✦ For everyone, free.

Practical knowledge for real and everyday life

Home

6.3.1 Tensor Order Slot Count

Tensor Order Slot Count defines the number of indices required to uniquely identify each component in a tensor, reflecting its rank and structural dimensionality.

Tensor Order Slot Count is the specific integer k = p + q obtained by adding a tensor's number of contravariant slots p to its number of covariant slots q, serving as the exact numerical quantity that tensor order classification names and organizes into the scalar, vector, matrix-like, and higher-order categories. It is the single arithmetic fact that must be computed correctly before any order-based name can be assigned, and it is defined so as to depend only on how many index positions a tensor's components carry, never on the values those indices take or on the dimension of the underlying vector space.


The Formula and Its Terms

Stating the Formula

k = p + q

where p is the number of upper (contravariant) index positions and q is the number of lower (covariant) index positions appearing in a component expression such as T_{j₁...j_q}^{i₁...i_p}.

Reading the Slot Count Off a Component Expression

Given a specific component expression, the slot count is obtained by literally counting the index symbols: T^{i} has k = 1, T^{i}_{j} has k = 2, T^{ij}_{kl} has k = 4, with no other computation required.

Slot Count as the Arity of a Multilinear Map

Since a tensor of type (p, q) corresponds to a multilinear map taking p covector arguments and q vector arguments (or, dually, p vector-valued and q covector-valued slots depending on convention), the slot count k is exactly the total number of arguments the map accepts.

T : ×k V F

Computing the Slot Count Under Operations

Slot Count of a Tensor Product

k (AB) = k (A) + k (B)

follows directly from concatenating the index lists of the two factors: an order-2 tensor tensored with an order-3 tensor produces an order-5 tensor, 2 + 3 = 5.

Slot Count After a Contraction

k ( contraction of T ) = k (T) 2

follows because contracting pairs one upper slot against one lower slot and sums over the shared index, removing both slots from the surviving expression: an order-4 tensor, after one contraction, becomes an order-2 tensor.

Worked Example: Computing the Slot Count of a Composite Expression

For the expression Sᵢⱼ = Σₖ Tᵏᵢⱼₖ, the original tensor T has slot count k(T) = 1 + 3 = 4 (one upper index k, three lower indices ijk), and after the single contraction implied by the repeated k index, the resulting tensor S has slot count k(S) = 4 - 2 = 2, matching its two visible free indices i and j.

Diagram of Slot Count Arithmetic

k=1 k=3 = k=4 → contract → k=2

What the Slot Count Does Not Depend On

Independence From the Split Between p and q

The slot count k is a single sum and does not by itself reveal how many of its slots are upper versus lower; the values p = 3, q = 1 and p = 1, q = 3 both give k = 4, so recovering p and q individually requires type classification, not the slot count alone.

Independence From Dimension

The slot count does not depend on dim(V); a k = 2 tensor over a 2-dimensional space and a k = 2 tensor over a 100-dimensional space are both order-2 by slot count alone, even though the former has 4 components and the latter has 10,000.

Independence From Component Values

The slot count is fixed by the index structure of a tensor's definition, not by the specific numbers stored in its components; scaling every component of a tensor by a nonzero constant, or setting them all to zero, leaves the slot count entirely unchanged.


Boundary and Special Cases

Slot Count Zero

k = 0 is the slot count of a scalar, the unique case in which there are no index positions to count at all, and the base case from which every other slot count is built up by repeated addition under the tensor product.

Slot Count Is Always a Non-Negative Integer

Because p and q are themselves non-negative integers (counts of index positions), their sum k is always a non-negative integer; a contraction is only ever applied when at least one upper and one lower slot remain, so k never becomes negative through the subtraction rule.


Why the Slot Count Matters

The Single Number That Anchors Order Classification

Every name attached by order classification — scalar, vector, matrix-like, higher-order — is a label attached to one specific value of the slot count k; without first computing k correctly, no order-based name can be assigned with confidence.

A Prerequisite Check for Tensor Operations

Verifying that two tensors have matching slot counts is the minimal prerequisite before checking whether an operation such as addition is even defined between them, since tensors of different slot count cannot be added component-wise regardless of any other property they might share.