6.7.5 Tensor Type Transformation Role

Tensor Type Transformation Role is the function that the pair of numbers characterizing a tensor's contravariant and covariant orders performs in fixing the precise rule by which the tensor's components must change when the underlying basis or coordinate system is changed. The type of a tensor, conventionally written as an ordered pair where the first number counts contravariant (upper) indices and the second counts covariant (lower) indices, is not a passive label. It is the generating rule for the transformation law itself: every upper index contributes one factor built from the Jacobian of the new coordinates with respect to the old, and every lower index contributes one factor built from the inverse Jacobian. The type therefore determines, index by index, how many of each kind of factor must appear, and in what order they must be contracted with the tensor's components.

The Structural Link Between Type and Transformation Law

Contravariant Indices and the Direct Jacobian

Each contravariant index of a tensor transforms using the Jacobian matrix that carries old coordinates to new coordinates. If a tensor of type having at least one upper index is denoted with components indexed in the old system, its transformed components in the new system are obtained by contracting each upper index with one factor of the partial derivative of a new coordinate with respect to an old coordinate. This is the same rule that governs how the differential of a coordinate itself transforms, which is why contravariant indices are said to transform "like coordinate differentials."

Covariant Indices and the Inverse Jacobian

Each covariant index, by contrast, transforms using the inverse Jacobian matrix, the partial derivative of an old coordinate with respect to a new coordinate. This is the rule obeyed by the gradient of a scalar field, which is why covariant indices are said to transform "like the components of a gradient" or "like a basis of one-forms."

Mixed Tensors and Compound Transformation

A tensor of general type carrying several upper indices and several lower indices acquires one Jacobian factor for every upper index and one inverse-Jacobian factor for every lower index, all multiplying the original components and all summed over the repeated old indices. The type pair therefore counts, exactly, how many factors of each kind must be assembled: it is a recipe for the transformation law rather than an incidental classification attached after the fact.

Why the Type Must Govern the Law

Preservation of Multilinear Pairing

A tensor of type with p upper and q lower indices is, structurally, a multilinear map that consumes p one-forms and q vectors and returns a scalar. Because a scalar produced by such a pairing cannot depend on which coordinate system was used to describe the inputs, the transformation of the tensor's components is not a free choice: it is forced to be exactly the one that cancels, index by index, the transformation of the vectors and one-forms being fed into it. The upper indices must transform oppositely to basis vectors so that components paired with basis vectors stay invariant; the lower indices must transform oppositely to dual basis vectors for the same reason. The type transformation role is thus the mechanism that keeps the underlying multilinear object well defined independently of coordinates.

The Group Property of the Transformation

Because Jacobian matrices compose under successive coordinate changes, and inverse Jacobians compose in the corresponding opposite order, the transformation law dictated by a tensor's type is consistent under repeated changes of coordinates. Transforming from one system to a second and then to a third produces exactly the same result as transforming directly from the first to the third. This composability is only guaranteed because every upper index consistently uses the direct Jacobian and every lower index consistently uses the inverse Jacobian; mixing the assignment inconsistently across indices of the same kind would break the group property and make the notion of "tensor" incoherent under composition of coordinate changes.

Consequences of Type for Transformation Behavior

Type (0,0): Invariance Without Transformation

A tensor of type with no upper and no lower indices is a scalar. Its transformation role is trivial: a true scalar field carries the same numerical value at a given point regardless of the coordinate system, so no Jacobian factor is attached at all. This is the degenerate case that anchors the whole scheme, since every higher type reduces to scalar invariance once all of its indices are saturated by paired vectors and one-forms.

Type (1,0) and (0,1): The Elementary Building Blocks

A tensor of type with a single upper index is an ordinary vector and transforms with one factor of the direct Jacobian. A tensor of type with a single lower index is a one-form and transforms with one factor of the inverse Jacobian. Every higher-type transformation law is built by multiplying together this many copies of these two elementary rules, one copy for each index, which is why the type pair is sufficient by itself to reconstruct the entire transformation law without any further information about the tensor's internal structure.

General Type (p, q): Additive Bookkeeping

For a general type, the transformation role scales additively: p independent factors of the direct Jacobian and q independent factors of the inverse Jacobian appear, contracted against p and q dummy indices respectively. Raising the contravariant order by one adds exactly one more direct-Jacobian factor to the law; raising the covariant order by one adds exactly one more inverse-Jacobian factor. The type completely determines the number and kind of factors, though not their order of contraction, which follows the fixed correspondence between each new index's position and its transformation factor.

Practical Significance of the Type Transformation Role

Distinguishing Tensors From Non-Tensorial Arrays

Many arrays of numbers indexed by coordinates fail to be tensors precisely because they do not obey the transformation law dictated by any consistent type assignment. Christoffel symbols are the standard example: despite carrying three indices suggestive of a type (1,2) tensor, their transformation law contains an extra inhomogeneous term involving second derivatives of the coordinate change, so no type assignment reproduces their actual behavior. Recognizing the type transformation role is therefore also a diagnostic tool, since verifying whether an indexed quantity obeys the homogeneous multiplicative law associated with some type is the standard test for whether that quantity is a genuine tensor.

Operations That Change Type and Their Transformation Consequences

Raising an index with the inverse metric or lowering an index with the metric changes a tensor's type by shifting one covariant slot to a contravariant slot or vice versa. Because the metric and its inverse are themselves tensors obeying their own type transformation roles, these operations are consistent with, and indeed are forced by, the transformation roles of all the tensors involved: the new object automatically obeys the transformation law appropriate to its new type once the contraction with the metric is carried out, with no separate verification required.

Role in Formulating Coordinate-Independent Physical Law

In physical theories expressed in tensorial form, equating two tensors of the same type in one coordinate system guarantees that the equality holds in every coordinate system, since both sides transform by the identical Jacobian factors dictated by their shared type and those factors cancel between the two sides. This is the deepest practical use of the type transformation role: it is the guarantee that a law written tensorially is a statement about the underlying geometry or physics, not an artifact of the coordinate system chosen to express it.