The possibility of iron and chromium reduction in the composition of complex oxides by gaseous carbon monoxide and solid carbon was studied in a laboratory resistance furnace with a graphite heater. Holding only a CO atmosphere does not reduce chromium and iron in both chromites and silicates. In a mixture with solid carbon, chromium and iron were completely reduced in chromites, and iron precipitated in an iron–magnesia silicate matrix in the part having undergone hydration in nature and dehydration upon heating. Continuous carbide shells, which exactly reproduce a surface relief, appear on the surfaces of solid oxide and carbon in a mixture of solid reagents. Similar shells consisting of a carbon-free FeCr alloy form on the surfaces of chromite crystals disseminated in a silicate phase. These results do not correspond to the concepts of both direct and indirect reduction, and they are analyzed in terms of electronic reduction theory developed by us. According to this theory, a low-temperature plasma appears in a mixture of solid reagents due to thermionic electron emission from the carbon surface. As a result, charged particles, namely, electrons and carbon cations, appear near the reagent surfaces. Their interaction with oxide anions causes the formation of CO and anion vacancies, which are charged by two electrons and are scattered in the anion sublattice of oxide. The saturated vacancy solution in the anion sublattice of oxide decomposes to form anion void at the site where the chemical potential of free electrons reaches the Fermi level of certain cations. These cations in anion void regroup to form a metallic phase. The high reducibility of solid carbon is explained by the exponential temperature dependence of the number of electrons emitted by it. This dependence takes place when the plasma flows around the surfaces of both reagents. The formation of carbide shells on the reagent surfaces is caused by the secondary interaction of carbon with a reduced metal on the oxide surface or of oxide lattice fragments with carbon on the corrosion resistance surface. The formation of carbide shells “poisons” the reagent surface and retards and stops reduction. This process renews after breaking the shells upon melting and flowing carbides. The detected laws manifest themselves during both direct and indirect reduction, make the concepts of both mechanisms closer, and eliminate the theoretical discrepancies observed between them in practice.