Chemists view most if not all ionic crystals as composed of autonomously stable negative and positive ions. These building block ions can be singly or multiply charged (e.g., NaCl(s) consists of Na+ and Clions, and MgSO4(s) contains Mg2+ and SO4 ions), but they are expected to be stable to the extent that they can be characterized in the laboratory. Dividing crystals, melts, and ionic molecules into positive and negative ions is usually performed via the octet rule. For example, K2SO4 salt is considered to be composed of K+ cations and SO4 anions, within which all constituent atoms possess full octets of valence electrons. Within this point of view, chemists interpret the physical, reactivity, optical, and other properties of crystals, melts, and liquid ionic salts in terms of properties intrinsic to the corresponding positive and negative ions. For example, in NH4Cl(s), one interprets Raman spectra in terms of crystal phonons as well as internal N-H vibrations belonging to the NH4 cation. This widely accepted approach in chemistry is called into question when one recognizes that many multiply charged anions that occur frequently in salts do not exist as autonomously stable species in which the extra electrons occupy valence molecular orbitals. How then can one view the composition of an ionic material such as a salt in terms of building blocks that cannot be isolated and thus fully characterized? It is this difficulty that forms a primary focus of the present Account. It is very difficult to experimentally prove or disprove the electronic instability of isolated multiply charged anions. The absence of corresponding signals in mass spectral data does not necessarily mean that the ions do not exist; it could be that, under the experimental source conditions, these anions are not formed. Alternatively, mass spectral peaks cannot prove stability; the species may be metastable yet long-lived enough to survive to the ion detector. The significant increase in the reliability of electronic structure computer techniques allows the question of the instability of multiply charged anions to be addressed using ab initio methods. Moreover, in recent years, substantial progress has been made in the ab initio study of multiply charged species, specifically. Three very good reviews on theoretical and experimental studies have appeared in the last three years.1,2 However, because the reviews were devoted primarily to studies of electronically stable multiply charged anions, many issues related to unstable anions were not included and thus need to be addressed in this work. When we discuss the stability of multiply charged anions, we consider three types of stability. The first is electronic stability of the anion. If An-, at its own optimal geometry, is more stable than the corresponding A(n-1)at the same geometry, we consider Anto be vertically electronically stable. If Anat its optimal geometry is more stable than A(n-1)at its own optimal geometry, we consider Anto be adiabatically electronically stable. In addition, there is the issue of geometrical stability. If Anhas all real vibrational frequencies at its optimal geometry, it is locally geometrically stable. Finally, if Anis more stable than any possible dissociation fragment, it is thermodynamically stable. Clearly, if an anion is thermodynamically stable, it must also be electronically and geometrically stable. Most small multiply charged anions are not thermodynamically stable, although many are electronically and locally geometrically stable and may have large barriers to dissociation or to autodetachment. Such anions can be studied experimentally because they are long-lived. However, many multiply charged anions that are only vertically electronically stable are very difficult to probe experimentally because of their very short lifetimes.