用户名: 密码: 验证码:  
站内检索:
经典文献
暂无分类
 
首页 >> 学术资源 >> 经典文献
 
Pragmatics and Modularity
【发布日期:2009-11-15】 【来源:】【字体: 】 【打印本页】 【关闭窗口
 
                                          Pragmatics and Modularity
DEIRDRE WILSON AND DAN SPERBER
 
 
1. INTRODUCTION
 
Grammar and pragmatics are alike in two respects: they fall within the domain of cognitive psychology, and they have to do with language. Apart from that, we will argue, they have virtually nothing in common. Grammar is a special-purpose modular system: pragmatics is not a cognitive system at all. There are no special-purpose pragmatic principles, maxims, strategies or rules: pragmatics is simply the domain in which grammar, logic and memory interact. Modular grammatical processes offer little direct insight into the nature of non-modular pragmatic processes. Indeed, the more we model pragmatics on grammar, the more mistaken we are likely to be.
 
2. WHY PRAGMATICS IS NOT A MODULE
 
Utterance interpretation involves a variety of processes, grammatical and pragmatic. By 'grammatical processes,' we mean the processes used to recover the semantic representation of the sentence uttered (or in the case of ambiguity, its set of semantic representations). By 'pragmatic processes,' we mean the processes used to bridge the gap between the semantic representations of sentences and the interpretation of utterances in context. The goal of the hearer, and, by extension, of the pragmatic processes, is to recover not just some arbitrary interpretation, but the interpretation intended by the speaker this is the only interpretation it is worth the hearer's effort to recover.
Pragmatic processes are involved in every aspect of utterance interpretation: in the recovery of explicit propositional content, implicit import and Elocutionary force. Often, the sentence uttered is ambiguous or ambivalent, as in (1):
(1) His food is not hot enough
The hearer of (1) must not only recover the semantic representation of the sentence uttered, but decide who the referential expression 'he' refers to, whether the ambiguous word ‘hot' means very worm or spicy, whether the vague expression 'his food’ refers to the food he cooked, the food he brought, the food he served, the food he is eating, etc., and what this food is claimed to be not hot enough for.
Utterances have not only propositional content but Elocutionary force, and ambiguities or ambivalences may arise at this level, as in
  (2) You re not leaving
The hearer of (a) must not only recover its explicit propositional content, but also decide whether it is intended as a statement, a question or an order.
Utterances have not only explicit content but implicit import, as in (3):
  (3) a. Peter: Do you want some coffee?
       b. Mary: Coffee would keep me awake.
The hearer of (3b) must recover not only its explicit content but also the implicature that Mary doesn't want any coffee (or, in some circumstances, that she does).
Utterances may be metaphorical or ironical, as in (4) and (5):
  (4) Their friendship blossomed.
  (5) I've never had such a lovely new.
The hearer of (4) or (5) must not only recover its explicit content but also decide whether it was literally or figuratively intended, and what its figurative interpretation was intended to be.
More generally, the style of an utterance may affect its interpretation-compare the mildly witty (6) with its heavy-handed explicit paraphrase (7):
(6) Two taxis collided and 30 Scotsmen were taken to hospital. [Woody Allen]
(7) Scotsmen are very mean. They travel in enormously overcrowded taxis to avoid paying full
  fare. Once two taxis containing 30 Scotsmen collided. The passengers were taken to hospital.
These stylistic effects on the hearer must be described and explained. Pragmatics, then, is a theory of the cognitive principles and abilities involved in utterance interpretation, and of their cognitive effects.
Is pragmatics a module? For Fodor (1983), the paradigm example of a modular system is the linguistic system. By hypothesis, this incorporates a grammar, i.e. a code, which relates phonetic representations of sentences to semantic representations of sentences. Here, we will simply assume that there is a necessary connection between the modularity of the linguistic system and the fact that it incorporates a grammar or code. The claim that pragmatics is a module is thus essentially equivalent to the claim that there is a pragmatic code.
Many linguists have assumed without question that speakers of English know a pragmatic code, analogous to a grammar, which enables them to recover the intended interpretation of utterances in English. They have assumed, in other words, that pragmatics is a module, an extension of the grammar. Underlying this assumption is a more general assumption about the nature of communication: that communication necessarily involves the use of a code. Both assumptions seem to us to be false.
In the first place, pragmatic processes are highly context dependent. Different contextual assumptions lead to different pragmatic interpretations. If utterance interpretation is a matter of decoding the speaker s intentions, there must be some algorithm for selecting the appropriate set of contextual assumptions, the ones the hearer was intended to use. Yet there are at least two types of case[s] which suggest that no such algorithm exists.
Sometimes, a hearer is simply unable to access a contextual assumption he was intended to use in interpreting an utterance. Consider (8):
  (8) a. Peter: Ozzy Osbourne’s coming to dinner.
       b. Mary: I’11 bring a bat.
To disambiguate the word ‘bat' in (8b), the hearer would need access to the information that Ozzy Osbourne is a rock musician mainly known for having bitten the head off a live bat on stage. What algorithm could he use to recover this assumption from the exchange in (8)?
More often, a hearer has access to a variety of alternative contextual assumptions, which yield a variety of incompatible interpretations, only one of which could have been intended by the speaker. Suppose that as I give a lecture, I make a slip of the tongue. You turn to your neighbor and whisper
