8.19.2 Tensor Contraction Upper Lower Pair

Tensor Contraction Upper Lower Pair is the specific notational configuration — one index occurrence in the upper position and one occurrence of the identical letter in the lower position, both within a single term — that the implicit summation convention recognizes as the trigger for contraction, distinguishing this one legitimate pattern of repetition from every other configuration a repeated letter might otherwise take. It is the elementary structural unit out of which every instance of tensor contraction, at any rank and in any context, is built.

Anatomy of the Pair

The Two Required Occurrences

An upper–lower pair consists of exactly two occurrences of a single index letter within one term: one occurrence attached as a superscript to some tensor factor, and one occurrence attached as a subscript to some tensor factor (which may be the same factor or a different one). In

the upper occurrence of $i$ sits on $A$ and the lower occurrence sits on $B$, together forming the complete pair that the summation convention interprets as an instruction to sum over $i$.

Opposition of Position Is the Defining Feature

What makes a repetition an upper–lower pair, rather than some other pattern, is specifically the opposition of position between the two occurrences; a letter appearing twice in the same position — twice upper, as in a hypothetical $A^{i}B^{i}$ without any accompanying lower $i$ — does not constitute an upper–lower pair and triggers no implicit summation under the standard convention, regardless of how visually similar the repetition may appear.

The Meaning Assigned to the Pair

Triggering Implicit Summation

Once an upper–lower pair is recognized within a term, the summation convention specifies that the term is to be summed over every value the shared index can take, from $1$ to $n$ in an $n$-dimensional space:

The pair is thus not merely a notational curiosity but the complete and sufficient signal needed to reconstruct this sum without any explicit summation symbol appearing in the written formula.

Rank Reduction Associated With the Pair

Each upper–lower pair present in a term removes one upper and one lower slot from the term's overall type, reducing its total rank by two. A term originally built from factors totaling type $(3,2)$ that contains one upper–lower pair reduces, after that pair is accounted for, to an effective free-index type of $(2,1)$, with the paired indices having vanished from the result entirely.

Where Upper–Lower Pairs Arise

Between Two Distinct Tensor Factors

The most common occurrence of the pair is between two separate tensors multiplied together, as in the vector inner product $A^{i}B_{i}$, where the upper occurrence belongs to one factor and the lower occurrence belongs to a different factor entirely, and the pair effectively links the two factors together into a single contracted result.

Within a Single Tensor's Own Indices

An upper–lower pair can also occur entirely within the indices of one single tensor, as in the trace-like expression $T^{i}{}_{i}$, where both occurrences of the letter belong to the same symbol $T$; this configuration contracts two of that tensor's own slots against each other, producing a scalar (or a lower-rank tensor, if $T$ carries additional free indices) directly from a single object rather than from the interaction of two factors.

Multiple Independent Pairs Within One Term

A single term can contain several upper–lower pairs simultaneously, each involving a different letter and each triggering its own independent summation, as in $R^{i}{}{i}S^{j}{}{j}$, where $i$ and $j$ form two separate pairs that are summed independently of one another within the same overall term.

Distinguishing a Valid Pair From a Collision

Exactly Two Occurrences, No More

A valid upper–lower pair requires that the letter in question occur exactly twice within its term; a third occurrence of the same letter, in any position, converts what would have been a clean pair into a repeated-index excess collision, for which the summation convention supplies no defined meaning.

The Pair Must Not Include a Free Occurrence

If one of the two occurrences forming an apparent pair is actually intended to remain a free index elsewhere in a larger governing equation, then treating it as part of a contraction pair — rather than protecting it as free — introduces a free–dummy collision rather than a legitimate contraction; recognizing a valid upper–lower pair therefore depends on confirming that neither of its two occurrences carries any other, conflicting role in the surrounding context.

Role Within the Index Contraction Pattern

The upper–lower pair is the smallest and most fundamental unit of the broader contraction pattern found throughout tensor algebra: every instance of contraction, from a simple vector inner product to a metric trace to the complex multi-index contractions found in tensors like the Ricci tensor, is ultimately built by combining one or more of these elementary pairs. Understanding the pair as the base case clarifies why contraction rules — one upper, one lower, exactly two occurrences — apply uniformly regardless of how many indices a given tensor expression may otherwise contain.