12.7 Tensor Equality Verification

Tensor Equality Verification is the procedure and set of conditions used to determine whether two tensors represent the same underlying multilinear object, requiring that they share the same type and that every one of their corresponding components, when expressed in the same basis, be identical.

Conditions for Equality

Matching Type

Two tensors $$A$$ and $$B$$ can only be considered equal if they share the same type $$(p,q)$$, meaning the same number of contravariant indices, the same number of covariant indices, and the same dimension in every index. Tensors of different type are automatically unequal, since they do not even occupy the same space of possible objects.

Componentwise Coincidence in a Common Basis

Given that both tensors have the same type, equality further requires that, when expressed in one common basis, every component of $$A$$ matches the corresponding component of $$B$$ at the identical index position:

for every valid combination of upper and lower index values.

Basis Independence of the Verification

Equality Holds in Any Basis Once Verified in One

If equality is verified componentwise in one particular basis, it automatically holds in every other basis as well. This follows because both tensors transform according to the identical rule dictated by their shared type, so if their components agree in one basis, the linear transformation applied to both sides preserves that agreement when passing to any other basis.

whenever $$A=B$$ in the original basis, since the same transformation applied to equal inputs produces equal outputs.

Practical Consequence

This basis independence means that tensor equality can be checked in whichever basis is most convenient for computation, without needing to separately verify the result in every other basis of interest.

Reduction to Zero Tensor Comparison

Equality via Subtraction

An equivalent way to verify tensor equality is to compute the difference of the two tensors and check whether that difference equals the zero tensor:

This reformulation is useful because it reduces the equality check to inspecting whether every component of a single computed tensor vanishes, rather than comparing two separate tensors component by component.

Common Pitfalls in Verification

Comparing Tensors in Different Bases Without Transforming

Attempting to compare the raw numerical components of two tensors expressed in different bases, without first transforming one into the basis of the other, can lead to an incorrect conclusion of inequality even when the tensors are, in fact, the same underlying object.

Overlooking Type Mismatch

Comparing only a subset of components, or failing to check that both tensors have the same number and placement of indices, can also produce an unreliable equality verification, since components might coincidentally match in value while belonging to structurally different tensors.

Illustration