  (9) That was interesting.
What algorithm could your neighbor use to decide that you intended to refer to the slip of the tongue I had just made, rather than, say, to the example I had just been discussing, the theoretical claim I had just made, or the fact that a strange bird had just Down past the window? But if there is no algorithm for identifying the intended set of contextual assumptions, there can be no algorithm for recovering the intended overall interpretation of an utterance.
A more general problem for modular approaches to pragmatics is their failure to deal with indeterminacies in interpretation. According to the code model, the speaker's thoughts, encoded into an utterance, should be replicated in the hearer by a decoding process. The result of verbal communication should be an exact reproduction in the hearer of the thoughts the speaker intended to convey. However, the most cursory examination of ordinary conversation reveals that in the case of implicit import, figurative interpretation and stylistic effects, such reproduction is rarely intended or achieved. The existence of indeterminacies in interpretation suggests a fundamental inadequacy in modular approaches to pragmatics. Where indeterminacy is involved, it seems that the most that communication can achieve is to bring about some similarity between the thoughts of communicator and audience: but the code model can provide no interesting account of those cases where similarity, rather than identity, is intended and achieved.
Finally, it is easy to show that communication is possible in the absence of a code. Consider (10):
  (10) a. Peter: How are you feeling today?
  b. Mary: [Takes a bottle of aspirin out of her bag and shows it to Peter.]
Here, Mary communicates that she is not feeling very well, even though there is no code or convention which says that showing someone a bottle of aspirin means that one is not feeling well. Intuitively, Peter does not need a code to understand Mary's behavior in (10) because he can use his knowledge of the world and his general reasoning abilities to work out what she must have intended to convey. On this account, communication is achieved not by coding and decoding messages, but by providing evidence for an Mended inference about the communicator s intentions. Might this not be the case for verbal communication too?
The assumption that pragmatic processes are modular thus seems to us neither necessary nor plausible. But what exactly does it mean to claim that they are not modular According to Fodor, the paradigm example of a non-modular process is scientific theorizing. Scientific theorizing is a non-demonstrative inference process. Like all such processes, it can be broken down into two stages: hypothesis formation and hypothesis confirmation. Both stages exhibit what Fodor regards as the defining feature of non-modular processes they are informationally unencapsulated That is, they have free access to contextual information. There is no piece of evidence, however remote, no hypothesis, however implausible, that might not turn out to have a bearing on their outcome. Grammatical process by contrast, have highly restricted access to contextual information: they are unaffected by the hearer's non-linguistic beliefs.
Pragmatic interpretation seems to us to resemble scientific theorizing in essential respects. The speaker's intentions are not decoded but non-demonstratively inferred, by a process of hypothesis formation and confirmation which, like scientific theorizing and unlike grammatical analysis, has free access to contextual information. The hearer's aim is to arrive at the most plausible hypothesis about the speaker s intentions; but the most plausible hypothesis, in pragmatic interpretation as in science, may still be wrong.
There is thus a fundamental difference between the modular and the non-modular approaches to pragmatics. A fundamental am gumption of the non-modular approach is that the interpretation of communicative behavior, like the interpretation of evidence in general, is always subject to risk. There are always alternative ways of interpreting a given piece of evidence, even when all the correct procedures for interpretation are applied. These procedures may yield a best hypothesis, but even the best hypothesis may not be the correct, i.e., the intended, one. By contrast, decoding procedures, when correctly applied to an undistorted signal, guarantee the recovery not only of an interpretation, but of the correct, i.e. the intended interpretation. The two approaches start from radically different assumptions about the nature of communication itself.
Fodor is sceptical about the amenability of non-modular processes to scientific study:
…the limits of modularity are also likely to be the limits of ~ hat we are going to be able to
understand ghoul the mind, given anything like the theoretical apparatus currently available.
(Fodor 1983:126)
As he rightly points out, virtually nothing is known about either the logic or the psychology of scientific hypothesis formation and confirmation. If pragmatic interpretation is just a special case of scientific theorizing, it seems that the search for an explanatory pragmatic theory is doomed to fail.
Certainly, recent work in pragmatics gives little cause for optimism. Formal pragmatists simply assume that pragmatics is a module, and provide mechanical analyses for a severely restricted range of data. Gricean pragmatists, who approach pragmatics in inferential terms, often fall back on the assumptions of the code model, and offer nothing approaching an explicit theory. In the last few years, we have been trying to develop an explicit, explanatory theory of pragmatic interpretation. We will try to show that it offers interesting solutions to a variety of empirical problems arising from more traditional accounts. (For a full account of this theory, and the assumptions behind it, see Sperber and Wilson 1986).
 
