8.18.3 Tensor Covariant Contravariant Balance

Tensor Covariant Contravariant Balance is the specific requirement that, for a tensor expression to reduce to a coordinate-independent scalar invariant, every contravariant (upper) index it contains must be paired off, through contraction, against an equal number of covariant (lower) indices, so that the inverse and direct Jacobian factors each index contributes under a change of coordinates cancel completely, leaving no net dependence on the choice of coordinate system. It is the aspect of the index balance rule concerned specifically with the interplay between upper and lower indices as it bears on invariance, rather than with the more general bookkeeping of free-index letters and counts.

Why Cancellation Requires Equal Numbers of Each Kind

One Contraction Cancels One Upper–Lower Pair

Each individual contraction between one upper and one lower index eliminates exactly one factor of the direct Jacobian and one factor of its inverse, since these two factors are matrix inverses of one another:

Because each single contraction removes exactly one of each kind, reducing an expression to a fully coordinate-independent scalar — with no free indices and no uncancelled Jacobian factors remaining — requires precisely as many contractions as there are upper indices, and precisely that many again matched against an equal number of lower indices.

An Unequal Count Leaves an Uncancelled Residue

If an expression contains, for example, two upper indices but only one lower index available to contract against them, at most one upper–lower pair can be cancelled, leaving one upper index with no partner; the resulting expression retains a free index and an uncancelled Jacobian factor, and is therefore not a scalar invariant but a genuine, coordinate-dependent tensor component of nonzero rank.

Balance as a Precondition for Full Contraction

Total Contraction Requires Type (p, p)

A tensor can be fully contracted down to a scalar, using only the pairing of its own indices against one another (without invoking an external metric), only if it has an equal number of upper and lower indices to begin with — that is, only if its type is $(p, p)$ for some $p$. A tensor of type $(2,1)$, having one more upper index than lower, cannot be reduced to a scalar by internal contraction alone, since one upper index will always remain unpaired.

The Metric Restores Balance When It Is Otherwise Absent

When a tensor's native type is not balanced between upper and lower — for instance, a purely contravariant tensor of type $(2,0)$ — the metric tensor can be introduced specifically to supply the missing lower indices needed to restore covariant–contravariant balance, converting an otherwise unbalanced expression into one that can be fully contracted. The familiar squared-length expression

illustrates this directly: the purely contravariant vector $A^{i}$, of type $(1,0)$, is paired against the metric's two lower indices, restoring the balance needed to produce a coordinate-independent scalar.

Balance in Multi-Term Scalar Expressions

Each Term Must Independently Achieve Balance

When a scalar expression is built from several additive terms, each individual term must achieve covariant–contravariant balance on its own, since a scalar sum can only be coordinate-independent if every one of its addends is already coordinate-independent; a term with unmatched upper and lower indices cannot be rescued by other, properly balanced terms elsewhere in the same sum.

Balance Does Not Require Matching Structure, Only Matching Counts

Achieving covariant–contravariant balance in a term does not require the upper and lower indices to originate from the same original tensor or to be paired in any particular grouping; it requires only that the total count of upper indices equal the total count of lower indices once every contraction within the term has been carried out, regardless of which specific factors originally supplied each index.

Distinguishing This Balance From the General Free-Index Balance Rule

A Narrower, Invariance-Focused Requirement

Where the general index balance rule concerns the consistency of free indices across the terms and sides of an equation, covariant–contravariant balance concerns a narrower and more specific question: whether a given expression's internal mix of upper and lower indices permits it to be reduced entirely to a scalar. An expression can satisfy general free-index balance — having a well-defined, consistent free-index inventory — while still failing to be reducible to a scalar simply because it retains free indices of its own by design, with covariant–contravariant balance becoming relevant only when full reduction to an invariant is the specific goal.

Role Within the Index Balance Rule

Covariant–contravariant balance is the specialized application of the index balance rule to the particular question of scalar invariance, isolating the requirement that a tensor's upper and lower indices occur in equal number wherever complete self-contraction, possibly aided by the metric, is intended to eliminate every free index and yield a coordinate-independent result. It supplies the precise numerical condition — equal counts of each kind — that must hold before the cancellation of Jacobian factors underlying tensor invariance can be carried out in full.