A tentative picture for the charging of the sapphire basal plane in dilute electrolyte solutions allows reconciliation of the available experimental observations within a dual charging model. It includes the MUltiSIte Complexation (MUSIC) model and auto-protolysis of interfacial water. The semi-empirical MUSIC model predicts protonation and deprotonation constants of individual surface functional groups based on crystal structure and bond-valence principles: on the ideal sapphire c-plane only doubly co-ordinated hydroxyl groups exist which cause quasi zero surface potential (defined as the potential in the plane of the surface hydroxyl groups) from pH 5 to 7 and rather weak charging beyond (compared to typical oxide behaviour). MUSIC predictions concur strikingly with recently published sum frequency data for the pH dependence of the so-called "ice-like" water band (interfacial water) and contact angle titrations. Zeta potential as well as second harmonic generation data reveal a sharp IEP of around 4 and a negative surface charge at the pristine point of zero charge predicted by the MUSIC model. New zeta-potential data corroborate (i) the low IEP and its insensitivity to salt concentration and (ii) the second harmonic results. We thus establish two groups of conflicting results arising from different techniques. A conventional model of the mineral electrolyte interface such as the MUSIC model is at odds with the negative zeta potentials in the pH range 5 to 7. Therefore an additional charging mechanism is invoked to explain all the observations. Enhanced auto-protolysis of interfacial water is the most probable candidate for this additional mechanism, in agreement with net water orientation observed with sum frequency generation and second harmonic generation. Our phenomenological explanation is further corroborated by the similarity of the zeta potential vs. pH curves of the c-plane with those of hydrophobic surfaces. Additional support comes from infrared spectroscopic data on thin water films on sapphire c-plane samples. Most stunningly, theoretical calculations on basal planes of this kind suggest a 2D water bilayer that makes such surfaces hydrophobic towards further adsorption of water. The proposed dual charging mode approach comprises the MUSIC model for protonation/deprotonation of the surface aluminols affecting the surface potential and the currently advocated enhanced auto-protolysis picture for hydrophobic surfaces controlling the zeta-potential and can explain the available information in a qualitative way. The respective contributions from the two components of this dual charging mechanism may be different for different single crystal cuts of alumina. Thus interplay between protonation/deprotonation of surface functional groups and auto-protolysis of interfacial water will cause the observed zeta potentials and isoelectric points. Repercussions of one mechanism on the other will result in the most favourable interfacial water structure, which can be followed by non-linear optic techniques like sum frequency generation.