It is commonly assumed that learning in both animals and humans represents a particular, since essentially non-genetic way to assimilate information from the environment. This view is even held by many theoreticians who, otherwise, repeatedly stress the importance of the genetic adaptation process by referring to the general validity of the so-called central dogma of molecular biology which forbids any directed instruction of the genome through phenotypic influences. If, however, one takes a closer look at what really happens in a number of most different learning processes, one quite rapidly discovers that they fully obey the central dogma and, associated with it, the mutationist principle of evolutionary theory which prescribes random variation as the sole source of evolutionary change, i.e., information gain. Consequently, learning taken to be a true gain in information is possible only through concrete genetic changes within the germ line, i.e., through non-somatic mutations, and has nothing to do with our common understanding of "learning" as an ontogenetic phenomenon.