15.12 Tensor Symmetric Basis Structure

Tensor Symmetric Basis Structure is the complete description of how a basis for the underlying vector space V gives rise to a corresponding basis for each graded piece Sym^n(V) of the symmetric algebra, together with the properties that make this induced basis well-behaved: linear independence, spanning, and a direct indexing by multi-indices that connects the tensor picture to the polynomial picture. This structure formalizes and organizes, as a single topic, the basis-related facts established piecemeal through the symmetric power basis construction and the multi-index notation used to label its elements.

Understanding the symmetric basis structure as a unified topic clarifies why choosing a basis for V is sufficient to fully coordinate-ize every symmetric tensor of every rank simultaneously, using one consistent scheme, rather than requiring a separate, unrelated basis choice for each rank individually.

Constructing the Basis for a Fixed Rank

From a Basis of V to a Basis of Sym^n(V)

Given a basis e_1 through e_d of V, the induced basis of Sym^n(V) consists of the symmetric products of basis vectors, one element for every multi-index alpha of degree n:

using the multi-index power notation to compactly denote the symmetric product of powers of each basis vector.

Independence From the Order of Combination

Because the symmetric product is commutative and associative, the element e^alpha is unambiguous regardless of the order in which the individual factors e_1^{odot alpha_1} through e_d^{odot alpha_d} are combined, a direct consequence of the algebra role established for the symmetric product.

Linear Independence and Spanning

Independence via the Polynomial Correspondence

The elements e^alpha, for distinct multi-indices alpha of the same degree n, correspond under the polynomial analogy to distinct monomials x^alpha, and distinct monomials are linearly independent as polynomials; this transfers directly to establish that the corresponding tensors e^alpha are linearly independent within Sym^n(V).

Spanning Established via Dimension Matching

The number of multi-indices of degree n equals the count relation C(d, n), exactly matching the dimension of Sym^n(V); since the constructed set is linearly independent and its size equals the dimension of the space, it must also span Sym^n(V), completing the verification that the e^alpha form a genuine basis.

The Basis Across All Ranks Simultaneously

A Single Basis for the Entire Graded Algebra

Collecting the bases e^alpha across every degree n, for every possible multi-index alpha of every possible degree, produces a single basis for the entire symmetric algebra Sym(V), since the graded structure guarantees every element of Sym(V) decomposes uniquely into finitely many homogeneous pieces, each expressible in the basis of its own degree.

Consistency With the Universal Property

This basis realizes concretely the abstract freeness expressed by the universal role of the symmetric algebra: specifying an algebra homomorphism out of Sym(V) reduces, once this basis is fixed, to specifying where each basis vector e_k of V is sent, with the images of every e^alpha determined automatically by multiplicativity.

Coordinates of a General Symmetric Tensor in This Basis

Coefficients as Rescaled Components

Expressing a general symmetric tensor T of rank n in the basis e^alpha assigns to each multi-index alpha a coefficient equal to T's independent component at the corresponding canonical index tuple, scaled by the multinomial factor n! divided by alpha!, reflecting the combinatorial relationship between the tensor's raw component value and the coefficient appearing when T is expanded in this particular basis.

Practical Use for Explicit Computation

Because this basis aligns exactly with the monomial basis under the polynomial analogy, expressing a symmetric tensor in the symmetric basis structure and expressing its associated homogeneous polynomial in the standard monomial basis are two descriptions of literally the same coefficient data, making this basis the natural default choice whenever symmetric tensors are to be manipulated using ordinary polynomial arithmetic.

Relation to the Broader Symmetric Tensor Framework

Basis Structure as the Computational Backbone

While the independent symmetric component structure addresses how many degrees of freedom a symmetric tensor has and how to organize them, the symmetric basis structure specifically supplies the concrete vectors, the e^alpha, that give those degrees of freedom a coordinate meaning within the vector space Sym^n(V), making the two structures complementary descriptions of the same underlying reduction from the full tensor space to its symmetric subspace.

Dependence on the Choice of Underlying Basis

Unlike the abstract, basis-independent characterizations of Sym(V) as a graded algebra, a quotient of the tensor algebra, or the solution to a universal mapping problem, the symmetric basis structure is inherently tied to a specific choice of basis for V, and changing that basis produces a different, though equally valid, basis for each Sym^n(V), related to the original by the same transformation rules that govern basis changes in ordinary linear algebra.