What's the relationship between pragmatics and grammar?

↵ Back to class homepage

Over the course of this class, we have seen that pragmatics works when people violate expectations and conventions (see the "Violating maxims" module), and that people use pragmatic indirectness to accomplish things like being [or not being] polite, maintaining deniability, being convincing, and changing the circumstances around them (see "Benefits of indirect utterances"). Pragmatics is fundamentally creative. Levinson (p. 27, my emphasis added) says pragmatics is "concerned precisely with such mechanisms whereby a speaker can mean more than, or something quite different from, what he actually says, by inventively exploiting communicative conventions."

Of course, if you want to break rules and subvert expectations, you have to first know what those rules and expectations are.

Conventions, creativity, and breaking the rules

If we want to break rules and subvert expectations in a purpose-driven way—i.e., if we want to break rules not just for the sake of breaking rules, but to convey something or to accomplish some special effect—we have to know what the rules and expectations are. We can see this in many aspects of life. For example, comedians often make us laugh by saying or doing things that play on our expectations; they would not be able to do that if they didn't understand our expectations well enough to make us expect one thing and then surprise us by saying something else. In Ali Wong's acclaimed stand-up set Baby Cobra [spoiler alert!], Wong spends an hour presenting herself as a gold-digger looking to trap a rich man in marriage, but then at the end makes a shocking reveal:

And then I meet this dream guy who's, like, way more handsome than me, out of my league, graduated from Harvard Business School... [I] worked hard to trap his ass, got him to propose to me, oh my god, then we got married, all my dreams coming true, and then we got pregnant, and recently we bought our first home together. And, uh, two weeks into the escrow process, I discovered that my beautiful, Harvard-educated husband was $70,000 in debt. And me, with my hard-earned TV money, paid it all off! ....So, as it turns out, he's the one who trapped me.

And it's funny because it usurps what the audience had expected, based on the image of herself that Wong had built up over the preceding hour of the set.

For another example of breaking rules on purpose after knowing what the rules are, think about the experience of learning a new language. I suspect anyone who's learned another language has had the experience of having to "unlearn" rules. Did you ever have some first-year textbook teach you some rule, and then years later you discovered that people don't actually follow that rule in colloquial speech? For example, when I learned Chinese, we were taught to always use a classifier (量词) between a noun and a number or demonstrative. In actual Chinese, though, people often say stuff like 这人 ("this person") rather than 这个人 ("this <classifier> person"). This is, again, an example of a situation where you first learn a rule before you can understand when and why to break the rule.

Or, for one last example, let's consider fiction writing. Terry Bisson's "60 rules for short SF (and fantasy)" gives prospective writers some "rules" for how to write good science fiction stories. Some of them are very specific, like rule #5 (don't use flashbacks), rule #7 (don't use "dialect"), rule #18 (don't use more than two points of view), etc. But if you read much science fiction, you will inevitably find some excellent stories that break one or more of these rules; in fact, for every "rule" here, there is probably at least one great science fiction story that breaks that rule. How is this possible? Look at Bisson's last two rules. Rule #59: "Ignore these rules at your peril." Rule #60: "Peril is the [science fiction] short story writer's accomplice, adversary, and friend." In other words, what Bisson is saying is: breaking these rules is risky, but also useful. A writer can do interesting things by breaking rules. But a writer who breaks a rule is doing it intentionally, trying to accomplish specific effects, while still adhering to many other rules. A single story that broke all of these rules would probably be an incomprehensible mess; a great science fiction story is often one that breaks one or two of these rules while observing the rest.

In fact, rules and rule-breaking are two parts of language that go hand in hand. Chomsky discusses this in the essay "Language and freedom", where he writes, "free creation takes place within – presupposes, in fact – a system of constraints and governing principles". What he means here is that we couldn't be creative if we didn't have rules to break. Linguistic rules (i.e., grammar) create the conditions for linguistic utterances and acts to have meaning, and create the conditions that we exploit when we break the rules to express new meanings. We would not be able to do that if all speech were just mumbling meaningless unconstrained nonsense.

(And, on the flip side, this also explains why a theory of pragmatics is necessary: a completely semantic or conventional explanation of meaning would never be able to explain how meaning works, because, as Levinson [p. 112] points out, as long as linguistic rules and conventions exist, people will break them in order to communicate creatively. What a boring world we would live in if all language use were literal and rule-abiding!)

In this way, we could view pragmatics as something that takes advantage of grammar: "pragmatics" happens when a person exploits grammatical rules and conventions to achieve communicative goals.