3. COGNITION: RELEVANCE
 
Grice made two major contributions to pragmatics. First, he outlined an inferential alternative to the code model of communication. Second, he suggested a method by which pragmatic hypotheses might be confirmed or disconfirmed. The idea was that, given a range of hypotheses about the communicator’s intentions, the hearer should discard any that are incompatible with the assumption that the cooperative principle and maxims of truthfulness, informativeness, relevance and brevity have been observed. For a variety of reasons, this suggestion does not amount to an explanatory theory: crucial terms such as 'relevance,' 'cooperation,' 'brevity,' 'required information,' 'purposes of the exchange, were left undefined; no account of the role of contextual information, nor of the processes of pragmatic hypothesis formation, were offered, and the origin of the maxims themselves was left unexplained. However, it does indicate what we should be trying to do. We should be trying to develop an explicit criterion powerful enough to eliminate all but a single hypothesis about the communicator’s intentions. The rest of this paper will be devoted to developing such a criterion and showing how it works.
We believe that this criterion has its source in some basic facts about human cognition. Humans pay attention to some phenomena rather than others: they represent these phenomena to themselves in one way rather than another: they process these representations in one context rather than another. What is it that determines these choices? Our suggestion is that humans tend to pay attention to the most relevant phenomena available; that they tend to construct the most relevant possible representations of these phenomena, and to process them in a context that maximises their relevance. Relevance, and the maximization of relevance, is the key to human cognition.
This has an important consequence for the theory of communication. A communicator, by the very act of claiming an audience's attention, suggests that the information he is offering is relevant enough to be worth the audience's attention. We would like to show that this simple idea-that communicated information comes with a guarantee of relevance- is enough on its own to yield an explanatory pragmatic theory.
But what is relevance? We claim that information is relevant to you if it interacts in a certain way with your existing assumptions about the world. Here are three examples of the type of interaction we have in mind.
Case A
You wake up with the following thought:
  (11) a. If it's raining, I’ll stay at home.
You look out of the window and discover
  (11) b. It's raining.
In this case, from your existing assumption (I la) and the new information (11b), you can deduce some further information not deducible from either the existing assumption or the new information alone:
  (11) c. I'll stay at home.
To deduce (11c), you have to use both old and new information as joint premises in an inference process. Intuitively, the new information (11b) would be relevant in a context containing assumption (1la). We claim that it is relevant precisely because it enables such a joint inference process to take place. Let us say that assumption (1la) is the context in which the new information (11b) is processed, and that (11b) contextually implies (11c) in the context (11a). Then we claim that new information is relevant in any context in which it has contextual implications, and the more contextual implications it has, the more relevant it will be.
Assumptions about the world may vary in their strength: you may have more or less evidence for, more or less confidence in, your assumption that it is raining. New information may affect the strength of your existing assumptions, as in the following case.
Case B
You wake up, hearing a pattering on the roof, and form the hypothesis that:
  (12) a. It s raining
You open your eyes, look out of the window, and discover that:
  (12) b. It Is raining.
Here, the new information (12b) strengthens, or confirms, your existing assumption (12a). It would also, intuitively, be relevant to you in a context containing assumption (12a). Vice claim that (12b) is relevant precisely because it strengthens an existing assumption of yours. New information is relevant in any context in which it strengthens an existing assumption; and the more assumptions it strengthens, and the more it strengthens them, the more relevant it will be.
If new information can achieve relevance by strengthening an existing assumption, it should also achieve relevance by contradicting, and eliminating, an existing assumption, as in the following case.
Case C
You wake up, as in case B, hearing a pattering on the roof, and form the hypothesis that:
  (13) a. It s raining
This time, when you open your eyes and look out of the window, you discover that the sound was made by leaves falling on the roof, and that actually:
  (13) b. It's not raining.
Let us assume that when new and old assumptions contradict each other, the weaker of the two assumptions is abandoned. Here, the new information (13b) would provide conclusive evidence against the old assumption (13a), which would therefore be abandoned. Intuitively, (13b) would be relevant in these circumstances. We claim that new information is relevant in any context in which it contradicts, and leads to the elimination of, an existing assumption: and the more assumptions it eliminates, and the stronger they were, the more relevant it will be.
These cases illustrate the three ways in which new information can interact with, and be relevant in a context of existing assumptions: by combining with the context to yield contextual implications; by strengthening existing assumptions; and by contradicting and eliminating existing assumptions. Let us group these three types of interaction[s] together and call them contextual effects. Then we claim that new information is relevant in any context in which it has contextual effects, and the greater its contextual effects, the more relevant it will be.
This comparative definition of relevance is inadequate in one respect, as the following example shows:
Case D
You wake up, thinking:
  (14) a. If it rains, I'll stay at home.
Then EITHER: You look out of the window and see:
  (14) b. It’s raining.
OR:
You look out of the window and see:
  (14) c. It's raining and there's grass on the lawn.
Intuitively, (14b) would be more relevant to you than (14c) in the context (14a). Yet (14b) and (14c) have exactly the same contextual effects in this context: they both have the contextual implication (14d), and no other contextual effect at all:
  (14) d. I'll stay at home.
If comparisons of relevance are based solely on contextual effects, then the difference in relevance between (14b) and (14c) is inexplicable.
This difference, we suggest, can be explained in terms of the intuition underlying Grice's Manner maxims, which itself derives from some basic facts about cognition. The intuition is that speakers should make their utterances easy to understand: in our terms, that speakers should make the contextual effects of their utterances easy to recover. Now it is clear that though (14b) and (14c) above have exactly the same contextual effects in the context (14a), you would have to work harder to recover them from (14c) than from (14b): since ( 14c) includes ( 14b) as a subpart, (14c) will require all the effort needed to process (14b), and more besides. This extra processing effort detracts from the relevance of the information in (14c), and of any utterance used to communicate it.
We thus propose the following comparative definition of relevance (developed in more detail in Sperber and Wilson 1986):
  Relevance:
    (a) Other things being equal, the greater the contextual effects, the greater the relevance.
    (b) Other things being equal, the smaller the processing effort, the greater the relevance.
An individual with finite processing resources, who is aiming to maximize relevance, should pay attention to the phenomena which, when represented in the best possible way, and processed in the best possible context, seem likely to yield the greatest possible contextual effects in return for the available processing effort. Relevance, and the aim of maximizing relevance, is the key to cognition.
 
