Several data accumulated over recent years on the mechanisms underlying the interactions between the brain, hormones and the immune system. These data concern two major avenues of research: the evidence that brain-controlled, behavioral parameters can modulate the response of immunocompetent cells, and an increasing awareness that a number of chemical signals - neurotransmitters, hormones or mediators of immunity - are not, as previously believed, specific of given sets of tissues or of functions, but that, on the contrary, they can be produced and recognized by cellular elements belonging to any of those three systems. There is indeed evidence to indicate that signaling molecules involved in cellular communication are 'banalized': that means that their receptors are liable to be expressed in almost any tissue by a wide variety of cells. This statement, together with the discovery that intercellular regulation is multifactorial - that is, depends at any given time upon messages built up by combinations of signal molecules rather than by isolated transmitters - raises a certain number of theoretical problems as to the manner by which cells extract messages out of an important background noise. In the present paper, some of those theoretical problems will be presented in a summarized form, and their relevance for the interpretation of neuroendocrine or neuroimmunological interactions will be discussed.