Pragmatics

 We called semantics the study of meaning, but clearly our understanding of the concept of meaning is more richly textured than what we've discussed as the truth conditional way of understanding sentences. When we think about meaning, it is also important to take into account the contribution of context. In simple terms, then, we can think about pragmatics as the study of the contribution of context to meaning.

1) Context: what is it?

Here are four subareas involved in fleshing out what we mean by context.

Physical context: we can think of this in terms of where the conversation is taking place, what objects are present, what actions are occurring, and so forth.

Epistemic context: the epistemic context refers to what speakers know about the world. For example, what background knowledge is shared by the speakers is crucially part of your epistemic knowledge when you have a conversation with someone else?

Linguistic context: the linguistic context refers to what has been said already in the utterance. For example, if I begin a discussion by referring to Jane Smith and in the next sentence refer to "her" as being a top notch athlete, the linguistic context lets me know that the antecedent of "her" (the person "her" refers to) is Jane Smith.

Social context: the social context refers to the social relationship among speakers and hearers.

Think about the discussion of the library scene in Language Files. Two people come into a library and they are talking really loud. They sit at your table and continue their babbling. So, you look up at them and say:

"Excuse me, could you please speak up a bit more? I missed what you said."

What do we know pragmatically about your utterance? What contributes to our understanding of why its literal meaning (i.e. please speak up) departs so much from what its intended meaning is (i.e. shut up!)? Consider some of the contextual properties of the utterance.
 


Contextual knowledge allows for the hearer to comprehend that the intended meaning is distinct from the literal meaning. With any luck, they'll even tone it down.

Besides understanding broad notions of how context is important, pragmatics also examines the kinds of Speech Acts that people employ when they speak. We'll cover the notion of speech acts here.

2) Speech Acts

Basically, we can divide the world into two flavors of Speech Acts: direct and indirect. Here is a sketch of direct speech acts, with indirect acts addressed later on.

Direct Speech acts: (3 sub-flavors to start with)

Note that imperatives and interrogatives are different from declaratives in that they don't have truth conditions. Rather, they seem better understood in terms of whether they are appropriately used or not. For example, if you say to your two year old brother, "Bobby, can you explain the basic idea behind string theory in contemporary physics?" it is more than likely the case that Bobby won't have a clue what you are talking about. The question itself is understandable, but not to Bobby. Thus, its use is inappropriate. This refers to the idea of "felicity", and we'll discuss this a bit more below.

Before leaving the world of direct speech acts, there's one more twist. A special kind of direct speech act is called a performative. Performatives are interesting in that they are basically verbs whose action is a speech act.

We can embed declaratives, interrogatives, and imperatives such as those in (1-3) in sentences.

These are interesting in that they call our attention to the fact that "speech action" is like any other action for us. What's up here is that the main verbs of these sentences are performing actions that are purely linguistic, asserting, asking, and ordering.

Okay, this wouldn't be a linguistics class if we didn't discuss a way of attempting to test for what kind of speech act you've got in front of you.

Consider the following three sentences: Here's the big question: Which of these is a PERFORMATIVE use of the verb promise?

Here's the answer: Only the first one. Why?

Note that the other sentences are still direct speech acts, just not performatives. How else can we test for a performative use of a verb? One trick is to try inserting "hereby". I don't know about you all, but only sentence (1) sounds fine to me. The insertion of "hereby" in the other two makes them seem somewhat less than entirely acceptable. Be careful with the "hereby" test, though, as it can also sound odd with some performatives. For example, if you say, "I hereby ask you to give me a spoon", that sounds odd. But note that it is even odder to insert "hereby" in non-performatives like #He hereby asks me to give him a spoon, or #I will hereby ask you to give me a spoon.

3) Felicity Conditions

Think about the many ways that a sentence can go wrong? It can be mispronounced. We can regularize an irregular verb. We can produce an ungrammatical sentence. All of these are errors that aren't exactly pragmatic errors in the sense that we want to discuss here. The kind of pragmatic error that we want to think about here is the situationally inappropriate use of a sentence. Inappropriate sentences can be perfectly well-formed, but they can nevertheless be situationally all wrong.

Linguists characterize the notion of situational inappropriateness in terms of what are called felicity conditions. The basic idea here is that felicity conditions allow us to determine under what circumstances it is appropriate to ask questions, give commands, and so forth. Note that we are just touching the tip of the iceberg here. Really delving into the area of felicity conditions requires a ton of work, but what we'll review here should give you a decent idea of how the thinking goes. One big thing to bear in mind is that if we get the felicity conditions down explicitly enough, we can pinpoint the nature of the inappropriate use of language in terms of the particular felicity condition or conditions violated.

As an example, here are some felicity conditions associated with questioning and requesting.

S questions H about P (some state of affairs)

What's going on with the felicity conditions for questioning? Well, for starters, we can see that appropriate use of questions must satisfy a number of criteria. Note that these felicity conditions don't hold of all questioning contexts. In class, for example, I might ask you a question about, say, compositional semantics or X-bar theory or derivational versus inflectional morphology. If I do so, the odds are good that I already know the answer. So, for teacher-to-student questions, the first condition is suspended. If we think about it, our knowledge of social and physical context allows us to understand why. Teachers, we know, ask questions of students in order to test what students know about topic X. Physically, we're in the classroom--the place where teachers are likely to be in this kind of questioning role. So, this is a kind of question for which those felicity conditions are suspended.

