Learning Systems[Systèmes Apprenants]

Learning is a characteristic of open systems, which contributes to their adaptation to their environment. Learning was initially studied in humans, giving rise to specific research domains of psychology, sociology, and specific disciplines, pedagogy, didactics, educational sciences. The study of processes was then extended to living beings in general, to highlight the function of learning in adaptation. Learning becomes possibly too powerful a term, because we discover that any living organism would be capable of it: even a blob, devoid of a brain, is capable of learning. Systemic approaches make it possible to extend the modeling of these processes to any system capable of storing and processing the information it receives from its environment, to adapt to it or act on it to adapt it to itself. This is how we can speak of learning organizations for groups such as companies or administrations which take advantage of their experience to develop. Or collective intelligence about an ant colony which identifies a resource and marks the route to bring it back to the anthill. Formalization makes it possible to apply processes to machine learning. It allows us to ask the question of whether viruses, bacteria, and other pathogens "learn" to circumvent the barrier measures against them. Yet, remarkably, the founding fathers of cybernetics and systems theory did not deal with learning per se. It is therefore one of the interests of this issue of the Cahiers de Systémique to propose taking up this entry.