For a linear endomorphism of a finite-dimensional vector space (or a trace-class operator in analytic contexts), the trace is the scalar given by the sum of diagonal entries of any matrix representing the endomorphism in a basis; it is a similarity-invariant equal to the sum of eigenvalues counted with algebraic multiplicity.

Principle

Trace is a linear functional tr: End(V) → k satisfying the cyclic property tr(AB) = tr(BA), is basis-independent, and provides a nondegenerate pairing between endomorphisms and forms in finite dimensions; it behaves functorially with respect to direct sums and tensor products under standard rules.

Demonstration

For the 2×2 matrix [[a,b],[c,d]] the trace is a + d; the trace of the identity on an n-dimensional vector space equals n; nilpotent matrices have trace zero, and the trace of a projection equals the rank of the projection when diagonalizable with eigenvalues 0 and 1.

Misapplication

Assuming trace zero implies nilpotent (false in general), treating the sum of diagonal entries in a non-basis-independent manner, or applying finite-dimensional trace identities to operators that are not trace-class in infinite-dimensional Hilbert spaces.

Consequence

Trace yields characters of representations, linear invariants used to distinguish conjugacy classes, and coefficients of characteristic polynomials (trace equals minus the coefficient of degree n−1 up to sign); it also participates in index theorems and invariants in geometry when extended analytically.

Reversal

Determinant is the multiplicative dual notion: while trace sums eigenvalues, determinant multiplies them; focusing on determinant emphasizes volume-scaling and multiplicative invariants rather than additive summaries.

Boundary

The standard trace is defined for endomorphisms of finite-dimensional vector spaces and for trace-class operators in analysis; over noncommutative rings or for general linear operators of infinite rank a naive diagonal-sum need not exist or be invariant, so extensions require care (Hattori–Stallings trace, von Neumann trace, etc.).

Semantic Tension

Trace competes with related concepts: matrix trace (elementary), categorical trace (abstract duality), and operator traces (analytic regularity); one must specify whether the context is algebraic, categorical, or analytic to avoid conflating definitions and existence criteria.

Synthesis

Trace extracts the canonical additive invariant of an endomorphism — the sum of eigenvalues or diagonal entries invariant under similarity — providing a linear map from endomorphisms to scalars that underlies characters, polynomial coefficients, and many index-type formulas.