Additionally, (2) and (3) don't exactly hold. The teacher doesn't so much want to know the truth about X as to know what the student knows about X. And, when asking the question, the teacher doesn't necessarily know whether the student knows the truth about X. Rather, the teacher often asks in order to see IF the student knows the truth about X.

Finally, note that I may stand in front of the class and ask the following question: "Does anybody have the time?" Immediately, you all will most likely shift back into understanding my question in terms of the felicity conditions in (1-3) above and give me an answer. That is, even though I'm the teacher, that's the not kind of question that would be asked in the teacher role. The question would be infelicitous only if I have a watch on that you know is working perfectly or if there is a huge clock with the time in front of me, because I'd be violating the first condition.

Anyway, think about how you do this kind of interpreting all the time. Think about how you are constantly sifting through your knowledge of context and of the felicity conditions behind questions in order to understand whether the question is used appropriately.

Here's a general set of felicity conditions on requests.

S requests H to do A (action):

When I told you all to raise your hands the other day in class, which of these was violated? That's right. Condition 4. I didn't really want you to stand up, except to illustrate how important (3) is. You stood up because in the context of the classroom, we all determined together than you all were willing to do things like stand up if requested by me, the teacher.

4) Indirect Speech Acts

That last observation gives us a nice transition to the notion of indirect speech acts. Consider the following sentences

Given the right context, all three can all three "mean" basically the same thing, i.e. "give me a ride to the mall." However, the hallmark of an indirect speech act is that its literal meaning is different from its intended meaning. People are intuitively aware of this, and that's why we can play around as follows:
 


Clearly, B has responded literally to A's question, while A's meaning was a request for B to pass the salt, and not a query regarding B's ability to engage in the activity of salt-passing.

How can we identifying indirect speech Acts?

How to do this (flow chart style)

S questions H about P (some state of affairs) If I say to you "Could you give me a ride to the mall?", I am violating (1), assuming that I know you could take me to the mall. So, if my question is felicitous, it is only felicitous with the understanding that it is really a request to take me to the mall and not a question about your ability to do so.

Note that the felicity conditions on requests aren't violated (assuming this to be true):

S requests H to do A (action):


So, this is a question literally, but intended as a request. It is an indirect speech act. This said, we might wonder why we use indirect speech acts at all. That is, why don't we always just go straight to the point, stop beating around the bush, etc? In this question/request, we can think of the indirect speech act as a way of communicating to the hearer that although the speaker believes that the hearer is willing to take her/him to the mall, there is also a recognition that asking this favor constitutes a kind of imposition. Framing the request as a question thus allows the hearer the opportunity to more gracefully decline. There are many subtle elements here. Perhaps the speaker has the power to make the hearer take him/her to the mall but does not want to make a show of that power by making a direct request. Perhaps the speaker is unsure about whether the hearer will be willing to take him/her to the mall.

What I want you to notice is that indirect speech is fundamental to the way in which we communicate, because much more is communicated than the so-called literal meaning of our utterances when we use language. That's why I'm often amused by people who think things would be better if we always got straight to the point. Beating around the bush provides speakers and hearers with valuable information. That's why we do it.

An exercise about Felicity Conditions:

Here's felicity conditions on PROMISE as given in Language Files.

S promises H to do A:

Now, what change(s) could we make to provide felicity conditions on THREATEN? Hint, I think you can do it by adding just a couple of words to one of the conditions in (1-4).
 

5) Rules of Conversation

How can we keep the ball rolling? How can we continue to find usefulness in communication through language? Imagine that people all told lies in random fashion (as opposed to for particular and often transparent reasons). How effective would language be as a communicative device?

Gricean Maxims

The Gricean maxims are a framework for understanding how humans cooperate socially in their use of language.

H.P. Grice first formulated the notion of the so-called cooperative principle. This is not to say that only one type of cooperation underlies all communication. Rather, it attempts to characterize the more generally cooperative aspect of human behavior across situations.

Have you ever spoken with a child and had the following type of exchange?

If you've had this kind of experience, you are ready to appreciate the Gricean Maxims. For a start, they'll help you understand why such conversations seem bizarre and/or funny.

In a nutshell, here are the Gricean maxims.

Maxims of Quality

Maxim of Relevance
  1. Be relevant.
Maxims of Quantity Your book gives you an interesting example. Literally, Gail can be said not to have lied in her first answer, since the fact that she can run 26 miles entails the fact that she can run 24 miles. But, she violated the 1st maxim of Quantity. Her answer wasn't informative enough. That Kim assumes that Gail is respecting this maxim is what underlies her second question (which is actually an indirect question).

Here's a violation of the maxim of quantity at a store.

Maxims of Manner


What do the following violate?


6) Language in Advertising

This section is interesting because it allows you to really examine closely the difference between two concepts: entailment vs. implicature, and it also allows us to think about how implicature requires reference to pragmatics in a way that entailment doesn't.

