifying V with V* directly, without going through V**, requires choosing a basis or an inner product, and different choices give different, non-canonical identifications. The natural pairing avoids all such choices by pairing V with its own dual rather than with itself.


Formal Statement

The Pairing Map

The natural pairing is the bilinear map

μ : V* × V F , μ f,v = f v

built entirely from the fact that each f in V* is, by definition, a function V -> F. No basis of V needs to be fixed, and no property of F beyond it being a field is required to write down this definition.

Naturality Under Linear Maps

The pairing's naturality can be stated precisely using how it interacts with linear maps. If T : V -> W is a linear map between vector spaces, with dual map T^* : W^* -> V^* defined by (T^* g)(v) = g(Tv), then the pairing satisfies

T* g , v = g , T v

for every g in W* and v in V. This identity, holding for every linear map T without any further restriction, is what makes the pairing a natural transformation between the relevant functors in the categorical formulation of duality.


Consequences of Naturality

Basis Independence Follows Automatically

Because the pairing is defined without reference to a basis, its value cannot depend on a subsequent choice of basis used only to compute components. This gives an alternative, more conceptual explanation for the coordinate independence of the evaluation operation: the invariance is not something that must be separately proven from component transformation laws, it is built into the pairing from the outset by virtue of how it was defined.

Canonical Double Dual Identification

The naturality of the pairing is also what underlies the canonical isomorphism V &#8773; V** in finite dimensions. The map sending v to the functional f &#8614; f(v) on V* uses only the natural pairing and no auxiliary choice, which is why this particular identification of V with V** is considered canonical, unlike the identification of V with V* itself, which is not natural and always requires an extra choice such as a basis or inner product.


Natural Pairing Versus Chosen Pairings

Inner Products Are Not Natural

An inner product <,> : V x V -> F pairs V with itself, but requires selecting a specific positive-definite symmetric bilinear form; changing the inner product changes the numerical values produced, even though the underlying vector space V has not changed. The natural pairing, by contrast, gives the same values for f(v) regardless of any such auxiliary structure, since it depends only on V and its algebraically determined dual V*.

When a Chosen Identification Is Still Useful

In practice, once a basis or a metric is fixed for other reasons, such as performing explicit calculations, it becomes convenient to identify V with V* and treat the natural pairing as if it were an inner product on V. This identification is useful and common, particularly in physics, but it should be understood as an additional structure layered on top of the natural pairing, not a replacement for it.


Diagrammatic Summary

V* x V -> F no choice required V x V -> F needs a chosen form Natural pairing exists for every vector space automatically; an inner product must be separately selected.

The diagram contrasts the natural, choice-free pairing between V* and V with a pairing on V alone, which necessarily depends on an extra selected structure.

Content in this section