In talking about the meaning of phrases and sentences, so far we’ve focused on the literal meaning of words and morphemes—their denotations—and how those literal meanings compose with one another in predictable ways. The study of these types of meaning is the domain of semantics. But there are other aspects of meaning that can’t be studied or understood in these terms, things we communicate not with the literal meanings of our words and how they’re arranged, but instead by the way we choose to say them, or the social context of our communication. These types of meanings are the domain of pragmatics, the study of meaning in its social or communicative context.
Probably all of us have had the experience of having our words misinterpreted, or taken “out of context”. This often happens even if someone definitely understood our literal words — they may simply have misread our intentions or goals. This type of misunderstanding is precisely the kind of thing we might investigate in pragmatics — both what we intend to communicate, and how someone else might interpret our communications.
In this unit we focus on one particular type of pragmatic reasoning, the the calculation of conversational implicatures on the basis of what are known as Gricean Maxims—these maxims were proposed by the philosopher H.P. Grice in a (1975) paper that proposed that in conversation we adopt a Cooperative Principle when interpreting what people say.
Terminology: Entailment, Presupposition, and Implicature
Before discussing the Cooperative Principle and the individual maxims, let’s introduce some terminology that will help us talk about the logical relationships between sentences—the different ways in which we can draw conclusions.
Consider the following pair of sentences:
- Jennice and Alice have both read War and Peace.
- Jennice has read War and Peace.
If the first of these sentences is true, the second one also has to be true! Put another way, if the second sentence is false, then the first sentence is also definitely false. This is the relationship of entailment.
- For two propositions (i.e. things that can be true or false) P and Q, P entails Q if whenever P is true, Q must also be true.
Entailment is technically a semantic relationship rather than a pragmatic one, but it’s useful to have it in mind to contrast two other relationships with: presupposition and implicature.
Consider next the following two sentences:
- Nadim’s brother is visiting.
- Nadim has a brother.
This might seem to be another case of entailment, but it works a little bit differently. If the first sentence is true, the second one also has to be true. But if the second one is false—if Nadim doesn’t have a brother—then it’s not just that the first sentence is false, it seems like we can’t even really interpret the first sentence. Here we say that the first sentence presupposes the second one.
Here’s another pair of sentences where the first sentence presupposes the second one:
- Lou stopped smoking.
- Lou used to smoke.
If someone asks you Have you stopped smoking in the last year? and you never smoked, you couldn’t answer “yes” or “no”—instead you might say “Hey, wait a minute! I never smoked!” This temptation to say something like Hey wait a minute! is a sign of a presupposition that isn’t satisfied.
- For two propositions P and Q, P presupposes Q if Q has to be true for P’s truth or falsity to be evaluated.
Finally this brings us to implicature, which is the relationship most relevant for our discussion of Gricean maxims.
Consider a final pair of sentences:
- Marie has two cats.
- Marie has exactly two cats.
If someone said to you “I have two pet cats.”, in most contexts you would assume that they didn’t have 10 cats—if it turned out that they did have 10 cats, you’d feel that they’d misled you somehow. But there’s nothing about the first sentence in the pair above that logically entails that Marie doesn’t have more than two cats.
Grice (1975) introduced the term implicature for the relationship between the first and second sentences in this pair.
- For two propositions P and Q, P implicates Q if a listener would infer Q on the basis of someone saying P, despite P not entailing or presupposing Q.
Implicatures, unlike presuppositions or entailments, are cancellable—that is, you can negate them without contradicting yourself or saying something infelicitous.
