8.13.5 Tensor Index Position Transformation Signal
Tensor Index Position Transformation Signal describes how tensor indices are repositioned to alter signal representation in algebraic computations.
Tensor Index Position Transformation Signal is the informational role that the vertical placement of an index — upper or lower — plays in announcing, at a glance and without further explanation, exactly which transformation law a given tensor component obeys under a change of basis or coordinates. The position of an index is not decorative; it functions as a signal, embedded directly in the notation, that tells a reader whether to apply the Jacobian matrix or its inverse when relating components in one coordinate system to components in another, without requiring any accompanying prose to state which law is in force.
What the Signal Communicates
Upper Position Signals the Inverse Jacobian
When an index appears in the upper position, the notation itself asserts that the component transforms with the inverse of the coordinate transformation matrix:
No separate statement is needed to specify that this particular law applies; the mere fact that $i$ sits above the baseline on $A$ conveys it directly.
Lower Position Signals the Direct Jacobian
Conversely, when an index appears in the lower position, the notation asserts the direct (non-inverted) transformation law:
The subscript alone is sufficient to fix which of the two mutually inverse Jacobian factors governs this component's behavior.
Why a Signal Is Necessary
Two Genuinely Different Transformation Laws Exist
Coordinate or basis changes admit two natural, mutually inverse ways for a set of numbers to respond: directly with the transformation matrix, or with its inverse. Since both behaviors occur constantly in tensor algebra — basis vectors transform one way, coordinate differentials transform the other — some notation is required to distinguish, term by term, which law applies to which quantity. Index position is the device chosen to carry exactly this distinction.
The Signal Is Local and Self-Contained
Because the signal is attached directly to each index rather than declared once for an entire equation, a single expression can mix both behaviors freely and remain unambiguous. In a mixed tensor $T^{i}{}_{j}$, the upper $i$ signals inverse-Jacobian behavior and the lower $j$ signals direct-Jacobian behavior within the very same symbol, with no risk of confusing which index obeys which law.
Reading the Signal in Practice
Verifying a Transformation Equation by Position Alone
Given any tensor transformation formula, the position of each index dictates, without further computation, which Jacobian factor must appear on the right-hand side and in which orientation. An expression is recognizable as ill-formed the moment an index in the upper position on the left is paired with the direct (non-inverted) Jacobian factor on the right, since the position signal has been violated.
Consistency Requirement Across an Equation
For a tensor equation to be valid, every occurrence of a given free index must appear in the same position on every term of the equation; if $i$ is upper on one side, it must be upper on every term where it appears freely. This requirement — sometimes called index balance — is a direct consequence of the transformation signal: mismatched positions across terms would imply that the two sides transform by different, incompatible laws, contradicting the assumption that they represent the same tensorial quantity.
Relation to Contraction
Position Signals Determine Contractibility
The transformation signal carried by index position is also what makes contraction well-defined. Contracting an upper index against a lower index works because their transformation laws are exact inverses of one another, so that the Jacobian factors cancel:
leaving a coordinate-independent scalar. This cancellation is only guaranteed to occur because one index's position signals the direct law and the other's signals the inverse law; two indices in the same position would produce Jacobian factors that do not cancel, and their product would not be an invariant.
Role Within Index Position Notation
The transformation signal is the semantic content underlying the purely syntactic rule that indices are written above or below the baseline. Everything else built on top of index position — the classification of tensors into type $(p,q)$, the rules for valid contraction, the requirement of consistent index placement across an equation — follows from treating index position as a faithful signal of transformation behavior. Recognizing this signal is what allows a reader to determine, from notation alone, whether a given tensor expression is coordinate-invariant, correctly balanced, and validly contracted.