What's entailment? In a sense, we've already seen entailment in hyponymy. The word "mare" for example, entails the concept of "horse", since all things that are "mares" are also "horses". Here's a couple of sentences that have an entailment relation:

Note that if (1) is True, then (2) also has to be True. So, sentence (1) entails sentence (2). (The reverse is clearly not the case, of course. If Sally drives a car, it is not necessarily the case that she drives an Accord, so driving a car does not entail driving an Accord.)

The concept of implicature is slightly different from entailment, but it is related in a way that makes the two concepts sometimes tricky to tease apart, at least as far as language use is concerned. Consider the following pair of sentences.

The heart of the matter here lies in the following question: Does sentence (1) entail sentence (2)?

By the normal interpretation, it would seem to. That is, if I say to you, "Not everyone is going to come", you understand me to be saying that at least someone is coming. For example, you invite the football team to a fund raiser, and the coach calls you and says, "not everyone is going to come." Would you be surprised to arrive and see that absolutely no one showed up? Of course. Why? Because the sentence "not everyone is going to come" implicates that SOMEONE is going to come. However, it doesn't entail it. Note that "Not everyone is going to come" would still be true in a situation in which "no one is going to come." Contrast this with the car sentence above. If "Sarah drives a truck", the sentence "Sarah drives an Accord" can't be true. (Of course, we are making some simplifying assumptions, such as the assumption that Sarah does not have both an Accord and a truck!)

BIG BPICTURE

For something to implicate something else, two things have to be the case:

In this case, we can see that the Gricean maxim of quantity plays a role. If the speaker had wanted to mean that no one is going to come, s/he would have said so (the message would have been as informative as required--note: we might also see this as following the maxim of manner "avoid ambiguity").

Bearing this in mind, now think of the language of advertising.

What do you understand from such a claim (in normal language use)? The answer is probably something like: If you use ABC filters, these filters will remove all the bad bacteria from your drinking water. But note that what you understand when you hear the sentence isn't actually entailed by the statement that ABC filters remove bacteria from your drinking water. The only entailment is that some bacteria is removed by ABC filters. In truth conditional terms, if any (even good) bacteria are removed by ABC filters, then the claim is True. But if that's the truth, I bet you would be bummed to have bought an ABC filter on the basis of the above claim, because you understood what was implicated, not what was entailed. Implicature is a hallmark of our communication, and we often understand implicature to be the same as entailment in our daily communication. It isn't.

At the risk of hammering you to death on this issue, let's take a closer look at the bacteria sentence in order to appreciate why what is happening is happening.

This sentence is called a generic because it has a possible general reading. Generic sentences have the property of making strong statements, as in "dogs bark", "cats meow", and so forth. That is, it is a general property of dogs that they bark. In the bacteria sentence, the generic reading, which is most often the default reading in such constructions, is "all bacteria". However, there is also a possible non-generic or literal reading of the sentence.

If the advertiser only means "some bacteria", there's a potential maxim of relevance violation, under the assumption that you are interested in the filter as a way of guaranteeing water purity. You can see how the advertiser can exploit the default nature of the generic reading in order to take advantage of your belief upon hearing the claim "ABC filters remove bacteria" that what is meant is all bacteria or all harmful bacteria.

The point is this: advertising often takes advantage of implicature to make claims that people interpret to be more powerful than they really are. Bearing this in mind, here is an examples of a real advertising claim taken from your text book. Think about what is implicated in the following case versus what is literally entailed.

Although the implicature is a bit vague here, it is completely reasonable to assume that the soup has one-third less salt than some relevant benchmark. That benchmark might be the salt level of a major competitor, of the average of all nationally sold soups, of an earlier incarnation of Campbell's soup. Gricean maxims lead us to this, because it would be totally irrelevant to claim that the soup has one third less salt than the Dead Sea. But, literally, all it would take would be for this to be true for the advertiser to claim that the sentence is literally true. Here's a good case in which communication can be misleading, since we as readers of the slogan reasonably make implicatures that aren't necessarily true.

Your book notes that the degree to which we are sometimes misled can be surprising. They note an old claim by Ford that "...the Ford LTD is 700% quieter." This is a great example, since Ford apparently admitted that they meant "quieter than the outside". You don't even have to be cynical by nature to see that Ford was testing the line between vague claims and deceit in this case.

Here's a few more slogans. What's implicated versus what is literally entailed?

Bottom line: The difference between logical entailment and implicature lies at the root of why many advertising claims are misleading. My point here is not to vilify advertising in general, but rather, to call your attention to the limitations of logical entailment for understanding how meaning works. Implicature plays a huge role in our daily communication with one another. Advertising takes advantage of this. So do politicians, and so do we when we choose to mislead without, literally speaking, lying.

Here are a couple of cases from last week's Bush/Gore debate that struck me as interesting. I'm paraphrasing here, but note the following claims made by each:
 

Think of these as advertising slogans. Both have implicatures that seem to be at odds with each other, yet both might be literally true. What's implicated in Gore's claim? What's implicated in Bush's?