- Marie has two cats, but (in fact) she has ten cats. (implicature → cancellable)
- #Lou stopped smoking, but they didn’t used to smoke. (presupposition → not cancellable)
- #Nadim’s brother is visiting, but Nadim doesn’t have a brother. (presupposition → not cancellable)
- #Jennice and Alice have both read War and Peace, but Jennice hasn’t read War and Peace. (entailment → not cancellable)
Grice distinguished two types of implicatures:
- Conventional implicatures: triggered by specific words
- Conversational implicatures: calculated based on the Cooperative Principle / specific maxims
We will be mostly concerned with conversational implicatures. Just for illustration, though, an example of a word that triggers a conventional implicature in English is the coordinator but. Consider the contrast between the following two sentences:
- Ruowen likes chocolate ice cream and Helen likes vanilla ice cream.
- Ruowen likes chocolate ice cream but Helen likes vanilla ice cream.
Logically speaking, both and and but mean the same thing—both of these sentences are true only if it’s true that Ruowen likes chocolate ice cream AND true that Helen likes vanilla ice cream.
Grice observed that the coordinator but implies that there’s a contrast between the two clauses, though, or that it’s somehow surprising to assert the second one. This is the conventional implicature of using but instead of and.
Gricean Maxims and the Cooperative Principle
What are we trying to do when we have a conversation with someone? There probably isn’t any one thing that we’re always trying to do, but often part of what we’re trying to do is exchange information.
Some of the things we communicate are not part of the logical or literal meaning of our words. For example, consider the following exchange (adapted from Grice 1975):
- A: How does your friend like working at the bank?
- B: Oh, pretty well. They like their colleagues, and they haven’t been sent to prison yet.
What does B mean by saying their friend hasn’t been sent to prison yet? They could mean a number of things: maybe B is given to telling jokes, or maybe they mean that their friend isn’t usually trustworthy, or maybe they mean that if you didn’t like working at a bank you’d steal money.
The actual meaning conveyed will depend on the context: what A and B both know, their relationship to one another and to B’s friend, and other factors. Whatever B means, though, it’s clear that by saying “they haven’t been sent to prison yet”, B is conveying something more than just the literal meaning of their words! Many many people haven’t been to prison! Why is B bringing it up?
The meaning conveyed by a utterance based on these kinds of considerations is its conversational implicature: implicatures that arise from the structure of discourse or conversation, based on our understanding of how communication works.
Conversational implicatures arise in the context of a general Cooperative Principle for communication. The idea of this principle is that we assume, when we speak to people, that we are working towards a common goal (or common goals), and we can interpret what people say in light of that. So in the mini dialogue above, A can assume that there was some reason that B mentioned that their friend hadn’t been sent to prison yet—they didn’t simply add a true but irrelevant statement to the conversation for no reason!
Under the general heading of the Cooperative Principle, there are four more specific maxims that Grice proposed:
- Maxim of Quantity
- Maxim of Quality
- Maxim of Relation
- Maxim of Manner
In the following sections we will review how each of these maxims works when it is successful, before turning to two ways in which someone can fail to follow the maxims in conversation: either by violating a maxim or by flouting one.
If we violate a maxim, then we simply fail to follow it. At best, violating a maxim results in being a confusing or uncooperative conversationalist. At worst, violating a maxim involves lying or being intentionally misleading.
If we flout a maxim, by contrast, we blatantly fail to follow it—we aim to communicate something precisely by making it very obvious that we have chosen not to follow the cooperative principle, and trusting that our audience will draw the intended conclusions.
Maxim of Quantity
The maxim of quantity states:
- Make your contribution as informative as is required.
- Do not make your contribution more informative than is required.
To follow this maxim, we make the strongest claim that’s both compatible with the facts and relevant in context.
For example, consider the following conversation:
- A: Does Elspeth have any siblings?
- B: Yes, she has a sister.
When hearing B’s response, A assumes that B is fully answering the question—that is, that B is being as informative as possible. So A would naturally assume that Elspeth has exactly one sister, and doesn’t have any brothers.
If it turned out that Elspeth has two sisters and a brother, A would feel that B had misled them—this would be an example of violating the maxim of quantity.
Changing the context can change how we calculate this implicature, though. Suppose that A needs to borrow a car in order to run an errand, and the following conversation ensues:
- A: Does Elspeth have a car I could borrow?
