Non-reactive pure metals do not wet ionocovalent oxides like alumina and silica. It is believed generally that a strong improvement of wetting can be produced by alloying elements presenting a high reactivity with the oxide substrate. In this study, existing data on contact angles are reviewed and analysed in order to answer the following three questions related with the modelling of reactive wetting: (1) Does a relation between wettability, which is a 2D phenomenon, and bulk reactivity (3D) really exist? (2) If so, what are the mechanisms by which an interfacial reaction can modify wettability? (3) What are the criteria for selection of alloying elements able to act as wetting agents? From the answers to these questions, a new description of the relationship between reactivity, wettability and adhesion in metal/oxide systems is given.