As Jørgensen pointed out in 1966 (Coord. Chem. Rev., 1966, 1, 164), a ligand is to be regarded as ‘innocent’ if it allows the oxidation state of a metal in a complex to be defined. In this respect, the vast majority of ligands are ‘innocent’ and, therefore, ligands that are ‘non-innocent’ have received special attention. Dithiolenes have been regarded as ‘non-innocent’ ligands since it is possible to consider a ligand of this type to be present in a complex as either: (i) an ene-1,2-dithiolate dianion or (ii) a neutral dithioketone. On this basis, the electronic structure of a dithiolene complex can be described by a set of resonance structures, each of which involves the dithiolene in one of the two forms with the oxidation state of the metal centre being adjusted accordingly. The relative importance of these structures is expected to be reflected in the corresponding molecular structure and spectroscopic properties. In this paper we present a theoretical study of the pair of related η⁵-cyclopentadienyl cobalt dithiolene complexes, [CpCo(S₂C₂(H)Ph)] and [CpCo(S₂C₂(H)Ph)(PMe₃)]. Density functional theory calculations successfully predict their different structures and NMR chemical shifts, which we have measured. These wavefunctions have been analysed, particularly in terms of Natural Bond Orbitals and Nucleus Independent Chemical Shifts in an attempt to understand how “innocence” or otherwise is reflected in the experimental data. To this end, a similar analysis is applied to the gold complexes [Au(S₂C₂(H)Ph)₂]⁻ and [Au(S₂C₂(H)Ph)₂].