The previous seven chapters largely concerned semantics, apart from some passingattention to the interpretation of original metaphor in Chapter 7. Chapters 8and 9 shift focus to pragmatics, and, following on from Chapter 7, it is worthexplaining more about the semantic–pragmatic distinction at theoutset.

To recap, a language is a code more or less shared by the members of a linguisticcommunity. Semantics attempts to describe the meanings of this code and therelations between the meanings of the items of the code, as the survey inChapters 3 to 6 shows. We compose sentences (messages) out of the items in thiscode, and semantics investigates what the sentences mean. Pragmatics, on theother hand, is about what a speaker means, that is, intends, by the utterance ofa sentence in a particular context.

There are three important ways in which semantic meanings differ from pragmaticmeanings. Firstly, pragmatic meanings are non-conventional: when sentences areuttered in context their conventional meanings may be pragmatically overridden.This means different contexts will produce different pragmatic implications. IfI see one of my twenty-year-old students with a Mickey Mouse pencil case and ask“How old are you now?”, I imply a criticism of their childishtastes rather than asking a real question as I would if I knew it was theirbirthday and uttered the same question. Second, pragmatic meanings arecalculable: they are computed through a process of logical inferencing (see10.3–5). And, thirdly, implicatures are defeasible. If my student replies“I don’t want to grow up too quickly”, as a response to my impliedcriticism of her immature taste, I can deny the implied criticism and say“I was only asking your age”. Semantic meanings, by contrast, areconventional, less variable according to context, do not need calculatingbecause they are simply decoded, and are non-defeasible: if you make astatement, relying largely on coded meanings, you cannot truthfully claim youdidn’t express its semantic meaning (Thomas 1995).