The border war between semantics and pragmatics has an early version in the dispute between Mates and Cavell. While Mates argues for a strict separation between semantic inferences and mere pragmatic regularities, Cavell argues for a “logic of ordinary language”, identifying the commitments following the act of saying something. This answer gives a clue to the contemporary debate between minimalists and contextualists: we may either think that pragmatic inferences are only effective after the proposition is grasped, or think that it is part of the determination of what is said. However, Cavell has also another answer, that seems much less amenable to logical regimentation: we understand what an agent is committed to as we evaluate what is appropriate to do in a given situation. In this latter case, there is no logic to backtrack to the very determination of what is said: the agent is attuned to the complexity of her contexts of action, but it does not lead to a local adjustment of concepts. This is precisely what a semanticist does, according to Predelli: she is attuned to the complexity of human purposes, and represents this contextual sensitivity at points of evaluation of sentences, but does not build this complexity into his semantic system. Predelli is at odds both with minimalists and contextualists in this respect, and we can see the specificity of his position as we construe different maps of the theoretical landscape. A bit of historical perspective opens up new ways to see some current discussions.