8.12.3 Tensor Implicit Contraction Expression

Tensor Implicit Contraction Expression is a notation used in tensor algebra to denote contraction without explicit indices, simplifying operations in multilinear algebra.

Tensor Implicit Contraction Expression is a tensor formula in which two indices — one upper, one lower — are written with the same letter and, by the summation convention, are automatically summed over their full range without any explicit summation symbol appearing in the notation. The expression itself looks like an ordinary product of tensor components, but the repetition of the index pair signals that the product is to be summed, reducing the total rank of the resulting object by two: one covariant and one contravariant degree of freedom are removed for every repeated index pair. It is the notational vehicle through which contraction — an algebraic operation independent of any particular symbol choice — is expressed compactly within index notation.

Structural Definition

The Repeated Index Pattern

An implicit contraction expression is recognized purely by its index pattern: a single letter occurring exactly twice within one term, once as a superscript and once as a subscript. For example,

A^{i} B_{i}

is an implicit contraction expression because the index $i$ appears once raised (on $A$) and once lowered (on $B$), signaling implicit summation:

A^{i} B_{i} = \sum_{i = 1}^{n} A^{i} B_{i}

Mandatory Opposition of Index Position

The pattern is only a valid implicit contraction expression when the repeated index occupies opposite positions — one up, one down. A repeated index appearing twice in the same position, such as both upper or both lower, does not constitute a contraction under the standard convention and is typically treated as a notational error unless a non-Euclidean metric convention is explicitly declared to justify it.

Contraction of Higher-Rank Tensors

Single-Index Contraction

When one index pair of a higher-rank tensor is repeated, the implicit contraction expression reduces the rank by two while leaving the remaining indices free. For a mixed tensor of type $(1,1)$,

T_{i}^{i} = \sum_{i = 1}^{n} T_{i}^{i}

produces a scalar — the trace of the mixed tensor. When the tensor carries additional free indices, such as $T^{i}{}_{i}{}^{j}$, contraction over $i$ leaves the free index $j$ untouched, producing a rank-1 object.

Multiple Simultaneous Contractions

A single term may contain several distinct repeated index pairs, each contracted independently. In

R_{i}^{i} S_{j}^{j}

the indices $i$ and $j$ are each contracted separately over the full implicit range, since they form two distinct repeated pairs rather than a single shared pattern.

Contraction via a Metric Tensor

Raising and Lowering Combined With Contraction

When both indices of a pairing are on the same level (both lower, for instance), the metric tensor is inserted explicitly to raise one index before contraction can occur:

g^{ij} A_{i} B_{j} = A^{j} B_{j}

Here the pair $i$ is contracted against the metric's first index, effectively raising $A_i$ to $A^j$, after which $j$ becomes the new implicitly contracted pair with $B_j$. This chained pattern is one of the most common uses of implicit contraction expressions in Riemannian geometry.

The Metric as a Full Contraction

Full contraction of a tensor against the metric on both indices reproduces its trace with respect to that metric:

g^{ij} T_{ij} = T

where the scalar $T$ on the right denotes the metric trace of the rank-2 covariant tensor $T_{ij}$.

Distinguishing Contraction From Mere Repetition

The Dummy Index Has No Independent Meaning

A hallmark of the implicit contraction expression is that the specific letter chosen for the repeated index carries no meaning outside the term in which it appears; it is a dummy index and may be renamed freely without altering the value of the expression, provided the new letter does not collide with another index already in use:

A^{i} B_{i} = A^{k} B_{k}

Free Indices Remain Untouched

Any index that appears only once within a term is a free index and is unaffected by contraction elsewhere in the same term; it persists into the result and must match, in name and position, on both sides of the governing tensor equation. Only the specific letters forming an implicit contraction pair are consumed by the operation.

Role in Tensor Algebra

The implicit contraction expression is the notational realization of tensor contraction, which is itself a linear map from tensors of type $(p,q)$ to tensors of type $(p-1,q-1)$. Within tensor index notation, every instance of an upper and lower index sharing a letter inside a single term is, by convention, exactly such a map applied to that pair of slots. This makes the implicit contraction expression one of the two core mechanisms — alongside implicit summation itself — by which compact index notation encodes operations such as matrix traces, inner products, divergence operators, and general tensor reductions without recourse to explicit summation signs or operator symbols.