下载掌阅APP，畅读海量书库
立即打开
Pragmatics is not the same thing for all of its practitioners. Some pragmatic theorists see it as the study of language use in general, some as the study of communication, others as an approach to the study of language via language's communicative function.
There is some agreement that questions about speaker meaning and how people communicate are at the centre of pragmatics, but even theorists who accept this have differing views of pragmatics' methods and goals. One of the leading pragmatic theorists, Deirdre Wilson, notes that there are three approaches, broadly speaking. Pragmatics can be seen as a part of philosophy: an attempt to answer certain questions about meaning, in particular the relation between what sentences mean and what speakers mean when they utter them. Alternatively, it can be seen as an extension of the study of grammar in order to take into account and codify some of the interactions between sentence meaning and context. On this view pragmatics belongs to linguistics. Finally, pragmatics can be pursued as an attempt at a psychologically realistic account of human communication; this would make pragmatics part of cognitive science.
Despite these differences about the scope, aims and methods of pragmatics, there is considerable agreement on four fundamentals, particularly among those who focus on communicative use of language. All four points derive from the work of the philosopher Paul Grice:
1. Communication involves a certain complex intention which is fulfilled in being recognized by the addressee.
2. The addressee has to infer this intention from the utterance, a form of inference to the best explanation.
3. Communication is governed by principles or maxims. It is usually assumed that these principles derive from more general principles of rationality or cognition. Griceans, neo-Griceans and relevance theorists propose differing principles.
4. There is a distinction between what a speaker conveys explicitly and what she implicates , which are both aspects of speaker meaning or 'what is communicated'. Many theorists would also claim that speaker meaning includes another component or components. The list of components that have been proposed includes presupposition, conventional implicature and illocutionary force.
These four fundamentals are explained below, starting with the distinction between what speakers state and what they implicate.
The central data for pragmatics are cases in which a speaker, in making an utterance, conveys something more than, or different from, the meaning of the words she uses. There are many examples, and these examples fall into different types. In example (1), the second speaker is answering the first speaker's question. But if we look at the words used, we see that what B has said does not in itself provide any answer. B has stated that he had a haircut yesterday, but that does not entail that he does not want another one, nor that he does. On the other hand, B clearly intended his utterance to convey an answer to A 's question. In the terminology introduced by the philosopher Paul Grice in his famous 'Logic and Conversation' lectures in 1967, pragmatic the orists say that B (or B 's utterance) implicates that B would not like a haircut.
(1) A: Would you like a haircut?
B: I had one yesterday.
An implicature is a implication that the speaker intended to convey, according to the simplest definition: not, in general, a logical entailment of the sentence uttered, but something that may be inferred from the fact that the sentence was uttered, and uttered in a certain way, in a certain context.
Once the distinction is made between what a speaker expresses explicitly with the words she says, and what she implicates in saying them, examples appear everywhere. One example that Grice gave has become particularly famous. A professor is asked to provide a letter of recommendation for a student who is a candidate for a philosophy job, and he writes the letter in (2):
(2) Dear Sir,
Mr. Jones's command of English is excellent, and his attendance at tutorials has been regular.
Yours etc.
In writing this, the professor implicates that the student is no good at philosophy, since if he had been able to say something good about his philosophical abilities, he should have done so.
The distinction between what a speaker's words mean and what she means is also displayed in figures of speech such as irony, metaphor, understatement and hyperbole. Imagine example (3) said by someone who has been waiting for a friend when that friend finally turns up, well past the prearranged time. In this example of irony , the speaker does not endorse what she seems to be saying, and means something quite different, although related in that it is a comment on the friend's punctuality.
(3) The best thing about you is that you are always on time.
Most pragmatic theorists would follow Grice in taking what the speaker actually means by (3) to be an implicature. Grice extended this treatment to other figures of speech, although here there is less agreement about whether they involve implicature or some other way in which a speaker's meaning differs from the meaning of the words she says. What is not in doubt is that in examples like (4), (5) and (6) the speaker's meaning is different from the standard meaning of the words. In (4) Shakespeare has Macbeth express metaphorically his reason for continuing to kill in order to hold on to power. Obviously Macbeth does not literally mean by his utterance what the words literally mean.
(4) I am in blood/Stepp'd in so far that, should I wade no more, / Returning were as tedious as go o'er.
Understatement , as in (5), and hyperbole , as in (6), are the converse of each other: in one the speaker means more than her words mean, in the other, less.
