Besides depth and temperature, CO2 and H2O, are the two most important variables in stabilizing partial melts in the Earth's mantle. However, despite decades of experimental studies on the roles of these two volatile species in affecting mantle melting, ambiguity remains in terms of the stability, composition, and proportion of volatile-bearing partial melts at depths. Furthermore, the difference in the influence of H2O versus CO2 in production of mantle melts is often inadequately discussed. Here I first discuss how as a function of depth and concentration of volatiles, the peridotite + H2O versus peridotite + CO2 near-solidus melting conditions differ - discussing specifically the concepts of saturation of volatile-bearing phases and how the mode of storage of ‘water’ and carbon affects the near solidus melting relations. This analysis shows that for the Earth's mantle beneath oceans and continents, deep, volatile-induced melting is influenced mostly by carbon, with water-bearing carbonated silicate melt being the key agent. A quantitative framework that uses the existing experimental data, allows calculation of the loci, extent of melting, and major element compositions of volatile-bearing partial melts beneath oceans and continents. How the domains of volatile-bearing melt stability are affected when possible oxygen fugacity variation at depths in the mantle is taken into account is also discussed. I show that trace amount hydrous carbonated silicate melt is likely stabilized at two or more distinct depths in the continental lithospheric mantle, at depths ranges similar to where mid-lithospheric discontinuity (MLD) and lithosphere-asthenosphere boundary (LAB) have been estimated from seismology. Whereas beneath oceans, hydrous carbonated silicate melt likely remain continuously stable from the base of the thermal boundary layer to at least 200 km or deeper depending on the prevailing oxygen fugacity at depths. Hotter mantles, such as those beneath oceans, prevent sampling strongly silica-undersaturated, carbonated melts such as kimberlites as shallower basaltic melt generation dominates. Thick thermal boundary layers, such as those in cratonic regions, on the other hand allow production of kimberlitic to carbonatitic melt only. Therefore, the increasing frequency of occurrence of kimberlites starting at the Proterozoic may be causally linked to cooling and growth of sub-continental mantles through time.