Pragmatics is the study of "how to do things with words" (the name of a well known book by the philosopher J.L. Austin), or perhaps "how people do things with words" (to be more descriptive about it).
We'll consider four aspects of pragmatics in this lecture: speech acts; rhetorical structure; conversational implicature; and the management of reference in discourse.
People use language to accomplish certain kinds of acts, broadly known as speech acts, and distinct from physical acts like drinking a glass of water, or mental acts like thinking about drinking a glass of water. Speech acts include asking for a glass of water, promising to drink a glass of water, threatening to drink a glass of water, ordering someone to drink a glass of water, and so on.
Most of these ought really to be called "communicative acts", since speech and even language are not strictly required. Thus someone can ask for a glass of water by pointing to a pitcher and miming the act of drinking.
It's common to divide speech acts into two categories: direct and indirect.
There are three basic types of direct speech acts, and they correspond to three special syntactic types that seem to occur in most of the world's languages. Examples are given in English, French and Buang (a Malayo-Polynesian language of Papua New Guinea
Although assertions, questions and orders are fairly universal, and most of the world's languages have separate syntactic constructions that distinguish them, other speech acts do not have a syntactic construction that is specific to them. Consider the English sentence,
(a) If you cross that line, I'll shoot you!
Most English speakers would have no trouble identifying such an utterance as a threat. However, English has no special sentence form for threats. The if-construction used in (a) is not specific to the speech act of threatening. Such a construction might also express a promise, as in:
(b) If you get all A's, I'll buy you a car!
or simply a cause and effect relationship between physical events:
(c) If you heat water to 212 degrees Fahrenheit, it will boil.
A consideration of the syntactic means available for expressing the various speech acts leads us to see that even for the three basic speech acts laid out in the table above, speakers may choose means of expression other than the basic syntactic type associated with the speech act in question.
To some extent, this just reflects the existence of a diversity of means of expression, but a more pervasive reason is that speakers may use indirect rather than direct speech acts.
Returning to the speech act of questioning, we can easily come up with a number of alternate ways to ask the same question by using sentence types other than interrogative. Let's look again at the interrogative sentence:
(d1) Did Jenny get an A on the test?
A positive answer ("yes") to that question would give the questioner the actual answer she wanted, but now consider (d2)
(d2) Do you know if Jenny got an A on the test?
This is still in the form of a question, but it probably is not an inquiry about what you know. Most of the time, the answer "yes, I do" would be ostentatiously uncooperative. The normal answer we would expect in real life would be "Yes, she did", or "No, she only got a B", or something of the sort. Here the reply is directed to the speech act meaning, not the literal meaning. A simple "yes" answer that responds to the literal meaning would usually be taken for an uncooperative answer in actual social life (for example "Yes, I do") would be heard as "Yes, I do, but I'm not necessarily going to tell you".
Other indirect ways of asking the same question, using the declarative form, are listed in (d3) and (d4).
(d3) I'd like to know if Jenny got an A on the test.
(d4) I wonder whether Jenny got an A on the test.
In the case of the speech act of requesting or ordering, speakers can be even more indirect. As in the case of questions, conventional indirect requests may, taken literally, be questions about the addressee's knowledge or ability. Here is a direct request:
Conventional indirect requests may be expressed as questions as in (e2) and (e3), or as assertions (e4). In context, (e5) and (e6) may also be immediately understood as a complaints, meant as an indirect request for action.
(e2) Could you close the window?
(e3) Would you mind closing the window?
(e4) I would like you to close the window.
(e5) The window is still open!
(e6) I must have asked you a hundred times to keep that window closed!
One subtype of direct speech acts exists in English and in many other languages, and allows us to expand the kinds of direct speech acts we can make beyond the three basic types that have their own special syntax. These are the direct speech acts that use performative verbs to accomplish their ends. Performative verbs can also be used with the three basic speech act types as exemplified in (f) - (h), associated with making statements, requests and commands respectively:
(f) I assert that Jenny got an A on the test.