(5) I am a bit hungry. (Said by someone who hasn't eaten for days.)
(6) I'm starving. (Said by someone whose last meal was a few hours previously.)
Loose use is a related phenomenon. If a speaker utters (7) she will be taken as committed to something less precise than her words might suggest. In most circumstances we would not find what she says misleading if she was in fact driving at 60.3 miles per hour, or even 58 miles per hour, but would not expect her to have uttered (7) if her speed was closer to 50 or 70 miles per hour.
(7) I was driving at 60 miles an hour.
As well as implicatures and figures of speech there are more mundane ways in which what a speaker means goes beyond what her words mean. Words like 'he', 'they', 'this', 'that', 'here' and 'there' are known as indexicals . The linguistic meaning of an indexical underspecifies (i.e. does not fully determine) the meaning that it conveys in use. Reading (8) out of context we do not know who the speaker met, where, or when it happened. We need to know something about the context in order to work out which person, time and location the speaker intended to refer to, that is to assign reference to the indexical words.
(8) I met her the previous day, just there.
The problem created for a hearer by ambiguous linguistic expressions such as the headline in (9) is somewhat similar. Here there is a choice between two structures that the speaker may have intended and the hearer must disambiguate : that is, choose the intended sense.
(9) Crocodiles alert as floods hit Australia.
Reference assignment and disambiguation are usually treated as necessary elements in recovering what a speaker says (to use Grice's terminology) or what she expresses explicitly (as some other theorists would prefer to say), in contrast to what she implicates. What ambiguous examples and examples with indexicals have in common with cases of implicature is that knowing the linguistic meaning of the words uttered is not enough for a hearer to know what the speaker meant.
There are some further ways in which speakers typically mean more than the linguistic meaning of words they have uttered. For example, an utterance of the sentence in (10) might be a promise, a threat, a prediction or an order, or, with different intonation, a question.
(10) Third battalion will retake the ridge by nightfall.
In pragmatics, the difference between a statement, an order and a promise is said to be a difference in illocutionary force , using terminology introduced by the philosopher J. L. Austin in his work on speech acts. Speech acts can be indirect: not every promise begins 'I promise to ...', not every prediction begins 'I predict that ...' and so on. Therefore the illocutionary force intended by a speaker may go beyond the words that the speaker has uttered (and typically does). A hearer of the utterance in (10) has to work out from clues in the context what force the speaker intended.
There are also examples where the speaker seems to take something for granted, or require the hearer to accept it as taken it for granted. For example, in uttering (11), a speaker is expecting his audience to take it from him that he has a cousin and his cousin has or had a grandmother. Similarly, a speaker uttering (12) apparently takes for granted (or ' presupposes ', to use the usual technical term) that John used to smoke and that he has been trying to give up.
(11) My cousin's grandmother was the first woman at the South Pole.
(12) Has John managed to give up smoking yet?
Attention was drawn to cases like (11) in the 1950s by the philosopher Peter Strawson, who had studied with Grice. From around 1970 a great deal of attention has been devoted to the phenomenon of presupposition by linguists and philosophers. It is often claimed that examples like these show that what is presupposed is a distinct level of speaker meaning, in addition to what a speaker expresses explicitly, her intended illocutionary force, and what she implicates. On the other hand, some theorists argue that information that is taken for granted is part of what is implicated.
The suggestion that this discussion has been implicitly making is that we could see pragmatics as the study of what is communicated (or what a speaker means) minus the linguistic meanings of the words uttered. Some clarification of what is meant by this is necessary. First, one might wonder what the 'linguistic meaning' of a word is. Roughly, the answer is that the linguistic meaning of a word is its stable meaning: it can be thought of as the meaning that it would have even if no one happened to use it, or as its meaning in the mental lexicon of a native speaker of the language.
A second point is that semantics , the study of the meanings of linguistic expressions, is of course more complicated than just a list of words with their meanings. For one thing, we should talk of lexical items rather than words, since (a) some idioms (fixed bits of language bigger than words) have fixed meanings—for example 'kick the bucket' has a meaning that cannot be predicted from the linguistic meanings of its parts: 'kick', 'the' and 'bucket'—and (b) so do some sub-parts of words (morphemes). What is more, the meanings of phrases and sentences are not formed just by adding up the meanings of the lexical items in them, as the following examples show. These sentences contain the same lexical items but have quite different (although related) meanings:
(13) Dogs hate cats.
(14) Cats hate dogs.
