Projective Adaptivism

Abstract Moderate invariantism is the orthodox semantics for knowledge attributions (i.e., sentences of the form ⌜S knows/doesn’t know that Φ⌝). In recent years it has fallen out of favour, in large part because it fails to explain why ordinary speakers have the intuition that some utterances of knowledge attributions are felicitous and others infelicitous (felicity intuitions) in several types of cases. To address this issue moderate invariantists have developed a variety of what I call non-semantic theories (aka error theories) which they claim account for the relevant felicity intuitions independently of moderate invariantist semantics. Some critics have responded by arguing that these non-semantic theories are implausible for one or more of the following reasons: (i) they do not have a basis in empirical data or established theory; (ii) they do not account for all of the relevant felicity intuitions; (iii) they are ad hoc; or (iv) they in fact explain too many felicity intuitions and thus undermine the case for moderate invariantism. I develop a new non-semantic theory––projective adaptivism––that I argue escapes issues (i) to (iv) above.