16.4.5 Tensor Antisymmetric Terminology Boundary
The Tensor Antisymmetric Terminology Boundary defines limits in algebra where antisymmetric properties govern tensor behavior and classification.
Tensor Antisymmetric Terminology Boundary is the precise line separating where the words "antisymmetric" and "alternating" may be used interchangeably from where they denote genuinely different conditions, marking the exact circumstances under which conflating the two terms produces an incorrect statement.
Where the Two Terms Coincide
The Standard Field Setting
Over the real numbers, the complex numbers, or any field whose characteristic is not 2, the two conditions below are logically equivalent for any multilinear tensor T:
Within this common setting, most references treat the boundary as effectively absent, and the terms are used as synonyms for the same underlying object.
Derivation Showing the Equivalence
Alternating implies antisymmetric by expanding T(...,u+v,...,u+v,...) = 0 bilinearly and canceling the two identical terms; antisymmetric implies alternating by setting u = v in the antisymmetric law, giving T(...,v,...,v,...) = −T(...,v,...,v,...), which forces the value to zero provided 2 ≠ 0 in the field.
Where the Boundary Becomes Real
Characteristic 2 Fields
The terminology boundary becomes substantive precisely in fields of characteristic 2, such as 𝔽₂, where the equation 2T = 0 is automatically true for every T and carries no information:
Here, "antisymmetric" (T(u,v) = −T(v,u) = T(v,u)) reduces to ordinary symmetry and no longer implies vanishing on repeated arguments. "Alternating" (T(v,v) = 0) remains a strictly stronger, separate condition. In this setting, every alternating tensor is antisymmetric, but not every antisymmetric tensor is alternating.
Why the Convention Still Uses Both Words
Because most applied contexts (geometry, physics, standard linear algebra) work over fields of characteristic 0, the terminology boundary rarely surfaces in practice; both words persist in common usage as a historical artifact, with "antisymmetric" favored in physics and matrix contexts and "alternating" favored in abstract algebra and differential geometry contexts, even though they describe the same condition there.
A Second Boundary: Rank Restriction
Antisymmetric Bilinear Forms versus General Alternating Tensors
A further terminology distinction exists at rank 2: some sources reserve "antisymmetric" strictly for rank-2 bilinear forms (matrices satisfying Tᵀ = −T) and reserve "alternating tensor" for the general multilinear notion at arbitrary rank k:
Under this narrower convention, "antisymmetric rank-3 tensor" would be viewed as loose terminology for what should properly be called an "alternating rank-3 tensor," even though the underlying mathematical content is identical.
Boundary with Related but Distinct Terms
Skew-Symmetric as a Synonym, Not a Boundary Case
"Skew-symmetric" is used purely as an alternative name for "antisymmetric" in matrix contexts, with no terminology boundary between them; the two words describe exactly the same condition Tᵀ = −T with no field-characteristic subtlety attached.
Contrast with "Totally Antisymmetric"
"Totally antisymmetric" is sometimes used to emphasize that the antisymmetry condition holds for every pair of indices simultaneously (as opposed to only some pairs in a mixed-symmetry tensor); this phrase reinforces rather than crosses the terminology boundary, since a rank-k alternating tensor is always totally antisymmetric in all k of its slots.