(g) I ask you who took the photos.
(h) I order you to close the window.
To these can be added performative verbs that allow us to directly convey promises, threats, warnings, etc.
(i) I advise you to keep up the payments on your car.
(j) I warn you not to step across this line.
(k) I promise you that I will pay the money back by the end of the month.
(l) I bet you a dollar that it'll rain on the parade.
In the last sentence, the utterance of the sentence actually accomplishes the act of betting (possibly along with setting aside the money for the bet), and as such, it belongs to the class of ceremonial utterances that accomplish other kinds of changes in the world:
(m) I now pronounce you husband and wife.
(n) I name this ship Sojourner.
(o) I dub thee Sir Galahad.
It is clear that not all uses of verbs that can be performative are actually performative in particular utterances. For example, if we change the person or the tense in any of the last seven sentences, they are no longer performative:
(i2) He advises you to keep up the payments on your car.
(n2) I named this ship Sojourner.
In both these cases, the utterance simply reports, and does not accomplish the act of advising or of naming.
The hereby test.
A test of whether or not a particular sentence is a performative utterance is whether or not you can insert hereby before the verb. If the resulting sentence doesn't make sense, it is not a performative:
(m3) I hereby name this ship Sojourner; but
(m4) *I hereby named this ship Sojourner.
Some researchers have extended the classical lists of "speech acts" to include many actions that are felt to be helpful in analyzing task-oriented dialogs, things like "answer", "accept", "reject" and so forth. One influential set of ideas about this is expressed in the so-called DAMSL ("Dialog Act Mark-up in Several Layers") proposal.
For another, funnier take on an extended set of speech acts, listen to this scene (transcript here) from Chicago's Neo-Futurist group. A similar idea applied to weblog posts is here. There's also Spamalot's "A Song Like This", Da Vinci's Notebook's "Title of the Song", news reports from The Weekly Wipe and The Onion, and the Trailer for every Oscar-winning movie ever.
The tradition of double-talk comedy depends on our ability to infer the rhetorical structure of a presentation without any lexical information at all -- here's John Cleeses doubletalk version of a neuroscience lecture:
The work of H.P. Grice takes pragmatics farther than the study of speech acts. Grice's aim was to understand how "speaker's meaning" -- what someone uses an utterance to mean -- arises from "sentence meaning" -- the literal (form and) meaning of an utterance. Grice proposed that many aspects of "speaker's meaning" result from the assumption that the participants in a conversation are cooperating in an attempt to reach mutual goals -- or at least are pretending to do so!
He called this the Cooperative Principle. It has four sub-parts or maxims that cooperative conversationalists ought in principle to respect:
(1) The maxim of quality. Speakers' contributions ought to be true.
(2) The maxim of quantity. Speakers' contributions should be as informative as required; not saying either too little or too much.
(3) The maxim of relevance. Contributions should relate to the purposes of the exchange.
(4) The maxim of manner. Contributions should be perspicuous -- in particular, they should be orderly and brief, avoiding obscurity and ambiguity.
Grice was not acting as a prescriptivist when he enunciated these maxims, even though they sound like prescriptions for how to communicate. Rather, he was using observations of the difference between "what is said" and "what is meant" to show that people actually do follow these maxims in conversation. We can see how this works in considering the maxim of quantity at work in the following made-up exchange between parent and child:
| Parent: | Did you finish your homework? |
| Child: | I finished my algebra. |
| Parent: | Well, get busy and finish your English, too! |
The child did not say that her English homework is not done, nor did she imply it in a legalistic sense. Nevertheless the parent is likely to draw this conclusion. The implicit line of argument is something like this: the child would have simply said 'yes', without mentioning any particular subjects, if that answer were true; the fact that she referred to algebra, and did not mention other subjects, suggests ("implicates") that the unmentioned subjects are not done.
