3.7.3 Tensor Vector Covector Order Convention

The Tensor Vector Covector Order Convention establishes a standard for arranging tensor components, clarifying how vectors and covectors interact in algebraic structures.

Tensor Vector Covector Order Convention is the set of notational rules that fix whether a vector or a covector is written first when expressing the pairing between V and V*, and, more broadly, the rules governing the order in which upper and lower indices are listed in tensor component notation. Because the scalar value of the pairing does not depend on which argument is listed first, order conventions are purely notational choices adopted for consistency across a text or field, not mathematical requirements imposed by the pairing itself.

Ordering the Pairing Arguments

Covector-First Convention

Many texts write the pairing as <f, v>, listing the covector first and the vector second, mirroring the function-application notation f(v), where the functional naturally appears before its argument. This convention emphasizes the pairing as an act of the covector operating on the vector.

Vector-First Convention

Other texts, especially those emphasizing the symmetry between V and V**, write <v, f>, listing the vector first. This convention emphasizes that the pairing is a single bilinear form on V x V* treated symmetrically, without privileging either factor as the one doing the acting.

Equivalence of the Two Conventions

Both conventions describe the identical scalar:

〈 f, v 〉 = 〈 v, f 〉 = f (v)

Confusion arises only if a reader assumes a particular ordering carries additional meaning beyond notation; within a single consistent text, the chosen order convention is simply a stylistic decision fixed at the outset.

Index Order Conventions in Component Notation

Upper Indices Before Lower Indices

For a general (p, q) tensor, the standard convention lists all p upper indices together, followed by all q lower indices together, as in T^{i_1 ... i_p}_{j_1 ... j_q}, rather than interleaving them. This grouping keeps the contravariant and covariant parts of the tensor visually and structurally separated.

T_{j_{1} \dots j_{q}}^{i^{1} \dots i^{p}}

Relative Position of Upper and Lower Blocks

Some conventions place all upper indices before all lower indices when read left to right, such as T^{i_1 ... i_p}_{j_1 ... j_q}, while others interleave upper and lower slots in a fixed pattern to record the specific order in which arguments were originally introduced in a multilinear map, particularly when the tensor is not symmetric and the order of indices affects which slot corresponds to which vector space factor.

Why Order Convention Matters for Non-Symmetric Objects

Symmetric Versus General Bilinear Forms

The vector-covector pairing itself has no ordering ambiguity in its value because it pairs one factor from V and one from V*, which are different spaces, so swapping notation order cannot be confused with swapping mathematical content. This differs from a general bilinear form on V x V, such as an inner product, where swapping the order of two vector arguments can genuinely change the value unless the form is symmetric.

Consistency Requirement Within a Single Tensor

For a tensor with several upper or several lower indices of the same variance, such as a (2, 0) tensor T^{ij}, the order of the two upper indices does matter numerically unless T is symmetric, since T^{ij} and T^{ji} may differ. In this setting, the order convention is not merely stylistic; it fixes which numerical entry corresponds to which pair of basis vectors, and must be tracked consistently throughout a calculation.

Practical Guidance

Fixing a Convention Once

Because both the covector-first and vector-first notations for the simple pairing are equivalent, the only practical requirement is to fix one convention at the start of a discussion and apply it uniformly, so that expressions built from repeated pairings remain unambiguous and comparable across a derivation.

Reading Mixed Literature

When comparing sources that use different order conventions, the safest approach is to translate every expression back to the underlying evaluation f(v), which has an unambiguous meaning independent of notation, before comparing formulas or verifying an identity.

Diagrammatic Summary

The diagram shows both order conventions converging on the same scalar value, reinforcing that the ordering choice is a matter of notation rather than mathematics.