It is the job of semantic theory (in combination with syntactic theory) to explain these kinds of facts about the meanings of phrases and sentences: how they depend on the meanings of their parts and the way that those parts are put together.
Pragmatics can then be defined as the study of what is communicated (or speaker meaning) minus the part that semantics deals with: PRAGMATICS = SPEAKER MEANING − SEMANTICS.
One rather odd consequence of this definition would be that on some views of what semantics does, this will make the fixed linguistic meanings of certain words fall into pragmatics. One definition of a semantic theory for a natural language is that it is something that tells you the truth conditions of the sentences in that language. Certain words have linguistic meanings that are wholly or partly non-truth-conditional: an example is 'but'. Examples (15) and (16) are both true in just the same circumstances: if (and only if) John likes cake and Mary loves biscuits. (In other words, they have the same truth conditions.) However, (16) conveys in addition that there is some kind of contrast between John's liking of cake and Mary's love of biscuits.
(15) John likes cake and Mary loves biscuits.
(16) John likes cake but Mary loves biscuits.
Grice suggested that the extra meaning contributed by 'but' over and above the truth conditions of the sentence is a conventional implicature . Conventional implicature differs from the implicatures discussed above, which are conversational implicatures , in that it arises from the linguistic meaning (also sometimes called the conventional meaning) of a word, rather than from the conversational situation.
Since 'but' contributes to truth conditions (as 'and' does) and also to non-truth-conditional meaning, on the proposed definition the study of its meaning would belong to both semantics and pragmatics, and in fact pragmatics is often taken to include the study of non-truth-conditional meaning. A different view of the distinction between pragmatics and semantics allocates all linguistically encoded meaning to semantics. On this view, the study of lexically encoded non-truth-conditional meaning falls under non-truth-conditional semantics.
There is a more fundamental objection to the view that pragmatics simply fills in what semantics cannot explain about linguistic communication. One sign of this complication is that we can communicate without using words at all (and without using non-linguistic signs with fixed meanings, like thumbs-up for 'OK'). Suppose Mary is eating and has her mouth full and John asks:
(17) John: What did you do today?
Mary mimes writing, sealing envelopes, sticking stamps on them .
Mary is communicating that she wrote letters that day. Since her action is intended to be communicative, it counts as an utterance, in the technical sense used in pragmatics. We can say that in making this utterance, she conveys that she wrote letters that day.
How does this work? How can John work out what Mary wants to convey? In a lecture Grice gave in 1948 (which was published as a paper in 1957), Grice set out his theory of meaning. According to Grice, the words or gestures that a speaker utters are a clue to speaker meaning, where speaker meaning is analysed in terms of an intention that the speaker has, which Grice called the speaker's M (eaning)-intention. This M-intention can be decomposed into a number of separate intentions, and these intentions are nested, or stacked. The basic intention is to produce a certain response in the addressee. In the example, Mary wants John to entertain the idea that she wrote letters that day. There is a further intention that the addressee realizes that the speaker is trying to get something across. Mary does not want John to just suddenly entertain the idea that she has been writing letters; she also intends him to realize that she wanted him to realize that. There may be still more levels of intention involved in communication: Grice postulated a third level and there has been much debate about whether the three intentions are necessary or sufficient.
The crucial point is that on Grice's account, the speaker's recognition of the communicator's M-intention fulfils that M-intention. Grice's examples make the point clear. Suppose there is someone in your room and you want him to leave. You might bodily throw him out, or you could (if you knew he was avaricious), throw some money out of the window into the street. If he left just because he wanted to go and collect the money, then no communication is involved. But you might get him to leave (or to consider leaving, at least) by letting him know that you wanted him to. You could do this linguistically, by saying something like 'I think you should leave now', or non-linguistically, by giving him a little push. In this last case he would wonder why you pushed him, and he might infer that you wanted him to realize that you wanted him to leave.
In explanations of this sort, the addressee witnesses the speaker (or communicator, more broadly) behave in a certain way, that is, to make a gesture or utter a phrase, and infers that the best explanation for that behaviour is that the speaker had a certain intention. According to this view of communication, even an uttered sentence is, strictly speaking, only a clue to what the speaker meant. Of course sentences are generally much more specific and detailed clues than gestures, since they can carry a great deal of linguistically encoded information. But in principle the addressee must infer what relation there is between the information encoded in the phrase uttered and the intention with which the speaker produced that phrase, since it is the M-intention that ultimately matters in communication. On this view, pragmatic inference is necessary for all communication, not just those cases where speaker meaning outstrips sentence meaning.