"Pragmatic competence"

Just as people have grammatical competence (i.e., implicit understanding of grammatical rules and ability to use them), it is clear by now that people also have pragmatic competence: in other words, listeners are able to systematically figure out messages that speakers communicate without saying explicitly, and speakers are able to convey their messages to listeners without always making it explicit. And this process is usually successful: while there are some exceptions, speakers and listeners usually understand each other, which suggests that they are working on the basis of shared assumptions and principles rather than just guessing meanings randomly and idiosyncratically. In fact, describing this ability that people have is perhaps the main goal of the study of pragmatics (Levinson, chapter 1).

Nevertheless, the concept of "pragmatic competence" is often used vaguely and loosely. I especially see this phrase turn up in research about second language learners (i.e., research about how learners achieve "pragmatic competence" in their second language) and in various clinical "neuroatypical" populations (e.g., research about whether "pragmatic competence" is impaired in people with dementia, people with an autism spectrum disorder, etc.). When people make claims about certain groups having impaired "pragmatic competence", they often try to support these claims with observations that people don't follow certain social or conversational expectations. For example, I once spoke with a student who was interested in measuring pragmatic competence in adults with dementia by conducting a structured interview, and this student described their aims like so:

"[The interview will test] whether you can engage in turn-taking conversation... [for example] if I ask you something and you answer with a very short 'yes' or 'no' and then you stop talking, when there's a lot of things to elaborate but you stop talking, that would be a [pragmatic] problem."

The problem with such a way of measuring "pragmatic competence", though, is that all the behaviours this student mentioned (not taking turns properly, saying less than what would be appropriate) are also behaviours that a pragmatically competent speaker would do! As we've seen above and throughout this class, pragmatically competent users make implicatures by appearing to violate conversational principles. Thus, for instance, if a person says less than they "should" say in a given context, does that mean they are pragmatically incompetent, or they are competently flouting the maxim of quantity in order to implicate something? Since both situations would yield the same overt behaviour (saying less than appropriate), this behaviour cannot itself be taken as evidence about whether or not a person is pragmatically competent.

"Pragmatic competence" is really about someone's ability to recover speakers' intended meanings (e.g., understanding what you intend to implicate when you violate some maxim) and their ability to express their own intended meanings by using pragmatic techniques like indirectness. Pragmatic competence is not about following rules or conventions; indeed, to be pragmatically competent you must be able to violate conventions, and to recognize what other people mean when they violate conventions. (And, in fact, almost all of us are indeed able to do those things; you don't need to study pragmatics to do all that stuff, it's a natural ability you already have as a language user.)

Furthermore, it is crucial to distinguish between behaviours that happen because someone lacks pragmatic competence, versus behaviours that happen because someone lacks social competence or grammatical (linguistic) competence. In the "Violating maxims" module we saw an example about a woman who doesn't speak much English. If someone asks this woman a question in English and she responds in English in what seems to be a pragmatically inappropriate way, that would not be evidence that she lacks pragmatic competence; it would be more likely that she just lacks the grammatical competence (syntax, vocabulary, etc.) to respond in the desired way. This goes back to what we saw about rules above: pragmatics is all about breaking rules, so you can't do pragmatics within a system where you don't know the rules. We can't look at someone who doesn't know the grammar of a language and conclude on that basis that they don't have pragmatic competence. That would be like taking a professional poker player who doesn't know how to play mahjong, making him play mahjong, seeing him do a bad job at it, and then saying "He's not good at games of deception". That would be a stupid thing to claim, because the reason he didn't do well at mahjong isn't because he doesn't know how to do deception (indeed, as a poker player, he probably is good at games of deception), it's because he doesn't know the rules of mahjong. Likewise, it wouldn't make sense to claim someone is bad at pragmatics if actually they just aren't proficient in the grammar they are using.

Turning to social competence, we can also imagine situations where a lack of social competence (e.g., insufficient knowledge about the social norms of a particular context) could cause behaviour that looks like, but actually isn't, pragmatic "incompetence". For example, consider a particular kind of quantity implicature: if someone refers to a person with a description rather than referring to the person by name, (e.g., if someone says, "The man who got killed"), I might infer that the speaker does not know the person's name. However, there are some cultural contexts where this implicature should not arise; for example, in some communities it is taboo to speak the name of a person who "died a violent or premature death" (see Thomas, chapter 3). Now imagine that I am speaking with someone in such a cultural context, but I don't know about this taboo. If a person says "Do you know who killed that man?", I might infer that the speaker doesn't know the name of the man. My inference, then, would be wrong, but it would not be a result of pragmatic incompetence; it's an inference that comes from carrying out a pragmatic procedure properly. The real source of my error would have been the lack of social competence (or we could call it cultural competence): because I didn't understand the taboo about saying the person's name, I would have recovered a pragmatic implicature in a situation where I'm not supposed to.

