recent experimental investigation in which a salt containing the unusual charge distribution H+ and Na- was synthesized and characterized prompted us to undertake an ab initio theoretical investigation. In the salt synthesized, the H+ is bound to the nitrogen center of an amine and the Na- alkalide is "blocked" from approaching the protonated amine site by steric constraints of a cage structure. Although one expects that the Na- would deprotonate an unprotected R3N-H+ cation, we decided to further explore this issue. Using extended atomic orbital basis sets and Moller-Plesset and coupled-cluster treatments of electron correlation, we examined the relative stabilities of the prototype (Me)(3)N + NaH, (Me)(3)N + Na+ + H-, (Me)(3)N-H+ + Na-, and (Me)(3)N-Na+ + H- as well as the ion pair complexes (Me)(3)NH+...Na- and (Me)(3)N-Na+...H-. The primary focus of this effort was to determine whether the high-energy (Me)(3)N-H+...Na- ion pair, which is the analogue of what the earlier workers termed "inverse sodium hydride", might be stable with respect to proton abstraction under any reasonable solvation conditions (which we treated within the polarized continuum model). Indeed, we find that such ion pairs are metastable (i.e., locally geometrically stable with a barrier to dissociation) for solvents having dielectric constants below similar to2 but spontaneously decompose into their constituent ions for solvents with higher dielectric constants. We suggest that amines with large proton affinities and/or metals with weaker MH bond strengths should be explored experimentally.