8.15.1 Tensor Dummy Index Renaming

Tensor Dummy Index Renaming is the operation of replacing every occurrence of a repeated (dummy) index within its scope with a different, otherwise unused letter, without changing the value or meaning of the tensor expression in which it appears. Because a dummy index exists only to signal an implicit summation and carries no significance beyond that role, it can be exchanged for any other symbol not already in use, and the resulting expression is not merely similar to the original but identically equal to it, term for term and value for value.

The Basic Operation

Substituting the Repeated Letter

Given an expression containing a repeated index, such as

dummy index renaming replaces both occurrences of $i$ — the upper one on $A$ and the lower one on $B$ — with a new letter, say $k$, producing

Both expressions expand to the identical sum $A^{1}B_{1} + A^{2}B_{2} + \cdots + A^{n}B_{n}$, confirming that the renaming has changed nothing about what the expression computes.

Why This Is Legitimate

The legitimacy of dummy index renaming follows directly from the definition of implicit summation: since $A^{i}B_{i}$ is shorthand for a sum over a fixed range, and the variable used to index that sum is a bound variable in the same sense as the variable of integration in a definite integral, the specific letter chosen to run through the range has no effect on the value of the completed sum. Just as $\int f(x),dx$ and $\int f(t),dt$ denote the same definite value, $A^{i}B_{i}$ and $A^{k}B_{k}$ denote the same scalar.

Required Conditions for a Valid Renaming

Avoiding Collision With Existing Indices

A dummy index may only be renamed to a letter that does not already appear elsewhere in the same term, whether as a free index or as a different dummy index. Renaming $i$ to $j$ in the expression $A^{i}B_{i}C_{j}$ would be invalid, since $j$ is already a free index in that term; the substitution would incorrectly merge the renamed dummy index with the pre-existing free index, changing the meaning of the expression rather than preserving it.

Renaming Both Occurrences Simultaneously

Because a dummy index by definition occurs exactly twice — once upper, once lower — a valid renaming must replace both occurrences together. Changing only the upper occurrence while leaving the lower one as the original letter destroys the pairing that made the index a dummy index in the first place, converting a single contracted expression into one with two unmatched free indices.

Practical Uses of Dummy Index Renaming

Preparing Expressions for Combination

Dummy index renaming is routinely used to prepare two separate tensor expressions for multiplication or further contraction when both happen to use the same letter for unrelated dummy indices. Given $A^{i}B_{i}$ and $C^{i}D_{i}$, multiplying them directly would create an ambiguous quadruple use of $i$; renaming the dummy index in one factor to $j$ beforehand, giving $C^{j}D_{j}$, produces an unambiguous product $A^{i}B_{i}C^{j}D_{j}$ in which the two independent sums remain clearly distinguished.

Simplifying and Standardizing Long Derivations

In multi-step tensor derivations, dummy indices are frequently renamed purely for clarity or to match the notation of a preceding or following line, without any computational necessity forcing the change. This cosmetic use of renaming relies on the same guarantee — that the value of the expression is unaffected — to justify altering notation freely in the service of readability.

Revealing Hidden Symmetries

Renaming dummy indices can expose that two expressions, which appear notationally different, are in fact identical. Given $A^{i}B_{i}$ and $A^{j}B_{j}$, renaming the second expression's dummy index $j$ back to $i$ immediately shows the two are the same quantity; this technique is used more substantively when comparing complex tensor identities that use different dummy letters purely as an accident of how each was originally derived.

Distinguishing Dummy Index Renaming From Other Index Operations

Contrasted With Free Index Substitution

Dummy index renaming must not be confused with substituting a specific numerical value for a free index, which is an entirely different operation: assigning $i = 2$ in $A_i = B_i$ extracts one particular scalar equation from the family the expression represents, whereas renaming a dummy index leaves the expression's meaning completely unchanged. Free indices are never subject to renaming in the sense described here, since they are not bound variables and their letters must match consistently across every term of a valid tensor equation.

Contrasted With Raising or Lowering

Renaming changes only the label attached to an index, never its position (upper or lower) or its numerical value; raising or lowering an index, by contrast, uses the metric tensor to convert a covariant component into a contravariant one or vice versa, altering the actual numerical content of the component. The two operations address entirely different aspects of an index — its arbitrary label versus its substantive transformation type — and are never interchangeable.

Role Within the Index Renaming Operation

Dummy index renaming is the paradigmatic instance of the broader tensor index renaming operation, and it is the case for which the justification is most direct: because a dummy index's letter carries no information beyond marking a pair for summation, any renaming that preserves the scope and pairing of that index leaves the expression's value provably unchanged. This property underlies nearly every other technique in tensor algebra that involves relabeling indices to combine, compare, or simplify expressions.