We can see a similar example in the case of "Black Lives Matter", discussed in the module "Alternatives and context". As we discussed there, some people misinterpret this phrase as meaning "It is not the case that all lives matter". This is a misinterpretation, but the misinterpretation is not due to any deficiency in pragmatic ability. If a person thinks the question under discussion is "do all lives matter?", then inferring that "Black lives matter" means "not all lives matter" is a standard quantity implicature, realized through perfectly normal pragmatic principles. So, when people make this implicature, they are not being incompetent about pragmatics. Instead, they are misunderstanding the context that led to the utterance in the first place ("Black lives matter" is not a response to the question "Do all lives matter?", it's a response to an assumption that Black lives don't matter [this is an assumption that is rarely made explicit, but that is implicit from the way police and the carceral state treat Black people]). So the misunderstanding does not arise from a failure to do pragmatics correctly, but rather it arises from applying correct pragmatic procedures to an incorrect initial premise. In other words, this misunderstanding does not appear to be a result of "pragmatic competence", or pragmatics, at all.

Why did I have this long digression on the concept of "pragmatic competence"? It wasn't to bash the body of research that uses this term. Rather, I think looking at this issue helps us see what the relationship is between pragmatics and other aspects of communication (including grammatical and social aspects).

Ultimately, "pragmatic competence" is something almost everyone already has; it's a natural part of using language. Students often begin this class thinking that learning pragmatics is going to help them communicate more effectively and understand people better, but I don't think it will do that; in fact, you already were good at understanding people's indirect pragmatic communication before you ever studied pragmatics. Just like you don't need to study biology or orthopaedics or sports medicine to enjoy playing basketball, and you don't need to study media and film studies to watch a movie, and you don't need to study linguistics to speak Chinese, I also don't think anyone needs to study pragmatics to communicate effectively. Almost all of us already can communicate effectively (i.e., we already have "pragmatic competence"); studying pragmatics is just a way to gain an explicit understanding of what we are already naturally able to do.

Where semantics ends and pragmatics begins

One philosophical question to end with: does semantics feed into pragmatics, or does pragmatics feed into semantics?

Philosophers and theoretical linguists like to think of various parts of linguistics (phonology, morphology, syntax, semantics, etc.) as functions; each of these takes in one thing and spits something else out. (See the "Truth-conditional semantics" module for details.) For example, linguists see syntactic theory as a function from sentences to grammaticality: a "syntax" is a function that takes any sentence (more accurately, any combination of morphemes) in a language, and tells us whether that sentence (that combination of morphemes) is grammatical or ungrammatical in that language.

As language comprehenders, I think we often tend to think of pragmatics as coming "after" semantics; i.e., we tend to think of semantics as a function that takes some sentence and tells us "what is said", then we put that "what is said" into the function of pragmatics, which tells us "what is meant". But that is not the only way to see the relationship between semantics and pragmatics.

As Levinson (chapters 1 and 2) points out, we could instead see pragmatics as feeding into semantics. According to some theories, pragmatics is a function from utterances and contexts to propositions. In other words, if you take a given utterance uttered in a given context, pragmatics tells you what proposition that utterance is really expressing. (The utterance "My arm is hurt" could express a very different meaning depending on whether it's uttered in a doctor's office in response to a doctor asking why you've come there, or uttered at a dinner table in response to someone who just asked "Can you pass the salt?") Propositions are what have truth conditions, and that is what semantics handles (the meaning of a proposition is a function from possible worlds to truth values—if you know what a proposition means, that means you know what conditions need to hold in the world for that proposition to be true or false). The main point here, abstracting away from the complicated philosophical mumbo-jumbo, is: rather than assuming that the output of semantics is the input to pragmatics, we could instead assume that the output of pragmatics is the input to semantics. Pragmatics takes a person's utterance and figures out what logical proposition the utterance actually expresses, and semantics then figures out what that proposition literally means.

We are not in a position here to determine which of those positions is right; pragmaticists, semanticists, and philosophers have been debating this stuff for decades. If it sounds confusing to you, that's good; this is confusing stuff, so if you feel confused now that's a sign that you're thinking seriously about what it all means.


We've reached the end of our foray into some classic pragmatics topics.

