2.7.3 Tensor Linear Combination Finite Sum

Tensor Linear Combination Finite Sum expresses how tensors can be combined through finite sums, forming foundational structures in algebraic operations.

Tensor Linear Combination Finite Sum is the requirement, built into the very definition of a linear combination, that only finitely many weighted tensor terms are ever added together, ensuring that the resulting sum is constructed entirely from the two elementary operations of tensor addition and scalar action, applied a finite number of times, without appeal to any notion of limit or convergence. It is what keeps the linear combination structure a purely algebraic construction, valid over an arbitrary field, rather than an analytic one requiring additional structure on the tensor space.

Definition via Iterated Binary Operations

Building the Sum Term by Term

Let $V$ be a vector space over a field $F$ , and let $T_{1}, \dots, T_{N} \in T_{s}^{r} (V)$ with coefficients $α_{1}, \dots, α_{N} \in F$ . The finite sum

\sum_{k = 1}^{N} α_{k} T_{k}

is defined recursively: for $N = 1$ , it is simply $α_{1} T_{1}$ , and for $N > 1$ , it is the sum of the first $N - 1$ terms, computed by this same rule, added by tensor addition to the final term $α_{N} T_{N}$ .

Only Two Operations Involved

At every stage of this recursive construction, only two operations are used, the scalar action producing each weighted term $α_{k} T_{k}$ , and tensor addition combining the accumulated partial sum with the next term. No operation beyond these two is required to define a finite sum of any length.

Well-Definedness of the Result

Independence from Grouping

Because tensor addition is associative and commutative, the recursive construction produces the same tensor regardless of the order in which the terms are added or how they are grouped into partial sums:

(α_{1} T_{1} + α_{2} T_{2}) + α_{3} T_{3} = α_{1} T_{1} + (α_{2} T_{2} + α_{3} T_{3})

so the summation symbol $\sum$ denotes a single unambiguous tensor, not a value dependent on how the recursive construction happened to be carried out.

No Ordering Dependence

Commutativity further guarantees that reordering the terms of the sum, without changing the set of terms and their coefficients, leaves the result unchanged, so a finite sum may be indexed in any convenient order without altering its value.

Why Finiteness Is Essential

Absence of a Convergence Structure

A general tensor space $T_{s}^{r} (V)$ , defined over an arbitrary field $F$ , carries no topology by default, and therefore no built-in notion of what it would mean for an infinite sum of tensors to converge to a limiting tensor. Restricting linear combinations to finite sums avoids this issue entirely, since finite sums require only the algebraic operations already defined on the vector space, never a limiting process.

Compatibility with Arbitrary Fields

Because finiteness makes no reference to convergence, the finite sum definition of a linear combination applies uniformly whether $F$ is the real numbers, the complex numbers, a finite field, or any other field, whereas an infinite-sum definition would require additional analytic structure not present, or not meaningful, in many such fields.

Finite Sums and the Dimension of the Tensor Space

Sufficiency of Finitely Many Terms

Since $T_{s}^{r} (V)$ is finite-dimensional, of dimension $n^{r + s}$ , every tensor in the space is reachable as a finite sum of at most $n^{r + s}$ basis tensor products, so the finite-sum restriction never prevents any element of the space from being expressed as a linear combination; the dimension itself bounds how many terms are ever actually necessary.

No Loss of Generality

Even in a hypothetical infinite-dimensional generalization of tensor spaces, the finite sum restriction remains the standard definition of a linear combination, since it is precisely finite linear combinations that generate the algebraic span of a set, with any further completion to include limits treated as a separate, additional structure layered on top of the algebraic one.

Practical Consequence for Computation

Termination of Computation

Because a linear combination always involves finitely many terms, evaluating it, whether symbolically or numerically, is guaranteed to terminate after a fixed, predictable number of applications of tensor addition and scalar action, with the number of steps determined entirely by $N$ , the number of terms specified in advance, rather than by any iterative or limiting procedure.