Very often, particular non-literal meanings are conveyed by appearing to "violate" or "flout" these maxims. If you were to hear someone described as having "one good leg", you would be justified in assuming the person's other leg was bad, even though nothing had been said about it at all.
A more elaborate taxonomy of types of conversational implicature, with illustrative examples, can be found here.
[Exercise for the reader: the last sentence implicates that you should go read the referenced site. What kind of implicature is that?]
Recent work in Relevance Theory builds on Grice's insight about the nature of communication:
"Relevance theory may be seen as an attempt to work out in detail one of Grice’s central claims: that an essential feature of most human communication, both verbal and non-verbal, is the expression and recognition of intentions. In developing this claim, Grice laid the foundations for an inferential model of communication, an alternative to the classical code model. According to the code model, a communicator encodes her intended message into a signal, which is decoded by the audience using an identical copy of the code. According to the inferential model, a communicator provides evidence of her intention to convey a certain meaning, which is inferred by the audience on the basis of the evidence provided. An utterance is, of course, a linguistically coded piece of evidence, so that verbal comprehension involves an element of decoding. However, the linguistic meaning recovered by decoding is just one of the inputs to a non-demonstrative inference process which yields an interpretation of the speaker's meaning."
Consider the following paragraph from the introduction to a 2003 magazine article:
Here is the problem with the college-admissions system. It is a vast and intricate bureaucracy designed to do one thing, and it does that very well; but it is under intense social and economic pressure to do something different—something more or less directly at odds with its supposed goal. The resulting tensions affect everyone involved: The high school guidance counselors who try to steer students toward the right school. The college admissions officers who sort through ever mounting piles of applications to choose an entering class. The college administrators who wonder how many of those accepted will enroll—and how many of them will need financial aid. The parents who contemplate what will be (after housing) the second largest financial outlay of their lives. And, of course, the students themselves.
-The Atlantic Monthly, 11/2003.
Just as a phrase is a structured combination of words, this paragraph is a structured combination of phrases. Let's review how this passage uses the kinds of syntactic and morphological structure that we've previously studied to express semantic relations like modification and predicate-argument structure ("who did what to whom").
In the paragraph above, "vast and intricate" modifies "bureaucracy", and "intense" modifies "social and economic pressure." This modification relation is expressed by a structure that we might write down something like this:
(NounPhrase
(Adjective (Adjective vast) (Conjunction and) (Adjective intricate)))
(Noun bureaucracy)))
[Exercise for the reader: how would you draw this structure as a tree?]
STEER is a predicate whose arguments may include an agent (the person or thing steering), a theme (the person or thing that gets steered), and a goal (the path or endpoint of the steering). In the passage above, there is a verb "steer" whose agent is "high school guidance counselors", whose theme is "students", and whose goal is "toward the right school". This predicate-argument relation is expressed by the fact that "steer" is a active verb whose subject is the agent, whose object is the theme, etc. In this case, the subject turns out to be "high school guidance counselors", after we untangle and interpret the syntactic structure (. counselors who try to steer . ). The same semantic relation might have been expressed by syntactic patterns such as "students being steered by guidance counselors", "the steering of students by guidance counselors", and so on.
When you read and understand this passage, you're parsing the structure of its phrases, and using the results to help you interpret local aspects of the meaning, such as modification and predicate-argument structure.
However, it should be clear to you that there is also a larger-scale structure relating the 11 clauses of this paragraph. We could indicate this structure as follows (I've used line breaks, indentation and square brackets for the structural relations, simplified the language to make it fit more easily on the page, and numbered the clauses for future reference):
[A
(1) Here is the problem with the college-admissions system.
