9.22.5 Tensor Coordinate Free Interpretation

Tensor Coordinate Free Interpretation offers a deeper understanding of tensors by focusing on their intrinsic properties rather than coordinate systems.

Tensor Coordinate Free Interpretation is the reading of a tensor as an intrinsic multilinear object defined independently of any basis, in which components and coordinate expansions are understood as optional, secondary devices for computation rather than as part of what the tensor fundamentally is.

The Tensor as a Multilinear Map

Definition Without Reference to a Basis

A coordinate-free tensor of type ((p, q)) is defined as a multilinear map taking (p) covectors and (q) vectors as arguments and producing a scalar, with no basis mentioned anywhere in the definition.

T : \underset{p copies}{⏟} V^{*} \times \dots \times V^{*} \times \underset{q copies}{⏟} V \times \dots \times V \to ℝ

This definition specifies the tensor completely as a function with certain linearity properties in each argument; a basis is not required to state it, and two tensors are considered identical exactly when they produce the same output on every possible input, regardless of how that output is computed.

Vectors and Covectors Without Components

In this reading, a vector is an element of an abstract vector space (V), and a covector is a linear functional on (V), neither of which requires reference to coordinates; only when a specific computation is desired does one introduce a basis and extract components.

Coordinate Expansion as an Optional Tool

Components Appear Only Upon Choosing a Basis

The familiar component array (T^{i_1 \dots i_p}_{\ j_1 \dots j_q}) arises only after a basis is chosen and the multilinear map is evaluated on the basis vectors and dual basis covectors as arguments; the coordinate-free interpretation regards this step as a convenience for calculation, not as a necessary part of the tensor's identity.

T_{j_{1} \dots j_{q}}^{i^{1} \dots i^{p}} = T (e^{i_{1}}, \dots, e_{j_{1}}, \dots)

Tensor Operations Defined Without Coordinates

Operations such as addition, scalar multiplication, and the tensor product itself can all be defined directly on the multilinear maps, without any mention of coordinates: the tensor product (S \otimes T) of two tensors is defined by how it acts on arguments, and only afterward does one compute its components in a chosen basis.

Why the Coordinate-Free View Matters

Guaranteeing Basis Independence by Construction

Defining a tensor coordinate-free from the outset guarantees, automatically and without further argument, that any statement made about it is basis independent, since no basis was used in the definition; this contrasts with starting from a component array and having to separately verify that operations on it transform correctly under a change of basis.

Clarifying What a "Tensor Equation" Really States

A coordinate-free tensor equation, such as (T = S \otimes R), asserts an identity between multilinear maps as functions; the corresponding component equation, (T^{ij} = S^i R^j), is simply this same identity evaluated on basis elements, and the coordinate-free interpretation clarifies that the component equation is a consequence of the abstract one, not an independent fact requiring its own separate justification in every basis.

Visual Illustration

Relation to Coordinate-Dependent Notation

Tensor coordinate free interpretation does not discard component notation; it reframes it as one of many valid, interchangeable shadows cast by a single underlying multilinear object. Practical computation still relies on choosing a basis and working with components, but the coordinate-free view supplies the guarantee that whichever basis is chosen, the resulting components describe the same intrinsic tensor, and any two component sets related by the correct transformation law are simply different views of that one object.