Conventional implicatures are meaning contributions from specific words (like 'but,' 'therefore,' 'moreover') that are non-truth-conditional and independent of conversational context. They differ from conversational implicatures by being conventional and typically non-cancellable.
From your study of Grice's conversational implicatures and the cooperative principle, you know that much of what we communicate exceeds what we literally say. Conversational implicatures arise because listeners assume speakers are being cooperative — relevant, truthful, clear, and appropriately informative — and then infer what the speaker must have meant beyond the literal content. The key feature of conversational implicatures is that they are cancellable: a speaker can retract the implicature without contradiction ("I finished some of the work — in fact, I finished all of it"). Now consider a different kind of extra meaning, one that is built into specific words by linguistic convention rather than generated by contextual reasoning.
Conventional implicature (Grice's term) refers to meaning components that are attached to specific expressions by convention but that do not contribute to truth conditions. The classic example is "but." Compare these two sentences: "She is poor and she is honest" vs. "She is poor but she is honest." Both are true under exactly the same conditions — in any world where she is both poor and honest. The word "but" doesn't change when the sentence is true. Yet the two sentences clearly communicate something different: "but" conventionally signals a contrast or tension between the two conjuncts — an implicit message that the combination is surprising or that it runs against an expected connection. That contrast-meaning is the conventional implicature.
Other conventional implicature triggers include "therefore" (which signals that what follows is a conclusion), "moreover" and "furthermore" (which signal additive or intensifying relations), "even" (which signals that the associated proposition is surprising or at an extreme), and "yet" (which signals temporal or adversative contrast). Notice that all of these carry their extra meaning regardless of conversational context — you don't have to calculate whether the speaker is being maximally informative to recover the contrast or causal meaning. It's baked into the word.
The critical difference from conversational implicatures is cancellability (or its absence). A conversational implicature can be cancelled without contradiction: "Some students passed — in fact, all of them did." But you cannot cancel a conventional implicature without incoherence: "She is poor but honest — I don't mean to imply any contrast or tension between those facts." This sounds contradictory or at minimum bizarre, because the contrast-meaning is part of what "but" conventionally contributes. The non-cancellability is the diagnostic signature of conventional implicature.
This matters for semantic theory because it forces us to distinguish multiple layers in what an expression contributes: its truth-conditional content (what makes it true or false), its conventional implicature (non-truth-conditional but conventionally encoded), and its conversational implicature (pragmatically derived from context). Understanding where each layer of meaning comes from — convention, logic, or pragmatic calculation — is central to both the philosophy of language and to linguistic semantics, and the conventional implicature concept precisely delineates the boundary between semantic encoding and pragmatic inference.
Topics in reflective domains aren't scored by quiz answers. Read, reflect, and mark when you've thought it through.