Analyticity and Logicality in Semantic Constraints

Logic is often said to be “formal” in some sense. This is generally taken to mean that logical inferences, and therefore valid logical arguments, are those where the conclusion follows based on the form of the constituent sentences alone. Catarina Dutilh Novaes (2011) and John MacFarlane (2000) cite this notion as one of the ways we can think about the formality of logic. More specifically, this notion of formality is often taken to mean that logic is formal because it abstracts from the particular attributes and identities of the objects involved in the arguments. This is intuitive to most people who have taken an introductory logic class; logically valid arguments do not depend on facts about the objects referred to by the terms in the argument. This is in fact the position taken by many philosophers of logic (Sher (1991, 2008), Tarski (1986), and MacFarlane (2000)). Gil Sagi (2014) refers to this principle as one of the tenets of formality; I will refer to this principle throughout as F1.