12.3.5 Tensor Addition Result Type Preservation

Tensor Addition Result Type Preservation is the property guaranteeing that when two tensors of a given type are added under the componentwise rule, the resulting object is itself a tensor of that exact same type, with the same rank, the same variance pattern, and the same dimensionality as the two original operands.

Meaning of Type Preservation

What "Type" Refers To

The type of a tensor is characterized by the ordered pair $$(p,q)$$, where $$p$$ counts the contravariant (upper) indices and $$q$$ counts the covariant (lower) indices, together with the dimension of the underlying vector space. Result type preservation asserts that if $$A$$ and $$B$$ both have type $$(p,q)$$, then:

also has type $$(p,q)$$, with no change in the number of upper indices, the number of lower indices, or the dimension of each index slot.

Transformation Law Verification

Under a change of basis described by a transformation matrix, each tensor's components transform according to the rule dictated by its type. Because $$A$$ and $$B$$ share the same type, they transform by the identical rule, and by linearity of that transformation, the sum of the transformed components equals the transformation of the summed components. This equality is precisely what confirms that $$C$$ transforms as a tensor of type $$(p,q)$$, rather than as some other object.

Why Preservation Matters

Closure of the Operation

Result type preservation is what makes tensor addition a closed binary operation on the set of tensors of a fixed type. Without this property, adding two tensors could in principle produce something that no longer qualifies as a tensor, which would make repeated or chained addition operations meaningless.

Foundation for Algebraic Structure

Because the sum always remains within the same type, the collection of all tensors of type $$(p,q)$$ over a fixed vector space can be treated as a vector space itself. Type preservation is the property that permits this collection to be closed under addition, which is one of the basic vector space axioms.

Enabling Further Tensor Operations

Many subsequent operations, such as contraction, tensor multiplication, or applying further linear maps, rely on knowing the exact type of the tensor being manipulated. If addition could alter the type unpredictably, these downstream operations would no longer have a determinate output.

Illustrative Cases

Same Type, Same Result Type

Adding two rank-2 tensors of type $$(1,1)$$ produces another rank-2 tensor of type $$(1,1)$$:

The result still carries one upper index and one lower index, matching the operands exactly.

Repeated Addition Preserves Type Indefinitely

Because each individual addition preserves type, summing any finite number of tensors of the same type, one after another, yields a final result still of that same type, regardless of how many terms are involved or in what order they are grouped.

Illustration