[A1
[A1.1 (2) It is a bureaucracy designed to do one thing; A1.1]
[A1.2
[A.1.2.1 (3) but it is under pressure to do something different— A.1.2.1]
[A.1.2.2 (4) something at odds with its supposed goal. A.1.2.2]
A1.2]
A1]
A]
[B
[B.1 (5) The resulting tensions affect everyone involved: B.1]
[B.2 (6) Counselors who steer students toward the right school. B.2]
[B.3 (7) Admissions officers who sort through piles of applications. B.3]
[B.4
[B.4.1 (8) Administrators who wonder how many will enroll— B.4.1]
[B.4.2 (9) and how many of them will need financial aid. B.4.2]
B.4]
[B.5 (10) Parents who contemplate their second largest outlay. B.5]
[B.6 (11) And, of course, the students themselves. B6]
B]
At the highest level, there are two units: A, there is a problem with college admissions; B, the results affect everyone involved. The first unit explains something about what the problem is; the second units lists the kinds of people affected and something about their roles in the process.
Some of this rhetorical structure involves the relationships of phrases inside sentences, and some of it involves relationships among sentences or groups of sentences. Like syntactic structures, these rhetorical structures usually seem to be "trees" -- that is, successive subdivisions of larger units into smaller ones.
One difference between syntax and rhetoric is scale -- syntax typically operates on a smaller scale, among words or small groups of words inside sentences, while rhetoric works on a larger scale, typically relating clauses, sentences and whole sections of a discourse.
Another difference is function. Syntactic structure mainly expresses semantic relations like modification, predication, quantification and so on. These are key parts of a basic account of "sentence meaning". Rhetorical structure typically expresses pragmatic relations like exemplification, concession, justification, summary and so on, things that are part of "speaker meaning", the way that people use language to inform or entertain or persuade.
For the past century or two, linguists have been much more interested in syntactic structure than in rhetorical structure. As a result, theories of syntax are much better developed and more widely known. However, there are some interesting accounts of rhetorical structure, including a body of work known as RST (for "rhetorical structure theory"). RST postulates a tree structure of rhetorical relationships, analogous to the tree structures of syntactic relationships; RST has been used in practical applications, especially in systems for generating text, and there's a set of 380 RST-annotated WSJ stories published as the RST Treebank. (For an interesting debate about whether rhetorical relationships are indeed tree-like, see this weblog post "Discourse: Branch or Tangle?", and the various papers linked therein.)
A different approach is taken in the Penn Discourse Treebank, which annotates a million words of WSJ stories with pairwise discourse relations marked by explicit connectives such as accordingly, because, by comparison, however, etc.; and also implicit relationships like these:
Several leveraged funds don't want to cut the amount they borrow because it would slash the income they pay shareholders, fund officials said. But a few funds have taken other defensive steps. Some have raised their cash positions to record levels. Implicit = because High cash positions help buffer a fund when the market falls.
The projects already under construction will increase Las Vegas's supply of hotel rooms by 11,795, or nearly 20%, to 75,500. Implicit = so By a rule of thumb of 1.5 new jobs for each new hotel room, Clark County will have nearly 18,000 new jobs.
Some linguists are skeptical that rhetorical structure is even a well-defined kind of mental representation that is intrinsic to language, of the kind that syntax seems to be. Thus rhetorical structures might simply be useful patterns that speakers and writers sometimes choose to create or borrow, just as architects and painters may choose certain stereotypical ways of arranging their materials.
On this view, it might not turn out to be possible to list all the possible rhetorical structures, or to uniquely and accurately classify every passage in terms of such structures, any more than we can list all possible hand tools, or necessarily classify all the tools that we find into a fixed taxonomy. After all, someone can always invent a new tool, or a tool that combines aspects of several old ones. However, it is still useful to have an inventory of the tools in common use at a given time, and (even if these skeptics are right) an inventory of common rhetorical structures has a similar sort of value.
In conveying a message, we have to consider more than just "who did what to whom." We also have to keep in mind what our listeners know, and how to lay the message out for them in an orderly and understandable way.
We have to be careful not to assume knowledge listeners don't have. If a stranger comes up to us on the street and says, out of nowhere, "what is the frequency?" we are likely to assume that he is crazy, or perhaps mistaking us for someone else. Young children make this sort of communicative mistake all the time, because their ability to model other people's knowledge and belief is not well developed.