Summarizing: according to Grice's theory of meaning, communication involves (a) inferential recovery of (b) certain speaker intentions. Most pragmatic theorists accept this general picture.
An obvious question is how hearers can recognize the relevant speaker intentions, and how speakers can have reasonable confidence that their intended meaning will be understood. Grice suggested (in lectures given in 1967) that conversation is governed by certain rules and principles, and that hearers understand speakers on the assumption that they are either conforming with these rules, or that if they are not they have a good reason. Specifically, Grice proposed a Cooperative Principle (CP) and several conversational maxims. The idea is that a rational speaker will try to be helpful and therefore she will generally aim to meet certain standards, described by the maxims. There are maxims of quality (truthfulness), quantity, relevance and manner: a speaker should try to give information that is true; will try not to give information that she does not have sufficient evidence for; will try to give as much information as is required, not too much and not too little; will try to give relevant information; and will try to make the way that she says things clear and easy to understand.
While the claim is that the CP and maxims govern conversation generally, Grice was particularly interested in showing that they could explain implicature, as in examples (1) and (2) above. If a speaker's utterance appears to violate the CP or one of the maxims, then the hearer may still assume that the CP is in effect. On the assumption that the speaker was trying to be helpful, she must have had a reason for saying something that apparently violates a maxim. What could that reason be? If what a speaker has said does not, in itself, conform to all the maxims, that could be because the speaker wanted to get across something else in addition to, or instead of what she was saying, namely an implicature. What is communicated overall is then still truthful, informative, relevant etc., and the assumption is vindicated that the CP is still in effect.
Most pragmatic theorists agree that conversation (and communication more generally) is governed by principles. Some accept the CP and maxims as Grice proposed them, but other theorists have proposed alternative systems, mostly with fewer principles and rules. A common criticism of Grice's theory is that there are too many maxims and too many ways of generating implicatures, including blatant violations of maxims, apparent violations and clashes between maxims.
Neo-Griceans propose two principles (Horn's system) or three (Levinson's). Horn has a Q-principle, which says that the speaker should be maximally informative, and a R-principle, which says that the speaker should not say too much. These principles are opposed but complementary, and are seen as manifestations of a fundamental tension in language and language use between explicitness and economy.
Relevance theory has a single Communicative Principle of Relevance: that each utterance raises a presumption of its own optimal relevance. Essentially the claim is that in making an utterance a speaker takes up some of her hearer's attention and this means that there is a fallible presumption that what she says will provide a good (in fact optimal) pay-off in information, relative to the cost involved in processing it. The Communicative Principle and presumption of optimal relevance are specific to communication, but they are argued to be instances of a more general tendency that cognition tends to be geared to maximize relevance.
The description of pragmatics given above provides some hints about its history, in particular Grice's centrality to its development. This section adds a few details and dates without attempting a comprehensive account. For the sake of simplicity, a crude division is made into three periods: (1) the prehistory of pragmatics, from antiquity until Grice's lecture on meaning; (2) a classical period from the 1940s to the 1960s, during which time Grice was still working on meaning and developing his theory of conversation, and Austin was working on speech acts; and (3) the modern period, starting with the dissemination of Grice's William James lectures from 1967. The middle, classical period is dealt with first.
Grice's work on meaning and conversation was not conducted in isolation. A number of other Oxford philosophers were actively involved in related work during the 1940s, 1950s and 1960s. The most important contributor apart from Grice was J. L. Austin, whose work on speech acts has already been mentioned. Another reason why Austin was a key figure is that he was the organizing force of a group of philosophers whose work became known as ordinary language philosophy. The prevailing method among these philosophers was 'linguistic botanizing': paying close attention to the distinctions made by ordinary language on the assumption that the way people speak makes many subtle distinctions that are worthy of philosophical investigation. These philosophers were not necessarily interested in studying language as such, as linguists are, but they found themselves drawn into thinking about such questions as what saying or stating involves and what else speakers do with language. Grice's theories of conversation and meaning and Austin's views on speech acts are, in effect, different (perhaps complementary) answers to these questions.
In addition to Grice and Austin, other philosophers from this group whose work has had an impact on pragmatics include Peter Strawson, J. O. Urmson, R. M. Hare and Stuart Hampshire. Of these Strawson has probably had the most influence: through reintroduction of the idea of presupposition, mentioned above, and because of an influential criticism that he made of Austin's conception of speech acts.
