9.2.3 Tensor Expansion Area

Tensor Expansion Area explores how tensors expand in algebraic structures, revealing their role in multilinear operations and geometric transformations.

Tensor Expansion Area is the classification of the distinct senses in which the word "expansion" is used across tensor bases and coordinates — basis-component expansion, writing a tensor as a weighted sum of basis tensor products; Taylor-type expansion of a tensor field's components as a power series in displacement from a base point; and multipole-type expansion, representing a tensor field as a sum of terms of increasing angular or structural complexity — together with the characteristic role each expansion area plays in tensor calculation. It distinguishes these related but conceptually separate uses of "expansion" so that a reference to a tensor's expansion is understood in the correct one of these several senses.

Basis-Component Expansion

Writing a Tensor as a Sum of Basis Tensor Products

The most fundamental expansion area is the representation of a tensor as a finite sum of basis tensor products weighted by its components, T = T^{i}_{\ j} eᵢ ⊗ eʲ, which is the expansion underlying all of index notation; its scope and validity depend on the completeness of the chosen basis, as governed by tensor basis expansion scope, and it applies uniformly to any tensor of a fixed type at a single point (or, for a tensor field, at each point separately).

T = T_{j}^{i} e_{i} \otimes e^{j}

An Algebraic, Not Analytic, Operation

Basis-component expansion is a purely algebraic decomposition, requiring only linear algebra and no notion of differentiability, continuity, or nearby points; it is exact and finite, involving no approximation or truncation, which distinguishes it sharply from the Taylor and multipole expansion areas discussed next.

Taylor (Power-Series) Expansion of Tensor Field Components

Approximating a Tensor Field Near a Point

A tensor field's components, as functions of the coordinates, can be expanded as a Taylor series around a chosen base point, expressing the components at a nearby point as the value at the base point plus successive derivative correction terms:

T_{j}^{i} (x) \approx T_{j}^{i} (x_{0}) + (x - x_{0}) \partial_{k} T_{j}^{i} (x_{0}) + \dots

This expansion area is inherently approximate and local, its accuracy improving the closer the evaluation point is to the base point, and it underlies techniques such as normal coordinates, in which the metric's Taylor expansion around a point is arranged to vanish to first order.

Truncation and Order of Approximation

Because a Taylor expansion is generally an infinite series, practical use of this expansion area requires truncating at some finite order, with the size of the neglected remaining terms bounding the error of the approximation; this expansion area is therefore always accompanied by an explicit or implicit statement of the order to which it has been carried out.

Multipole and Structural Series Expansion

Decomposing a Field Into Terms of Increasing Complexity

A tensor field is sometimes expanded not in powers of displacement but in a series of terms distinguished by their angular or structural complexity — a monopole term, followed by dipole, quadrupole, and higher terms, each associated with a tensor of successively higher rank capturing successively finer structural detail of the field being represented.

Convergence and Truncation in This Area

As with Taylor expansion, a multipole-type expansion is generally infinite, and practical use retains only the leading terms judged most significant for the problem at hand, with the omitted higher terms representing detail the truncated expansion does not capture; unlike Taylor expansion, the ordering here reflects structural or angular complexity rather than proximity to a single base point.

Diagram Distinguishing the Three Expansion Areas

Relating the Three Areas Within a Single Calculation

Basis-Component Expansion Underlies the Other Two

Both Taylor and multipole expansion areas ultimately produce their successive correction terms as ordinary tensors, each of which must itself be represented via basis-component expansion in order to be written out in explicit index notation; the exact, algebraic expansion area is therefore the common substrate on which the two approximate expansion areas are built, rather than a competing alternative to them.

Choosing an Expansion Area for the Task at Hand

Whether a given calculation calls for basis-component expansion alone, or additionally requires a Taylor or multipole expansion, depends on whether the task is to represent a tensor exactly at a single point (basis-component expansion suffices) or to approximate a tensor field's behavior near a point or at a distance from a source (Taylor or multipole expansion becomes necessary), and recognizing which expansion area is actually called for is the first step in setting up the calculation correctly.