He was interested in answering the question: what do we mean when we say something? And how can we mean things that aren't explicitly in our speech - things which aren't in the semantics?
Implicatures
Grice introduces the word implicate to mean, "communicate without saying exactly". It is basically a synonym for imply / suggest / mean / ..., and stops us from having to choose one in technical discussions.Grice divides implicatures in to 2 categories: conversational and conventional.
Conversational implicature
Conversational implicatures are carried by context. So for example, if we have a conversation like this:Q. " Do you want to come for dinner on Sunday? "
A. " I've got a lot of work to do that weekend."
A has not said whether or not they want to come. In fact, they haven't even said that they can't - they have merely implicated it.
The test for implicatures is simple: can you say "...but [not the implicature]" or "...and [the implicature]" after the sentence, and have it be both meaningful and not contradictory?
So using the example above:
"I've got a lot of work to do that weekend ... but I would still like to come" : meaningful
"I've got a lot of work to do that weekend ... and I will not be able to come" : meaningful
"I've got a lot of work to do that weekend ... but I will not have a lot of work" : contradictory
"I've got a lot of work to do that weekend ... and I will have a lot of work": repetitive
So we see that "I have a lot of work" is a proposition of the sentence, which can't be cancelled or reinforced. But "I can't come" is an implicature of the sentence, which can be cancelled or reinforced.
Conventional implicature
Conventional implicatures are ones which are carried by the words of the sentence - the way that language is conventionally used, we assume that these implicatures are true unless told otherwise.His example of an conventional implicature is this:
"He is an Englishman; he is, therefore, brave" implicates that braveness is a consequence of Englishness. But if that turns out to not be true, it may still be true that 'he' is an Englishman, and 'he' is brave. So it is not an intrinsic part of the meaning of the sentence; it is conventionally implicated by it.
I don't particulary like this example myself - if "therefore" doesn't make it explicit, what would? But I suppose you could follow it up by "Not that all Englishmen are brave, of course...", which would cancel the implicature. Do you contradict yourself in doing so? Discuss.
The Co-operative Principle
Conventional implicatures could be learnt along with the rest of word meanings. But how do we recover conversational implicatures from "what is said"?
Grice proposed that all communication obeyed the Co-operative Principle, which he formulated like this:
"Make your conversational contribution such as is required, at the stage at which is occurs, by the accepted purpose or direction of the talk exchange in which you are engaged"
Now, whilst this is pretty good advice, I personally see a few holes in it as an explanation of how human communication actually works. For starters: what is the purpose of conversation with your friends? When is it acceptable to change the subject? How do we account for people who ramble? How do you decide what is required?
But before we get on to more recent theories of pragmatics, let's look a bit more at what Grice thought this meant.
Grice's Maxims
Grice fleshed out his Co-operative Principle in a bit more detail, with the following maxims:
- Quantity
- Say enough
- Don't say too much
- Quality
- Tell the truth
- Don't say things you don't have enough evidence for
- Relation
- Be relevant
- Manner
- Avoid obscurity of expression
- Avoid ambiguity
- Be brief (avoid unnecessary prolixity)
- Be orderly
- Obey the maxims
- Violate the maxims (mislead the listener)
- Opt-out (explicitly say you are disobeying one of them)
- Experience clash (when you can't obey them all at once - e.g. in a situation where you can't both be informative and have sufficient evidence for what you are saying)
- Flout the maxims (blatently but not explicitly disobey them)
Flouting the maxims is supposed to give rise to conversational implicatures.
Objections
There are two objections to Grice's Co-operative Principle. The first is that humans are not necessarily being co-operative when they speak. To what extent our co-operativity sets us apart from animals in our communication is discussed at length in the literature (and I may return to it in a future post).The second objection is that the Maxims don't really follow from the principle; that you whilst you can derive conversational implicatures from the flouting of certain maxims, you can also derive a lot of other things. In short, that this isn't a coherent theory.
This objection centers on Grice's own hesitancy about the maxim "Be relevant". What does this mean? How do we decide what is and isn't relevant?
Sperber and Wilson explore this issue in a book called Relevance Theory, which expands it out to be an entire theory of communication on its own - not just in spoken language, but in all human gestures, with and without communicative intent. Relevance Theory will hopefully be the subject of a future post.
References
Grice, H. P. (1975) Logic and communication http://www.ucl.ac.uk/ls/studypacks/Grice-Logic.pdf