True clauses and false connections

Abstract Indicative conditionals—that is, sentences typically, though not exclusively, of the form “If p, (then) q,”—belong to the most puzzling phenomena of language. One of the puzzles that has recently attracted attention of psychologists of reasoning stems from the fact that on the majority of accounts of indicative conditionals, “If p, (then) q” can be true, or at least highly acceptable, even when there is no meaningful connection between p and q. Conditionals without such a connection, dubbed missing-link conditionals, however, often seem very odd. A standard pragmatic account of their oddity rests on an observation that, whenever missing-link conditionals come out as true, these are situations in which speakers are justified in asserting stronger, more informative statements. Asserting a less informative statement is odd because it is a violation of the Maxim of Quantity. This paper reports four experiments that present a challenge to the Gricean explanation of why missing-link conditionals are odd. At the same time, we will argue that these findings can be reconciled with general principles of Gricean pragmatics, if the connection is treated as a part of a conventional, “core” meaning of a conditional.