6.1.5 Tensor Component Organization Scope

Tensor Component Organization Scope defines how tensor elements are structured and indexed in algebraic contexts.

Tensor Component Organization Scope is the delimitation of the subject matter that belongs to how a tensor's individual numerical components are arranged once a basis has been chosen, covering multi-dimensional array layout, index-ordering conventions, and addressing schemes for retrieving or writing a specific component, as distinct from the abstract, basis-free classification of a tensor by slot count, order, type, or valence. This scope statement fixes component organization as a representational matter — how numbers are laid out and addressed — and separates it from the classification schemes that describe a tensor independently of any chosen basis or storage layout.

What Falls Inside This Scope

The Component Array as a Basis-Dependent Object

Within scope is the fact that once bases are fixed for each argument slot of a tensor, its components form a multi-dimensional array indexed by one integer per slot, with the size of the array along each dimension equal to the dimension of the vector space supplying that slot's basis.

T_{i_{1} \dots i_{k}}, with each i_{m} ranging from 1 to \dim (V_{m})

Index-Ordering Conventions

Within scope are the conventions that fix, once and for all within a given exposition, the order in which slot indices are written and read: row-first versus column-first addressing for order-2 tensors, and the corresponding leftmost-slot-varies-slowest or leftmost-slot-varies-fastest convention for higher-order tensors stored in flat memory.

Nested Array Visualization

Addressing a Single Component

Within scope is the mechanics of retrieving or assigning one specific component given a full index tuple (i₁, ..., i_k), including the flattening formula used to map a multi-dimensional index tuple onto a single linear offset in contiguous memory:

offset = \sum_{m} i_{m} \prod_{l > m}^{k} \dim (V_{l})

Organizing Components by Slices and Blocks

Within scope is the practice of viewing a higher-order tensor's component array as organized into lower-order slices or blocks — fixing one index and letting the rest vary — which is a purely organizational device used to display, traverse, or process a large component array in manageable pieces.

What Falls Outside This Scope

Basis-Free Classification by Order, Type, or Valence

Whether a tensor has order 2, type (1,1), or a particular valence sequence is determined without reference to any chosen basis or array layout, and belongs to the classification scopes covering order, type, and valence rather than to component organization, which only begins once a basis has already been fixed.

Transformation of Components Under a Change of Basis

How the numerical values of components change when a different basis is chosen is governed by the multilinear transformation law and belongs to tensor transformation, a separate topic from the static question of how components already computed in one fixed basis are arranged and addressed.

Symmetry Patterns Among Component Values

Whether the numerical values obey a symmetry or antisymmetry pattern under index exchange concerns the values themselves, not the organizational scheme used to store or address them, and is treated under a separate symmetry classification.

Computational Storage Formats for Sparse Tensors

Specialized sparse storage formats that avoid storing zero components explicitly are an implementation-level optimization on top of the dense organizational scheme addressed here, and belong to computational tensor methods rather than to this scope.

Boundary Cases Addressed Within This Scope

Organization of a Scalar's Single Component

A scalar, having zero slots, is within scope as the degenerate case of an array with no indices at all, a single stored value with no addressing scheme required beyond direct reference.

Organization of Vector and Covector Components

A vector or covector, having one slot, is within scope as the simplest non-trivial case, a one-dimensional array addressed by a single index, with the row-versus-column display convention (as a column of numbers or a row of numbers) treated purely as a organizational choice with no bearing on the vector's type.

Consistency of Organization Across Equal Orders

Within scope is the requirement that two tensors of the same order sharing the same sequence of underlying vector spaces be organized with array dimensions in the same respective order, so that component-wise operations such as addition can be carried out index by index without a mismatch in array shape.

Purpose of Fixing This Scope

Separating Representation From Abstract Structure

By confining this scope to array layout and addressing, the classification schemes for order, type, and valence remain free of representational detail, allowing the same abstract tensor to be organized differently in different bases or storage schemes without altering its classification.

Providing a Concrete Basis for Computational Discussion

Fixing precise conventions for component organization gives later, more computational discussions of tensor contraction, tensor product implementation, and basis-dependent formulas a stable, unambiguous starting point for referring to "the component at index tuple (i₁, ..., i_k)."