11.2.4 Tensor Slot Variance Area

Tensor Slot Variance Area is the domain of theory concerned with treating each individual index position of a tensor as an independently classified slot, studying how a multilinear object's argument positions are organized, labeled, and manipulated according to the variance type assigned to each slot separately from the others.

Core Area: The Slot as the Unit of Classification

Slots as Argument Positions of a Multilinear Map

A tensor of a given type is naturally viewed as a multilinear map accepting a fixed number of vector arguments and a fixed number of covector arguments, and each argument position is called a slot. Slot variance area treats these positions, rather than the tensor as an undivided whole, as the basic unit to which a variance type, contravariant or covariant, is assigned.

Independence of Slot Behavior

Because each slot corresponds to a distinct factor in a tensor product of copies of the vector space and its dual, the transformation behavior of one slot under a change of basis has no bearing on the transformation behavior of any other slot in the same tensor, a structural independence that is the defining feature explored in this area.

Area: Slot Ordering and Labeling Conventions

Fixed Ordering of Slots

This area addresses the convention that the slots of a tensor are given a fixed order, typically all contravariant slots listed before all covariant slots, or interleaved according to a stated convention, so that a numerical component array can be unambiguously associated with a specific assignment of arguments to slots.

Symmetry and Antisymmetry Restricted to Same-Variance Slots

An important consequence studied in this area is that symmetrizing or antisymmetrizing a tensor over a group of its slots is only meaningful when those slots share the same variance type, since permuting a contravariant slot with a covariant slot would not correspond to a well-defined operation on the underlying multilinear map.

Area: Slot Contraction

Pairing an Upper Slot With a Lower Slot

Contraction within slot variance area is described as selecting one contravariant slot and one covariant slot from a tensor and summing over the index shared between them, reducing the total number of slots by two and producing a new tensor with the remaining slots retaining their original individual variance types.

Restriction Against Same-Type Contraction

A structural rule emphasized in this area is that two slots of the same variance type can never be directly contracted with the summation convention alone, since doing so would violate the requirement that a summed index appear once as a superscript and once as a subscript; converting one of the two slots first, typically through the metric, is required before such a pairing becomes possible.

Area: Slot Assignment in Tensor Products

Tracking Slots Through a Product Operation

When two tensors are combined by a tensor product, this area studies how the resulting object's slots are formed by concatenating the slot lists of the two factors, with each slot in the product retaining the exact variance type it had in its original factor, unaffected by the variance types present in the other factor.

Slot Reordering Under Relabeling

This area also covers the bookkeeping involved when a chosen convention reorders the slots of a tensor product for notational convenience, confirming that such a reordering is a purely notational relabeling that leaves each slot's individual variance type, and the overall multilinear map itself, unchanged.

Practical Role of Slot-Level Thinking

Diagnosing Malformed Expressions

Focusing on slots individually, rather than on a tensor as an undifferentiated array, is the standard diagnostic approach for catching malformed index expressions, since verifying that every contraction pairs a genuinely contravariant slot with a genuinely covariant slot is most naturally carried out one slot at a time rather than by inspecting an entire expression at once.