6.23.1 Tensor Component Count by Dimension

Tensor Component Count by Dimension is the study of how the component count formula N = d^(p+q) behaves as the dimension d of the underlying vector space is varied while the tensor's order p + q is held fixed, isolating the role that dimension specifically plays in the growth of the number of independent scalars needed to specify a tensor. Where the component count relation establishes the formula itself, this topic focuses on its behavior as a function of d — how quickly it grows, what happens at the smallest dimensions, and why dimension and order interact multiplicatively rather than additively in determining total component count.

Fixing Order, Varying Dimension

The Count as a Power Function of d

For a tensor of fixed order n = p + q, the component count

is a power function of d with fixed exponent n. Viewed this way, dimension acts as the base and order acts as the exponent, so doubling the dimension multiplies the component count by 2ⁿ, while doubling the order (at fixed dimension) squares the component count — the two parameters enter the formula in structurally different ways.

Concrete Growth Table for a Fixed Order

For an order-two tensor (n = 2), the component count as d increases is d = 1 → N = 1, d = 2 → N = 4, d = 3 → N = 9, d = 4 → N = 16; the count grows quadratically in d, reflecting the familiar fact that a d × d matrix has d² entries.

Fixing Dimension, Varying Order

Contrast: Order as the Exponent

Holding d fixed instead and varying the order n produces exponential rather than polynomial growth: for d = 3, the sequence of component counts as n = 0, 1, 2, 3, 4 is 1, 3, 9, 27, 81, each term three times the last. This exponential growth in order, contrasted with the polynomial growth in dimension at fixed order, is the key asymmetry that the dimension-focused view of the component count relation is meant to highlight.

Diagram Comparing the Two Growth Regimes

Smallest-Dimension Special Cases

Dimension One Collapses All Orders

When d = 1, the component count formula gives N = 1ⁿ = 1 for every order n, meaning that in a one-dimensional space, tensors of any order have only a single component; this degenerate case shows that the exponential growth in order is entirely a phenomenon that requires d ≥ 2 to be visible, since the base of the power d^n must exceed 1 for increasing the exponent to have any effect.

Dimension Zero and the Trivial Space

The formula N = 0ⁿ is 1 when n = 0 (matching the convention that the trivial vector space still has a single scalar) and 0 for every n ≥ 1, consistent with the fact that a zero-dimensional vector space contains no nonzero vectors from which to form nontrivial tensor components.

Why This Dimensional Perspective Matters

Guiding Feasibility of Explicit Computation

Separating the roles of dimension and order in the component count clarifies which parameter is more responsible for computational infeasibility in a given setting: for tensors of modest order but very large dimension (as in high-dimensional data analysis), the polynomial-in-d growth is the dominant cost driver, whereas for tensors of modest dimension but high order (as in multilinear algebra over small spaces), the exponential-in-n growth dominates instead, and recognizing which regime applies informs the choice of computational strategy.