Austin was particularly interested in how certain speech acts create social facts, for example, the speech act of naming a ship. Once the act has been successfully performed, the ship has its new name, by virtue of social conventions. Austin pointed out that there are conditions that have to be met for a speech act to be successful: felicity conditions. The felicity conditions for naming a ship and for other institutional speech acts, such as declaring a defendant guilty and passing sentence, are also social in their character: the act must be performed by the right person, at the right time, in the right way, using a proper form of words and so on.
According to Strawson, Austin's interest in these institutional cases led him to neglect the important point that many speech acts are not in this sense social. A more Gricean view is that what is important for successful communication is the recognition of the intention of the speaker to perform a particular speech act. For example, if a speaker utters an interrogative sentence, did she mean what she said as a genuine request for information, or a rhetorical question, or with some other force?
Another difference between Austin and Grice's views is that Grice was more inclined to separate use of language from meaning. Grice saw this as a desirable corrective to the ordinary language style of philosophizing. For example, Austin was aware of something similar to what is sometimes known as 'the division of pragmatic labour'. In his article 'A plea for excuses' he expounded the idea that there is 'no modification without aberration'. The idea is that language has a natural economy, so we can only use a modifying expression 'if we do the action named in some special way or circumstance'. As an illustration of the point, he said that it would be redundant and misleading, rather than false, to say that someone sat down intentionally , unless there was some special reason to say it that way. People normally sit down intentionally, but we would only go to the trouble of including the modifier 'intentionally' if there was some doubt about whether the action was deliberate or not: perhaps it looked accidental.
While Grice agreed with the general point (and used a similar argument in a paper on perception, written around the same time), he wrote in a private note (cited in Siobhan Chapman's biography of Grice) that the principle as Austin states is wrong. In many cases it is more natural to use a modifier than not. Grice noted that 'aberrations are only needed for modifications that are corrective qualifications'. As he wrote, no aberration is required to justify the phrase 'in a taxi' as a modifier in 'He travelled to the airport in a taxi.'
In addition, Austin interpreted the principle as concerning language, not language use. He said that use of the modifier in the wrong circumstances is not only not required, but actually impermissible. Here he missed the essentially Gricean point that a speaker can always say things in a way that the hearer might find uneconomical or surprising, and that if she does, then typically the hearer will look for an interpretation that justifies the use of the unexpected choice of words.
The post-war Oxford philosophers did not generally use the term 'pragmatics' in their work on language use, although it had already been proposed as a label for the study of meaning in use by the American philosopher Charles Morris. In his Foundations of the Theory of Signs (1938) he distinguished between syntax, the study of the formal relations of signs to each other, semantics, the study of the meaning of signs in terms of the objects that they denote or might denote and pragmatics, 'the science of the relation of signs to their users' (p. 29). He expanded on this in his Signs, Language, and Behavior (1946): pragmatics 'deals with the origins, uses, and effects of signs within the total behavior of the interpreters of signs' (p. 219). Morris' views are one origin of the very broad conception of pragmatics as the study of language use in general mentioned at the beginning of this introduction.
Of course, interest in language use, communication and the difference between what is said and what is meant did not start with Morris' definition or the work of the Oxford philosophers. Since antiquity, philosophers and rhetoricians have been interested in cases in which speakers mean something different from what they say. We might (somewhat flippantly) call this period the prehistory of pragmatics. The linguist Larry Horn has traced some of the central concerns of modern pragmatics back to the work of earlier writers, for example in his book A Natural History of Negation and a more recent article, 'Presupposition and implicature'.
Classical rhetoricians were aware of figures of speech in which the speaker means something different from the words produced. According to Horn, the distinction between what is said and what is meant, and therefore between what is said and what is meant but not said, goes back at least to the fourth century rhetoricians Servius and Donatus, whose description of understatement is as a figure of speech in which we say less but mean more. Similarly, the classical definition of verbal irony is as a figure in which the meaning is the opposite of what one's words mean.
Horn has shown that in the nineteenth century, John Stuart Mill and Augustus de Morgan distinguished between the strict logical sense of 'some', which is compatible with all , and its use in common conversation, where use of 'some' often suggests not all , and that their explanations for the difference are thoroughly Gricean:
If I say to any one, 'I saw some of your children today', he might be justified in inferring that I did not see them all, not because the words mean it, but because, if I had seen them all, it is most likely that I should have said so: even though this cannot be presumed unless it is presupposed that I must have known whether the children I saw were all or not. (Mill, writing in 1867, p. 501)
There is also a prehistory to the concept of presupposition. Before Strawson, Gottlob Frege also thought that use of a singular referring expression presupposed the existence of the individual described, and Horn has shown that another nineteenth-century philosopher, Christoph von Sigwart, had a rather modern view of the subject. As Horn says, Sigwart's view 'that a presuppositionally unsatisfied statement is misleading or inappropriate though true foreshadows the pragmatic turn to come.'
