8.23.1 Tensor Index Notation Definition Boundary

Tensor Index Notation Definition Boundary is the limit of the preconditions that must already be in place — a vector space with a fixed basis, a well-defined dual space and dual basis, a finite and definite index range — before an index expression can be considered defined at all, marking the difference between a case where index notation simply has not yet been set up and a case where it has been set up but is merely difficult or subtle to work with. It concerns the prior question of whether the notation's basic vocabulary (upper index, lower index, dummy index, contraction) even has a referent, as opposed to the boundary of what operations can be validly performed once that vocabulary is already available.

The Boundary at Missing a Fixed Basis

No Components Without a Basis Choice

Index notation's components are only defined relative to a chosen basis of the vector space and the corresponding dual basis of its dual space; before any basis is fixed, an abstract tensor exists as a coordinate-free multilinear map, but the very act of writing T^i_j presupposes that a basis has already been selected. An expression using index notation without any basis specified or implied by context has not yet crossed into the domain where the notation is defined — it is a placeholder for a translation that has not been completed.

Requiring a Compatible Dual Basis

Even once a basis of V is fixed, index notation additionally requires the dual basis of V* — the unique basis satisfying eⁱ(eⱼ) = δⁱⱼ — to be well defined; in settings where the dual space or a compatible pairing between V and V* is not available or not canonically determined (as can occur in some infinite-dimensional settings where the dual space is much larger than the original space), the upper-index/lower-index distinction that index notation depends on cannot be set up as straightforwardly, placing such settings at this boundary.

The Boundary at an Undefined or Non-Finite Index Range

The Range Must Be a Determinate, Fixed Set

Index notation's definition presumes that each index ranges over a determinate set — ordinarily {1, ..., n} for a finite-dimensional space of dimension n; if the dimension of the underlying space is not fixed (for example, in a family of spaces of varying dimension considered together, or when the space is defined only up to an unspecified finite dimension), an index expression written before this dimension is pinned down is not yet a fully defined instance of index notation, even though it may be a useful schematic template for one.

Ill-Posed Ranges From Inconsistent Context

An index letter used inconsistently, so that its intended range is unclear or contradictory from context — for instance a letter meant simultaneously to range over a full spacetime index set in one part of an expression and a restricted spatial-only index set in another, without a clarifying convention distinguishing the two — sits at this boundary because the notation depends on every index's range being unambiguous; resolving this typically requires introducing separate letter conventions (such as Greek versus Latin indices) precisely to keep ranges determinate.

The Boundary at the Underlying Object Not Being Genuinely Multilinear

A Prerequisite the Notation Assumes Without Stating

Index notation is defined for multilinear maps — objects linear in each argument separately — and its component-extraction procedure (evaluating on basis elements and using linearity to reconstruct the whole) is valid only because multilinearity guarantees that behavior on basis elements determines behavior everywhere. If the object under discussion is not actually multilinear in each of its slots, the entire component-extraction step that defines what its "components" even are breaks down, meaning index notation was never truly applicable to that object in the first place, independent of any operation later performed on it.

Diagram of the Definitional Prerequisites

Distinguishing the Definition Boundary From the Contraction Boundary

A Prior Question, Not a Downstream One

The definition boundary concerns whether index notation's basic apparatus has been legitimately set up at all for a given object, while the contraction boundary (and other operation-specific boundaries within the general index notation boundary) concerns whether a particular operation applied to already-defined index expressions remains valid; an object can fail at the definition boundary — lacking a fixed basis or failing to be multilinear — well before any question of contracting its indices could even be posed.

Resolving Definition-Boundary Cases

Cases at the definition boundary are resolved not by extending the rules of contraction or transformation, as with operation-specific boundaries, but by supplying the missing prerequisite directly: fixing a basis, establishing a workable dual pairing, pinning down a finite dimension, or confirming multilinearity; once these prerequisites are secured, the object crosses fully into the domain where index notation is defined, and only then do the further, operation-level boundaries become the relevant concern.