11.1 Tensor Variance Behavior Scope

Tensor Variance Behavior Scope is the delineation of exactly which mathematical objects, operations, and index slots are subject to covariant or contravariant transformation rules, and which related quantities lie outside this classification entirely, clarifying the boundaries within which the labels covariant and contravariant carry precise meaning.

What Falls Inside the Scope

Individual Tensor Index Slots

Variance behavior applies at the level of a single index slot on a tensor, not to the tensor as an undifferentiated whole. Each upper index slot transforms contravariantly and each lower index slot transforms covariantly, independent of how many other slots the same tensor carries.

Basis Vectors and Dual Basis Vectors

The natural basis vectors associated with a coordinate system transform covariantly, while the dual basis one-forms transform contravariantly. This pairing is what allows a tensor's abstract, basis-independent value to remain fixed even as its component array changes under a change of basis.

Any Multilinear Object Built From Tensor Product Slots

Variance behavior scope extends to any object built as an element of a tensor product of copies of the underlying vector space and its dual, since each factor in the tensor product contributes one index slot with a definite variance type, following the same rules as an ordinary tensor.

What Falls Outside the Scope

True Scalars

A genuine scalar, meaning a quantity with no free indices at all, does not transform under a change of basis and therefore has no variance type to assign; it is neither covariant nor contravariant, but invariant. Referring to a scalar as having zero-order variance is a common shorthand, but the scalar itself does not participate in the transformation machinery that defines covariance and contravariance.

Non-Tensorial Quantities

Certain quantities that carry indices, such as the Christoffel symbols used in defining a connection, fail to transform by the pure tensor transformation rule at all, picking up an extra inhomogeneous term under a change of basis. Because they do not obey either the covariant or the contravariant rule cleanly, they lie outside the scope of tensor variance behavior even though they superficially resemble tensors in their index notation.

Densities Without Their Weight Factor Accounted For

A tensor density transforms like an ordinary tensor only after an additional power of the Jacobian determinant is included alongside the usual Jacobian factors. Considered without this weight factor, a density's raw component array does not obey the plain covariant or contravariant rule, placing the unweighted array outside the strict scope of ordinary tensor variance behavior, even though the fully weighted density transformation is well defined in its own extended framework.

Scope With Respect to the Underlying Space

Finite-Dimensional Vector Spaces and Manifolds

The scope of variance behavior as classically formulated presumes a finite-dimensional vector space at each point, or a finite-dimensional tangent space at each point of a manifold, since the index counting, the Jacobian matrix, and the summation convention all rely on a finite, well-defined dimension.

Extension and Its Limits

Generalizing variance behavior to infinite-dimensional settings requires replacing finite index sums with integrals or operator formalisms, and not every property that holds in the finite-dimensional scope, such as a well-defined determinant for the transformation, carries over automatically, so care must be taken when extending variance classification outside its original finite-dimensional scope.

Practical Use of the Scope Boundary

Confirming a Quantity Is a Tensor Before Assigning Variance

Because variance behavior is only meaningfully assigned to genuine tensors, the first step before labeling any indexed quantity as covariant or contravariant is confirming that it obeys the homogeneous tensor transformation law, without any extra additive term, since only then does describing its index slots by variance type carry the intended precise meaning.

Guarding Against Overextension of the Labels

Recognizing which quantities lie outside the scope of tensor variance behavior prevents the common overextension of applying covariant or contravariant transformation rules to objects, such as connection coefficients or raw density components, that were never designed to obey them, which would otherwise lead to systematically incorrect transformation results.