The use of the word 'pragmatics' to describe a separate field of study, on a par with syntax and semantics, was established during the 1970s. Around this time the term was being used in a different way by philosophers concerned with formal languages. For the formal semanticist Richard Montague, writing in the late 1960s and following the way the linguist and philosopher Yehoshua Bar-Hillel used the term in the 1950s, pragmatics was the study of any language containing indexical terms. As Levinson observed in his classic textbook, this would make the study of all natural language fall under pragmatics, since all natural languages have indexical elements.
The modern use of the term 'pragmatics' was emerging by the late 1960s in philosophy. Robert Stalnaker's 1970 article 'Pragmatics' gives a definition: 'pragmatics is the study of linguistic acts and the contexts in which they are performed' and contrasts pragmatics with semantics, which, for Stalnaker, is the study of propositions.
According to Gerald Gazdar, 'pragmatics had become a legitimate subdiscipline in linguistics by the late 1970s but it wasn't in the early 1970s.' The key factors in the emergence of linguistic pragmatics appear to have been the impact of Grice's 'Logic and Conversation' lectures, circulated in mimeograph form from the late 1960s; the publication of some of the lectures as standalone papers; and the publication during the 1970s of a number of Ph.D. theses and subsequent works by linguists concerned with pragmatic topics, including Larry Horn, Ruth Kempson and Deirdre Wilson.
The current state and recent history of pragmatics are too diverse and complex to describe briefly. A few areas of interest may be picked out.
Early in the modern period, disagreement about the principles that govern communication led to fragmentation of the field into Griceans, neo-Griceans and relevance theorists, among others. There are also pragmatic theorists who work primarily on speech acts. In addition, the Journal of Pragmatics and the International Pragmatics Association represent a very wide variety of work falling under the broad conceptions of pragmatics as the study of language use in general and the study of language through its use.
From the 1970s many theorists have been interested in developing formal accounts of phenomena where this seems possible, particularly scalar implicature, presupposition and conventional implicature. This work is now known as formal pragmatics, and has close links to dynamic approaches to semantics such as Discourse Representation Theory.
In more cognitively oriented work, including relevance theory, there has been interest in the structure of the mind and in how pragmatic inference is performed. The proposal in psychology that the mind/brain is massively modular has been influential across the cognitive sciences, and Sperber and Wilson have postulated that there is a dedicated pragmatics module. Work in psychology on mindreading (or 'theory of mind'), the ability humans have to infer other's mental states from observation of their actions, also appears to be of direct relevance to pragmatics.
Pragmatic inference is fast and seems not to be hugely effortful. As the linguist Gilles Fauconnier says, there is an 'illusion of simplicity', given that the task performed is actually rather complex. Some pragmatic theorists have recently been exploring the possibility of adopting insights from research into fast and frugal heuristics. This research programme aims to show that cognition uses simple, well-adapted mechanisms to solve complex problems rapidly and accurately.
Another very recent development is the new field of experimental pragmatics, coming into being at the intersection of pragmatics, psycholinguistics, the psychology of reasoning and developmental pragmatics, the last of which is itself a relatively new area of work.
Finally, there has been a great deal of interest in the linguistic underdeterminacy thesis, the idea that the linguistic material in an utterance often or typically underdetermines the proposition expressed by the speaker. After considerable controversy, this thesis (in some version) is becoming a new orthodoxy, although rearguard actions are still being fought by those who want to maintain at all costs that propositional logical form mirrors sentence form. This debate has gone hand in hand with another about the relations between several distinctions: between semantics and pragmatics, the explicit and the implicit, the truth-conditional and the non-truth-conditional, and coding and inference. This research raises fundamental questions concerning the nature of human communication and the scope of pragmatic theory.
Robyn Carston's book, Thoughts and Utterances (2002) is an excellent summary of the debates around underdeterminacy and at the same time a major contribution to them from an influential pragmatic theorist. It addresses a wide range of issues in modern pragmatics and would be a good place to learn more about the current state of the art.
There is also a list of key works starting on page 266 below, and works that are especially suitable as introductory texts are marked.