This paper addresses the issue of the “true” and the “real” in strictly semiotic terms, and does so by comparing our view, both by analogies and differences, with the reflections on “reality” carried out in the fields of physics, astronomy and cosmology — Nobel Prize winner Roger Penrose is here as a reference author. The semiotics of Saussurian derivation, and the sociosemiotics inaugurated by the studies of Émile Durkheim, present a specific epistemological conception and a specific concept of “real” which, if on the one hand has clear ancestry in Kant’s philosophy, on the other elaborates the idea of a social construction of what we call “reality”. Essential stages of the discussion are placed around the concept of “model”, concept of particular richness and ambiguity, since it is necessary to distinguish three fundamentally different types of entities: cultural models, which, although not directly observable, are present in social reality (for instance, the rules of grammar present in the collective subjectivity of the speakers of a language); the models built at the scientific metal level for the purpose of accounting, at least approximately, for the functioning of real semiotic processes (for instance, Propp’s compositional scheme); the model cases, or prototypes, which with their strength and their claimed “authenticity” support the perception that the world of experience is at the same time ordered, real, and endowed with meaning. The logical conclusion of a semiotic path is that “reality” is not a set of entities and events external to us, but a sense–building device in which we are active subjects. What is most surprising is that even a physics scholar reaches very similar conclusions: that reality is not something that exists “out there” but that corresponds to the experience we are building; the universe is, upon closer inspection, a gigantic symbolic processor, a configurator of meanings.