- B: Yes, she has a car.
In this context A will conclude that Elspeth has at least one car. Even if it turns out that Elspeth has two cars, A won’t feel like B misled them—because the second sub-maxim above says that you shouldn’t be more informative than a conversation requires, and in the relevant context all A needs to know is whether there’s a car they can borrow.
Flouting the maxim of quantity can be done in a few different ways! Grice gives the example of a reference letter for a job as a Philosophy professor that says, in its entirety:
- “Dear Sir,1 Mr. X’s command of English is excellent, and his attendance at tutorials has been regular. Yours, etc.” (p. 52)
This letter is ostentatiously much shorter than a reference letter would usually be, and so gives rise to the implicature that there is nothing else that the writer can say about Mr. X that would be positive.
A possibly more subtle example of flouting the maxim of quantity might be something like the following:
- Student: When is Assignment 2 due?
- Professor: You can find that information in the syllabus, which is posted on the course website.
In this exchange, the professor hasn’t actually provided an answer to the student’s question—in that sense it is an uncooperative response. The professor intends to communicate that the student should be able to answer their own question on the basis of information available to them. (The professor’s response probably also involves flouting the maxim of relevance, since they have not directly answered the question asked.)
Maxim of Quality
The maxim of quality states:
- Do not say what you believe to be false.
- Do not say that for which you lack adequate evidence.
In some ways the first of these points is the most basic maxim for the Cooperative Principle: communicating in good faith seems to require that we are—or at least try to be—truthful.
The second point—don’t say that for which you lack adequate evidence—is a bit harder to judge, and what counts as “adequate evidence” varies a great deal from context to context.
Violating the maxim of quality involves lying—intentionally saying things that are untrue—or else saying things that you don’t have enough evidence for.
If your housemate asks you what day garbage is being collected this week, and you can’t really remember but you think it might be Tuesday or Wednesday, you would be violating the maxim of quality if you confidently replied: “Garbage pickup is definitely Wednesday this week.”
Flouting the maxim of quality usually involves irony or sarcasm. For example, consider the following mini-dialogue between a child on a road-trip and their parent:
- Child, asking for the 20th time: Are we there yet?
- Parent, fed up with answering: Nope, we’re just going to keep driving in this car for the rest of our lives.
In this case the parent doesn’t intend their child to take their words literally; they’re flouting the maxim of quality to convey an implicature that the question was unwelcome.
Metaphors or idioms are also cases of flouting the maxim of quality! If I say a scarf is as light as a feather, this is not literally true—but I don’t intend for it to be taken as true!
Maxim of Relevance
The maxim of relevance states:
- Be relevant.
The idea behind this maxim is that when we converse, we shouldn’t introduce irrelevant topics—we try to stick to the topic of conversation, and we assume that our contributions will be interpreted in that light.
Consider the following exchange:
- A: Are you visiting family this weekend?
- B: I have a term paper due on Monday.
A natural interpretation of this exchange is that B is saying that they do not plan to visit family this weekend, and that the reason is that they have to work instead.
But this interpretation is an implicature, because if we think only about the literal meaning of B’s words, this interpretation is a bit mysterious—B doesn’t actually directly answer A’s question, but introduces new topic that doesn’t have anything to do with travel or families.
If we assume that B does intend to be relevant, though, we can explain the implicature: for the term paper to be relevant to the question about travel, it must be that working on the paper controls whether B is able to travel to visit family.
Indeed, suppose we know that B finds it easier to write term papers at home for some reason. In that case we might interpret their statement above as meaning that they do plan to visit family. This illustrates the type of context dependency that’s typical of conversational implicatures!
Violating the maxim of relevance means making irrelevant contributions. You might do this because you’re absent-minded, or because you aren’t actually paying attention to what the other person is talking about, but you can also violate the maxim of relevance more subtly. Consider a slightly different dialogue:
- C: Are you free to hang out this weekend?
