12.4.5 Tensor Subtraction Result Type Preservation

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

Meaning of Type Preservation

What Is Being Preserved

The type of a tensor is characterized by the ordered pair $$(p,q)$$, counting contravariant and covariant indices respectively, together with the dimension shared by every index. Result type preservation for subtraction asserts that if $$A$$ and $$B$$ both have type $$(p,q)$$, then:

also has type $$(p,q)$$, with no alteration to the number of upper indices, the number of lower indices, or the dimension of any index slot.

Derivation from Addition Type Preservation

Since subtraction is defined as $$A+(-B)$$, and the additive inverse $$-B$$ necessarily shares the type of $$B$$, result type preservation for subtraction follows directly from result type preservation for addition. There is no need for a separate proof, since subtraction is simply addition applied to a same-type operand.

Verifying the Transformation Law

Consistency Under Change of Basis

Under a change of basis, both $$A$$ and $$B$$ transform according to the identical rule dictated by their shared type. Because this transformation is linear, it commutes with subtraction:

This equality confirms that the difference tensor transforms exactly as a tensor of type $$(p,q)$$, matching both operands.

Why Preservation Matters

Closure Within a Fixed Type

Result type preservation ensures that subtraction is a closed operation on the set of tensors of a fixed type. Without it, subtracting two tensors of a given type could in principle yield something outside that set, which would undermine any further algebraic manipulation relying on a stable type.

Support for the Vector Space Structure

Because the collection of tensors of type $$(p,q)$$ forms a vector space, and every vector space must be closed under subtraction of its elements, this preservation property is what allows subtraction to be a legitimate internal operation on that space rather than an operation that could escape it.

Enabling Further Operations

Operations that follow subtraction, such as further contraction or multiplication with other tensors, require the resulting tensor's type to be known and fixed. Preservation of type under subtraction ensures that such subsequent operations remain well defined.

Illustrative Cases

Same Type Difference

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

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

Chained Subtraction

Because each subtraction preserves type, subtracting a sequence of tensors of the same type one after another produces a final result still of that same type, regardless of how many terms are subtracted.

Illustration