2.18.5 Tensor Complex Conjugation Awareness
Tensor Complex Conjugation Awareness examines conjugation's role in tensor algebra, preserving symmetry and defining dual spaces in mathematical contexts.
Tensor Complex Conjugation Awareness is the requirement, built into the handling of tensors over a complex vector space, that every operation performed on a tensor track whether and how complex conjugation acts on its arguments, its components, and its resulting output. Because complex conjugation is an operation with no counterpart in real tensor algebra, a computation that ignores it silently produces the wrong object whenever conjugation is actually present in the definition of the tensor being manipulated, such as a Hermitian form, a conjugate-linear map, or a tensor built from a mix of V and its conjugate space \bar{V}. Conjugation awareness is therefore not a separate algebraic structure by itself but a discipline: a set of rules for propagating the conjugation status of each tensor slot correctly through sums, products, contractions, and changes of basis, so that expressions remain well-defined and consistent with the algebraic identities that genuinely hold over C.
The Conjugation Operation on a Complex Vector Space
Conjugation as a Map
Complex conjugation on a complex vector space V equipped with a chosen real structure sends a vector v to \bar{v}, satisfying \overline{\alpha v} = \bar{\alpha}\,\bar{v} for every complex scalar \alpha, and \overline{v + w} = \bar{v} + \bar{w}. This map is conjugate-linear (also called antilinear), not linear, since it does not commute with complex scalar multiplication in the ordinary sense; this single fact is the root cause of every extra bookkeeping rule that conjugation awareness must enforce.
Conjugation Extended to Tensors
Conjugation extends from vectors to tensors slot by slot: if T is built from arguments in V and V*, its conjugate \bar{T} is the tensor obtained by conjugating every scalar coefficient of T when expressed in a fixed basis, equivalently by applying the vector-level conjugation map to each tensor factor. A tensor is called real when it equals its own conjugate, T = \bar{T}, generalizing the notion of a real number as a complex number fixed by conjugation.
Why Awareness Must Be Explicit
Conjugation Does Not Commute Freely With Multilinear Operations
Because conjugation is antilinear, it does not pass through a complex multilinear operation the way an ordinary scalar or linear map does. For a complex bilinear form B, conjugation awareness requires distinguishing \overline{B(v, w)} from B(\bar{v}, \bar{w}); these coincide only when B has real coefficients in a real basis, and are not equal in general for an arbitrary complex bilinear form.
in general, and conjugation-aware manipulation must track which of these two distinct expressions is actually being computed at each step.
Sesquilinear Forms as the Primary Motivation
The Hermitian inner product h(v, w), linear in one argument and conjugate-linear in the other, is the most common source of conjugation-sensitive tensor manipulation. Awareness here means recognizing that swapping the two arguments of a Hermitian form does not return the same value but its conjugate, h(w, v) = \overline{h(v, w)}, a rule that has no analogue for ordinary symmetric bilinear forms and must be applied correctly whenever such a form is contracted or combined with other tensors.
Rules Governing Conjugation-Aware Contraction
Contracting Conjugated and Unconjugated Slots
A conjugation-aware contraction only pairs a covariant slot with a contravariant slot of matching conjugation status; contracting an unconjugated index against a conjugated one is not a well-defined tensor operation unless the underlying pairing has been explicitly extended to allow it, as happens deliberately in the construction of Hermitian metrics.
Conjugation Under Tensor Products
The conjugate of a tensor product distributes over the factors, \overline{S \otimes T} = \bar{S} \otimes \bar{T}, which must be applied consistently whenever a composite tensor built from several factors is conjugated as a whole rather than factor by factor, to avoid inadvertently conjugating only part of the expression.
Consequences for Real and Hermitian Tensors
Real Tensors as Conjugation-Fixed Points
A tensor T satisfying T = \bar{T} is real, and conjugation awareness explains why this condition, rather than merely having real-looking numerical entries in one particular basis, is the correct basis-independent criterion for a complex tensor to represent a genuinely real geometric object.
Hermitian Tensors and Self-Adjointness
A tensor with one unconjugated and one conjugated covariant slot, h_{i\bar{j}}, is called Hermitian when h_{i\bar{j}} = \overline{h_{j\bar{i}}}, and conjugation awareness is precisely what allows this identity to be checked and preserved under changes of basis, since the transformation of the barred index must be conjugated relative to the transformation of the unbarred one.
Diagrammatic Summary
The diagram illustrates that conjugating an argument before applying a complex multilinear operation generally yields a different result than conjugating the output afterward, the exact discrepancy that conjugation awareness is designed to track and control.