Similarly, we have to be careful not to introduce familiar things as if they were new. Aside from being insulting, this can be confusing, since our listeners may try to find a new interpretation to match our implication of novelty. If your roommate says "there's a letter for you on the table", and it's the same old letter that both of you know well has been there for several days, you may waste some time looking for another one.
There are many aspects of language that help to indicate whether a particular piece of information is "old" or "new", and to manage the amount of detail that we use in talking about it, and to make it more or less salient for our listeners or readers. For example, "old information" (part of the earlier content of a discourse, for instance) is referred to using a pronoun, and occurs early in a sentence. What is "new" typically occurs as a noun, and occurs later in the sentence:
"When John appeared at the party, he was introduced to Pearl.
She had arrived with her friend Julie."
In this text fragment, John turns into 'he' when John is "known", and this pronoun occurs at the beginning of the clause that introduces Pearl as new. When Pearl becomes known, she also gets converted to the pronoun 'she' in the next sentence, occupying a slot at the beginning of the next sentence, which in turn introduces the new character, Julie, in the typical sentence-final position.
Here's a more realistic example, taken from a transcript of conversation about fashions that took place in 1991 (sw4746):
| B.72: | [Sniffing] One thing I've noticed is [sic] come back here are clogs. |
| A.73: | Really? |
| B.74: | Yeah. They're starting to make a comeback. You see them in the stores more and more and I said I didn't think I'd ever see those again. [laughter] |
The new information "clogs" is put at the end of the phrase that introduces it, and then referred to with a pronoun at the start of the next full sentence that discusses it. Consider how odd it would be to do the opposite, switching the structures of the first and second of B's sentences::
| B.72: | [Sniffing] Clogs're starting to make a comeback.. |
| A.73: | Really? |
| B.74: | Yeah. One thing I've noticed is come back here are them. You see them in the stores more and more and I said I didn't think I'd ever see those again. [laughter] |
Something similar often happens with indefinite and definite noun phrases ("a man" or "some people" vs. "the man" or "those people"). Here's a real example from another transcribed conversation (sw4787) , this one about family reunions (overlapping speech is marking with #. #):
| B.52: | And well they elect officers every year and #they have a# |
| A.53: | #You're kidding.# I have never heard of this. [laughter] |
| B.54: | Yeah, they have a,- they have a- a president. Usually they try to elect a family and inside that family, there'll be the president and- or the chairman or whatever and then each person has an assignment to- to you know, carry out one part of the thing. |
Here speaker B starts out by saying "they have a president ", and then, in adding more information, switches to " the president ". The same sort of switch from indefinite to definite occurs in saying "usually they try to elect a family and inside that family , there'll be. " As this switch occurs, nothing is changed about the nature of the concepts that the phrases are naming -- the only thing that changes is the listener's familiarity with them.
Another way to study how we organize and package information according to the communicative context is to look at the usage of different sentence forms with very similar meaning.
(o) I need a nickel.
(p) It's me that needs a nickel.
(q) What I need is a nickel.
(r) A nickel is what I need.
Now, imagine yourself standing next to a phone booth fishing for change. Someone trying to be helpful might say:
(s) What are you looking for?
(t) Here's a dime.
Which of (o)-(r) are appropriate responses to each of these?
Studying such potential question and answer pairs shows us that sentences can express the same semantic content and still have different pragmatic circumstances of appropriate usage. This is because language has many devices for indicating what is given and what is new, and questions (explicit or implicit) set up expectations that are respected in the answers.
Some reflections of these issues in philosophy of law:
Lawrence Solan, "Private Language, Public Laws", Georgetown Law Review, 2004 (preprint here).
Antonin Scalia, Review of Law's Quandary by Steven Smith, First Things, November 2005
Mark Liberman, "A result that no sensible person could have intended", Language Log, December 2005.