adjoint operator

The transpose of a matrix swaps rows and columns, but that is a coordinate gimmick. The adjoint is the coordinate-free version: T^* is the operator that lets you move T from one side of the inner product to the other. It is the inner product's own notion of transpose.

Definition: given T : V -> W, its adjoint T^* : W -> V is the unique operator satisfying <T u, v> = <u, T^* v> for all u in V and v in W. Existence and uniqueness come straight from the Riesz theorem: fix v, the map u -> <T u, v> is a functional, so it equals <u, w> for a unique w, and we name that w = T^* v.

In coordinates with orthonormal bases, T^* is exactly the conjugate transpose: (T^*)_{jk} = conj(T_{kj}). Over the reals the conjugate vanishes and T^* is the plain transpose. This is why every fact about transposes — but stated invariantly — really lives at the level of adjoints.

A warning: the matrix-equals-conjugate-transpose identity requires ORTHONORMAL bases. In a skewed basis the adjoint picks up the Gram matrix as a correction. The adjoint is defined by the inner product, not by the coordinates, and only orthonormal coordinates report it cleanly.