8.17.4 Tensor Index Type Matching

Tensor Index Type Matching is the specific requirement, within the broader index matching rule, that every term added together or set equal within a tensor equation must carry the identical overall type $(p,q)$ — the same number of upper indices and the same number of lower indices — even before checking whether the individual index letters and positions themselves align. Where the general matching rule concerns letter-by-letter and position-by-position agreement, type matching concerns the coarser, aggregate classification that must hold as a precondition before any finer-grained matching can even be meaningfully checked.

Statement of the Requirement

Type as the First Compatibility Check

Before comparing which specific letters occupy which specific positions across the terms of a tensor equation, each term's overall type must first agree: a term with two upper and one lower free index cannot be added to a term with one upper and one lower free index, regardless of what letters either term happens to use, because their types $(2,1)$ and $(1,1)$ do not match. Type matching is therefore the coarsest and first-applied layer of the full index matching rule.

Type Is Read From the Free-Index Structure

The type relevant to this requirement is determined by the free indices of each term, since dummy indices are contracted away and do not contribute to the type of the resulting expression. A term such as $T^{ij}{}_{k}A^{k}$, despite containing three total index occurrences, has type $(2,1)$ once the contraction over $k$ is accounted for, since $k$ is dummy and does not survive into the free-index structure of the term.

Why Type Must Match Before Anything Else

Addition Requires Identical Type

Tensor addition is only defined between two tensors of the same type, since addition is performed component by component, and two tensors of different type do not have components that correspond to one another in any well-defined way. A type $(2,0)$ tensor and a type $(1,1)$ tensor, even if both happen to have the same total rank, have no natural component-by-component correspondence, since one has two contravariant slots and the other has one of each kind.

Equality Requires Identical Type on Both Sides

Because an equation asserts that two tensors are the same object, expressed perhaps in different ways, both sides of a valid tensor equation must have matching type for the same underlying reason that two summands must: an equation between a type $(1,0)$ object and a type $(0,1)$ object cannot hold as a coordinate-independent statement, since the two sides would transform by different laws under a change of basis, one requiring the direct Jacobian factor and the other its inverse.

Type Matching Versus Position Matching

Type Matching Is Necessary but Not Sufficient

Two terms can share identical type — the same counts of upper and lower indices — while still failing the fuller index matching rule if the specific letters or their relative ordering disagree. A term of type $(1,1)$ written as $A^{i}{}{j}$ and another of type $(1,1)$ written as $A^{j}{}{i}$ match in type but not necessarily in the more detailed sense required for them to be added together meaningfully as representing the same free-index slots, since the roles of $i$ and $j$ have been swapped between the two.

Type Matching as a Fast Preliminary Filter

Because type is determined purely by counting index occurrences and classifying them as free or dummy, checking type matching is computationally and conceptually simpler than checking full letter-and-position matching, making it a useful first filter: an equation failing to match in type can be immediately identified as invalid without needing to trace through the specific letters used in each term.

Type Matching Under Metric Operations

Raising and Lowering Change Type but Must Change It Consistently

Applying the metric to raise or lower an index changes a term's type, and if such an operation is applied to only one term of a multi-term sum while the others are left unmodified, type matching across the sum is broken. Converting $A_{i}$, of type $(0,1)$, into $A^{i}$, of type $(1,0)$, by raising with the metric requires that every other term intended to be added to it also be raised to type $(1,0)$, or the resulting sum will no longer satisfy type matching.

Restoring Type Matching via the Metric

When two terms of an intended equation have mismatched type due to one having an index in a different position than the other, inserting an explicit factor of the metric tensor to raise or lower the mismatched index is the standard technique for restoring type matching before the rest of the equation can be verified or manipulated further.

Role Within the Index Matching Rule

Type matching functions as the coarse-grained layer of the general index matching rule, verifying the raw counts of upper and lower free indices before the finer question of specific letter and positional correspondence is examined. Any tensor equation that fails to satisfy type matching necessarily fails the full matching rule as well, making type matching a convenient and often sufficient first test for identifying malformed tensor equations before more detailed verification is undertaken.