2.20.5 Tensor Covariant Slot Preparation

Tensor Covariant Slot Preparation is the procedural machinery for setting up, labeling, ordering, and combining the covariant argument positions of a multilinear tensor before or during its construction, ensuring that each position designated to accept a vector from V is correctly built from a copy of the dual space V*, correctly ordered relative to the other slots, and correctly compatible with operations such as currying, symmetrization, and tensor product assembly that act on those positions. Where the dual vector space context supplies the abstract pairing between V and V*, and the covector space context treats V* as a whole space, covariant slot preparation addresses the concrete bookkeeping problem of arranging q such copies of V* into a working multilinear argument structure for a (p, q)-tensor.

What a Covariant Slot Is

Slot as an Argument Position

A covariant slot is a designated argument position in a multilinear map that accepts an element of V, and correspondingly is built, in the tensor product perspective, from one factor of V* in the tensor product T^p_q(V). A tensor of type (p, q) has exactly q covariant slots, each independently accepting a vector, alongside its p contravariant slots, each accepting a covector.

Why the Slot Is Built From V*, Not V

A slot that is meant to accept a vector v from V must itself be an element of V*, since applying a linear functional to v is the only way to extract a scalar from v alone. Preparing a covariant slot therefore means allocating a copy of V* in the tensor product, so that when the tensor is fully saturated, each V* factor evaluates against the vector supplied in the corresponding argument position, in accordance with the natural pairing.

Ordering and Labeling of Covariant Slots

Slot Order Matters

Because tensor multilinear maps need not be symmetric, the order in which covariant slots are listed is significant data, not an incidental labeling choice: swapping the arguments in positions j and k generally produces a different tensor unless the tensor happens to be symmetric in those two positions. Covariant slot preparation therefore fixes an explicit ordering, conventionally indexed j_1, ..., j_q, matching the order in which lower indices are written in component notation, T_{j_1 ... j_q}.

Slot Labeling for Mixed Tensors

For a mixed (p, q)-tensor, covariant slots and contravariant slots are prepared and labeled independently, since they draw from different spaces and transform by different rules; a common convention interleaves them only in written index notation, using upper indices for the p contravariant slots and lower indices for the q covariant slots, while the underlying tensor product itself keeps the V factors and V* factors as separate, explicitly ordered blocks.

Currying: Reducing Slot Count by Partial Application

Fixing One Slot at a Time

A (p, q)-tensor, regarded as a multilinear map, can be partially applied by fixing a single covariant slot's argument to a specific vector, producing a new multilinear map with one fewer covariant slot, of type (p, q - 1). This process, curry-ing the tensor in one argument, is the mechanism by which the linear-functional context's slot-wise evaluation extends to tensors of arbitrary rank: fixing all but one covariant slot reduces the tensor to a single linear functional in the remaining slot.

denotes the curried map obtained by fixing the first covariant slot to v_1 while leaving the remaining q - 1 covariant slots and all p contravariant slots free.

Currying as an Isomorphism of Tensor Spaces

The set of all curried maps obtainable this way corresponds to the canonical isomorphism T^p_q(V) ≅ Hom(V, T^p_{q-1}(V)), identifying a tensor with one fewer prepared covariant slot with a linear map from V into a lower-rank tensor space. This isomorphism justifies treating slot preparation and currying as reversible operations: any tensor can be viewed either as a fully saturated multilinear map or as a map producing lower-rank tensors one slot at a time.

Symmetrization and Antisymmetrization of Prepared Slots

Symmetrizing a Block of Covariant Slots

Once several covariant slots are prepared and ordered, they may be symmetrized, replacing the tensor with the average of its values over every permutation of those slots' arguments,

producing a new tensor whose covariant slots are interchangeable. Antisymmetrization applies the same averaging but with a sign determined by the parity of the permutation σ, producing an alternating tensor whose covariant slots flip sign under any transposition.

Why Slot Preparation Precedes Symmetry Operations

Symmetrization and antisymmetrization act on the block of already-prepared, already-ordered covariant slots; they are not meaningful without a fixed prior ordering to permute. Covariant slot preparation is therefore a logical prerequisite to defining symmetric tensor spaces and exterior powers, both of which are built as quotients or subspaces of the space of tensors with q prepared, ordered covariant slots.

Slot Compatibility When Composing Tensors

Concatenating Slots Under Tensor Product

When two tensors S, of type (p_1, q_1), and T, of type (p_2, q_2), are combined via the tensor product S ⊗ T, their covariant slots are concatenated into a single ordered list of q_1 + q_2 slots, with S's slots conventionally preceding T's. Correct slot preparation ensures that this concatenation is unambiguous, since each slot retains a record of which original tensor and which position within that tensor it came from.

Slot Alignment for Contraction

Before a contraction can be applied between a covariant slot of one tensor and a contravariant slot of another, or between two slots of the same tensor, both slots must be prepared as compatible pairs, one drawing from V* and one from V, in a way that allows the natural pairing to be applied directly. Covariant slot preparation is what guarantees this alignment is tracked explicitly, so that contraction can be performed on the correct pair of indices without ambiguity about which slot pairs with which.