The clonal selection theory (CST) is the centrepiece of the current paradigm used to explain immune recognition and memory. Throughout the past decades, the original CST had been expanded and modified to explain new experimental evidences since its original publication by Burnet. This gave origin to new paradigms that govern experimental immunology nowadays, such as the associative recognition of antigen model and the stranger/danger signal model. However, these new theories also do not fully explain experimental findings such as natural autoimmune immunoglobulins, idiotypic networks, low and high dose tolerance, and dual-receptor T and B cells. To make sense of these empirical data, some authors have been trying to change the paradigm of immune cognition using a systemic approach, analogies with brain processing and concepts from second-order cybernetics. In the present paper, we review the CST and some of the theories/hypotheses derived from it, focusing on immune recognition. We point out their main weaknesses and highlight arguments made by their opponents and believers. We conclude that, until now, none of the proposed theories can fully explain the totality of immune phenomena and that a theory of everything is needed in immunology.