13.9.2 Tensor Inner Product Slot Pairing

Tensor Inner Product Slot Pairing is a method in algebra that defines how tensors interact through inner products, pairing slots for structured computation.

Tensor Inner Product Slot Pairing is the specific assignment of correspondence between the index slots of two tensors, or of a tensor and a metric, that determines which contravariant slot on one factor is linked to which covariant slot on another when computing an inner product through contraction. It identifies the exact matching pattern used to align the indices of the participating objects, since an inner product between higher-rank tensors generally requires each index of one factor to be paired with a specific corresponding index of the other, rather than an arbitrary or unstated correspondence.

Conceptual Basis

Necessity of an Explicit Pairing

For vectors alone, there is only one index on each side, so the pairing is unambiguous. For tensors of rank greater than one, several indices exist on each factor, and the inner product slot pairing specifies which index of the first tensor is summed against which index of the second, since different pairings can yield different scalar results.

Pairing Through the Metric

Because an inner product often requires one contravariant and one covariant index to be joined by summation, and two tensors being compared may both carry contravariant indices, the slot pairing typically proceeds by first invoking the metric tensor to lower one index before the direct contraction with the corresponding index of the other tensor is carried out.

Pairing as a Bijective Correspondence

A complete inner product slot pairing between two tensors of the same rank associates each index slot of the first tensor with exactly one index slot of the second, forming a one-to-one correspondence that exhausts every index on both sides.

Formal Description

Pairing for Rank-Two Tensors

Given two rank-two tensors $A^{i j}$ and $B^{k l}$ , one valid inner product slot pairing links $i$ with $k$ and $j$ with $l$ , using the metric to lower each index of one tensor before contracting:

⟨ A, B ⟩ = g_{i k} g_{j l} A^{i j} B^{k l}

Alternative Pairing

A different, equally valid slot pairing on the same pair of tensors links $i$ with $l$ and $j$ with $k$ :

⟨ A, B ⟩ = g_{i l} g_{j k} A^{i j} B^{k l}

producing, in general, a different scalar value from the first pairing unless the tensors possess symmetry that equates the two results.

Requirement of Compatible Dimension

Each individual pair identified by the slot pairing must range over spaces of the same dimension, since the metric tensor or direct contraction used to join them is defined only when both slots share an identical index range.

Properties

Non-Uniqueness for Higher-Rank Tensors

Unlike the vector case, where only one slot pairing exists, tensors of rank two or higher admit multiple distinct slot pairings, each constituting a separate, well-defined inner product unless the tensors' symmetry properties force equivalence among them.

Preservation Under Simultaneous Relabeling

If both tensors undergo the same relabeling of their index names, the slot pairing and the resulting scalar remain unchanged, since relabeling is a notational convenience rather than a change to which slots are actually paired.

Interaction With Tensor Symmetry

When both tensors involved are symmetric under exchange of their indices, distinct slot pairings that differ only in the order of matching produce identical results, collapsing what would otherwise be multiple distinct inner products into a single value.

Practical Considerations

Explicit Specification in Ambiguous Cases

Whenever an inner product is computed between tensors of rank two or higher, the specific slot pairing intended must be stated explicitly, since the notation for the inner product alone does not disambiguate among the several valid pairings available.

Consequences of an Incorrect Pairing

Selecting a slot pairing that does not respect matching dimension, or that inadvertently pairs two indices of the same variance without an intervening metric, produces an ill-defined or invalid computation rather than merely an alternative valid result.

Role in Defining Norms of Higher-Rank Tensors

When the inner product slot pairing links each index of a tensor with the corresponding index of itself, the resulting scalar defines a generalized squared norm for that tensor, extending the vector norm construction to objects of arbitrary rank.