- D: I have a term paper due on Monday.
Imagine this dialogue uttered in a context where D does actually have time to hang out, but for whatever reason doesn’t want to spend time with C. Assuming D does have a term paper due on Monday (and therefore is not violating the maxim of Quality), their response would violate relevance: they’re saying something true but irrelevant, in the hopes that C will draw the (incorrect) implicature that D doesn’t have time to hang out because of the time needed to work on the term paper.
Flouting the maxim of relevance involves saying something obviously irrelevant, often to communicate that you want to change the topic of conversation.
For example, if a conversation starts getting awkward and you interject by saying: “How about that hockey game last night?” (when nobody had been talking about sports, never mind about hockey), then you would be flouting the maxim of relevance in the hopes that your audience would understand that you were trying to convey: “Can we please talk about something, anything, else?”
Maxim of Manner
The maxim of manner states:
- Avoid obscurity of expression. (That is, don’t use words or phrases that are hard to understand.)
- Avoid ambiguity.
- Be brief.
- Be orderly.
This relates not to the content of what you say, but the way you express yourself.
It is easiest to discuss each of these sub-maxims in turn, because they have slightly different effect in conversation.
Avoid obscurity of expression
To follow this maxim, we try to use words and expressions that our audience understands. For example, the terminology you would use when explaining a topic in a university class is different from the terminology you’d use if talking about the same thing to a Grade 1 class!
Violating this sub-maxim involves using words and phrases your audience doesn’t know. We often do this accidentally—for example, your linguistics professor might accidentally use a grammatical term that they haven’t defined in class—but if you use a word that you suspect your audience will misinterpret, with the goal of misleading your audience, that would also be a violation of this sub-maxim.
Flouting this sub-maxim would involve intentionally using words that you don’t expect your audience to understand—in most contexts, this would be with the goal of conveying something like: “I’m smarter / know more than you do”, which is pretentious and rude.
To follow this sub-maxim, we try to avoid saying things that can reasonably be interpreted in more than one way.
It’s very easy to violate this sub-maxim accidentally, because often you don’t see the ambiguity in something you say until it’s pointed out to you! But again, you can be intentionally ambiguous in the hopes of misleading people—this is an uncooperative way of talking.
Flouting this sub-maxim often happens in certain kinds of jokes, as in the following:
- A man walks into a bar. Ouch!
This joke turns on two things: 1. being familiar with the common joke set up: “Someone walks into a bar.” and 2. intentionally using the other meaning of the ambiguous word bar.
To follow this sub-maxim, we avoid going on at great length when a shorter statement would do.
Violating this sub-maxim involves saying or writing something much longer than is needed.
Flouting this maxim is more subtle. One example is avoiding a single word and instead using a long paraphrase, as in:
- What did you have for dinner last night?
- Well, we combined all the ingredients listed in a recipe for risotto milanese, in the indicated order, and the result was edible.
By using this long paraphrase, instead of saying “We made risotto milanese.”, the second speaker’s response gives rise to the implicature that the recipe didn’t turn out as intended, or wasn’t very good.
To follow this sub-maxim, we list or relate things in an order that makes sense. For example, when telling a story, we usually start at the beginning and then relate events in the order they happened in.
Violating this sub-maxim can be very confusing, as you’ll know if you’ve ever had to interrupt someone for clarification about the order of events in a story they’re telling!
Flouting this sub-maxim is not something we would do very often. A possible example might be intentionally relating events out of order when writing a fictional story, to convey something about the mood or the narrator’s state of mind.
Grice, H. P. (1975). Logic and conversation. In Syntax and Semantics 3: Speech acts, ed. Cole et al. (pp. 41–58). Brill.
- This is a now–outdated greeting for a formal business letter when you don’t know the name of the person who will receive the letter. It would now be rude to assume that the recipient would be a man. ↩