11.19 Tensor Variance Convention

Tensor Variance Convention is the overall system of naming, notational placement, and summation rules by which tensor components and their transformation behavior are consistently expressed, encompassing the choice of upper versus lower index position, the implicit summation over repeated indices, and the terminology of covariant and contravariant that together allow tensor equations to be written and interpreted unambiguously.

Foundational Setting

Why a Convention Is Needed

Tensor algebra involves objects that transform in several distinct ways under a change of basis or coordinates, and without an agreed notational system, every equation would need to be accompanied by a separate verbal explanation of how each symbol behaves. The variance convention exists to make this behavior visible directly in the symbols themselves.

The Three Pillars of the Convention

The convention rests on three interlocking components: the assignment of upper or lower position to each index according to its transformation type, the implicit summation rule applied whenever an index is repeated in matched upper and lower positions, and a fixed terminology, covariant for lower indices and contravariant for upper indices, used to describe these behaviors in prose.

The Origin of the Terminology

Historical Motivation

The terms covariant and contravariant were introduced to describe how certain quantities vary together with, or against, a change of basis. A covariant quantity was so named because its components change in the same direction as the basis vectors, while a contravariant quantity changes in the opposing direction, requiring the inverse transformation.

Persistence of the Terms

Despite alternative, purely structural descriptions being available, such as referring simply to upper-index and lower-index behavior, the historical terms covariant and contravariant remain the standard vocabulary across nearly all treatments of tensor algebra, and the convention preserves this vocabulary alongside the notational index placement.

The Summation Rule Within the Convention

Einstein Summation Convention

A central part of the overall variance convention is the agreement that whenever an index letter appears exactly once as an upper index and once as a lower index within the same term, summation over that index is implied automatically:

Free Index Matching Rule

The convention also requires that any index appearing only once, a free index, must appear in the identical vertical position on every additive term of a valid equation, since a mismatch would imply that the two terms transform according to different rules and could not be legitimately added.

Visual Summary of the Convention

The Three Components Together

Consistency Enforced by the Convention

Type-Correct Equations

Because the convention ties index position directly to transformation behavior, any tensor equation that respects the convention automatically guarantees that both sides transform identically under a change of basis, which is what makes the equation meaningful as a coordinate-independent statement.

Detecting Violations

An expression that violates the convention, for instance by placing the same free index as upper on one side of an equation and lower on the other, signals immediately, without any further computation, that the equation as written cannot represent a valid tensor identity.

Variation in Convention Across Contexts

Alternative Notational Choices

Some treatments favor abstract, coordinate-free notation that avoids explicit indices altogether, while others adopt differing conventions for where summation symbols are written explicitly versus left implicit. Despite these surface differences, the underlying logical content, upper indices behave contravariantly and lower indices behave covariantly, remains the same across essentially all standard sources.

Orthonormal Simplifications

In contexts restricted to orthonormal bases, some authors relax the convention and write all indices as subscripts, since the distinction between covariant and contravariant components collapses numerically in that special case, though the underlying transformation distinction persists in principle.

Summary of Key Traits

Defining Characteristics

• The convention links upper index position to contravariant behavior and lower index position to covariant behavior.
• Repeated indices in matched upper-lower position imply summation without an explicit symbol.
• Free indices must match in position across every term of a valid equation.
• The convention's core logic remains stable across different notational styles and simplifies, without disappearing, in orthonormal-basis contexts.