Ultimately, as we have been seeing since the beginning of the class, many of the topics in pragmatics are still under debate and don't have any clear right answer. Almost every concept we have seen this semester is challenged or rejected by some theories. And there are some influential approaches, like conversation analysis (see Levinson, chapter 6) and Relevance Theory (see Noveck, chapter 2, or Zufferey et al., chapter 3), that I have not done justice to in this class, and that would analyze many of our examples and phenomena in a very different way than what we have done. There are also huge aspects of pragmatics that we haven't explored at all, such as politeness theory, reference, and deixis. But I hope examining the issues that we did examine has given you a taste for how to keep an open mind to different ways of looking at a complicated issue; that's a skill that will serve you well not just in pragmatics, but for life in general. I also hope that the concepts and phenomena we have seen throughout this class will provide you with some useful tools to help understand and explain interesting things that happen in communication, and appreciate the variety of fascinating linguistic phenomena that are out there.

Video summary

In-class activities

I always discourage students from seeking "official" definitions of anything. I strongly believe that most things worth defining are too complicated to have any straightforward definition. A better way to appreciate something is to look at lots of definitions and see what's similar and different, or to look at examples and phenomena that illustrate the thing (this latter approach is what we've been doing for the whole class).

That advice goes double for pragmatics. There is no official definition of "pragmatics". Levinson (chapter 1) spends 30 pages trying to define pragmatics, and ultimately pretty much concludes that he can't. Another textbook begins each chapter with a completely different definition of pragmatics. These examples should illustrate how hard it is to define "pragmatics".

That being said, what I'd like you to do in this activity is put students into groups and have them try to come up with a definition for "pragmatics". No looking at other sources; they should try to show their own ideas. There are no wrong answers. Once they've had a try, they can compare and contrast their definitions with other groups', and think about if there are any phenomena from this class that are left out of some definitions, or any phenomena that don't seem pragmatic but would fall under one of the definitions, or any other ways a given definition might be vague or problematic.

Throughout this class you have learned about many different pragmatics concepts and theories. You might have sometimes felt confused about the differences between them, as sometimes different theories or frameworks use different terms to refer to similar phenomena (for example, "what is said" and "utterance meaning", "force" and "implicature", etc.).

In fact, it is often the case that different approaches to pragmatics can offer different explanations for the same phenomenon. Neither is necessarily right or wrong; they're just different ways of looking at the same thing. Here I'll share one example, and ask students to try to brainstorm other examples that can also be analyzed through different pragmatics concepts.

The example I have in mind comes from a scene in the show Better Call Saul. The context of the conversation is that Kim, a lawyer, has gotten in trouble with her boss at work and is being punished for it. Her boyfriend, Jimmy, offers to talk to her boss, Howard, to help smooth things over. But Kim doesn't want his help (because he's the one who got her in trouble in the first place, and because she wants to solve the problem on her own). They argue for a while, and the following exchange occurs:

Kim's utterance is meant to warn Jimmy that if he talks to her boss, she will end their relationship. But Jimmy interprets another piece of information in her utterance: in addition to the message I just mentioned, he also interprets her message as meaning that their relationship is not over yet. How does he arrive at this interpretation?

I can think of at least two ways we could explain this interpretation, one based on the Gricean theory of implicatures and one based on the theory of speech acts.

In terms of implicatures, we can see that Kim chose not to directly say "We're through", but instead uttered something else that doesn't entail they're through; this should elicit a clausal implicature meaning that Kim doesn't believe they are through. We also could try to analyze this through other conversational maxims as well (for example, if they are already through, then describing the extra condition "if you go to Howard..." is irrelevant—since they are already through regardless of whether or not Jimmy goes to Howard—so if we assume Kim is following the maxims of relation and quantity and only saying relevant information then we should also assume that Kim and Jimmy's relationship is not "through" yet). Regardless of which maxim we focus on, the important point is that in order to assume Kim is obeying the Cooperative Principle, we'll probably have to assume that she means to implicate that she and Jimmy are not through yet.

On the other hand, we could analyze the same interpretation in a very different way with the theory of speech acts. When Kim says "If you go to Howard, we're through", the perlocutionary effect she hopes to accomplish is to stop Jimmy from going to Howard. She does that by performing an illocutionary act of threatening: i.e., she tells Jimmy that if she does some thing that she doesn't want, something bad will happen to Jimmy. A threat is only felicitous if you're threatening something that hasn't already happened; if the "something bad" is already happening, then the threat won't "work". So, if we assume that Kim intends her utterance to have the illocutionary force of a threat, then we have to assume it meets the felicity conditions for threatening, and thus we have to assume that she and Jimmy aren't through yet.

There's nothing particularly unique about this example; it's just a simple illustration that we can use very different pragmatics concepts to explain the same interpretation of the same utterance. Many of the examples throughout this class could also be re-analyzed in similar ways.

Have students try to look at examples (it could be previous examples from the modules of this class, examples from previous discussions, or new examples) and try to come up with multiple explanations for the same example.

⟵ Grammar

by Stephen Politzer-Ahles. Last modified on 2022-03-30. CC-BY-4.0.