4. COMMUNICATION: THE PRINCIPLE OF RELEVANCE
 
If humans pay attention only to relevant information to communicator, by claiming an audience's attention, gives a guarantee of relevance. He guarantees, in particular, that the information he is attempting to convey, when processed in a context he believes the audience has accessible, will be relevant enough to be worth the audience's attention. But how relevant is that? What exactly is the guarantee of relevance that accompanies each act of inferential communication?
On the contextual effect side, the guarantee is one of adequacy. In the most straightforward cases of verbal communication, the speaker guarantees that the proposition he intends to express, when processed in a context he expects the hearer to have accessible, will yield enough contextual effects to be worth the hearer's attention. How much is required in the way of contextual effects will vary from individual to individual and occasion to occasion. How the level of adequacy is fixed and varies is an interesting question, but intuitions about particular examples are clear enough.
On the processing effort side, as Grice's Manner maxims suggest, the guarantee is of more than adequacy. A speaker who wants to achieve a certain range of contextual effects must make sure that they are as easy as possible for the hearer to recover that is, he must make sure that his utterance puts the hearer to no unjustifiable processing effort. This is in the speaker's interest as well as the hearer's, for two reasons: firstly, the speaker wants to be understood, and any increase in unjustifiable processing effort required of the hearer is an increase in risk of misunderstanding; secondly, any increase in processing effort detracts from overall relevance, and might cause the overall relevance of the utterance to fall below an acceptable level.
Let us say that an utterance (or more generally an act of inferential communication) which, on the one hand, achieves an adequate range of contextual effects, and on the other hand, achieves it for the minimum justifiable processing effort, is optimally relevant. Then Grice's maxim of relevance can be replaced by the following principle of relevance:
Principle Relevance
    Every act of inferential communication carries a guarantee of optimal relevance.
We believe that this single principle (or rather a more technical version developed in Sperber and Wilson 1986) is enough on its own to yield an explanatory pragmatic theory.
The fact that an utterance carries a guarantee of optimal relevance does not mean that it will actually be optimally relevant to the hearer. A guarantee may be given mistakenly or in bad faith: I may tell you something in the mistaken belief that you do not already know it, or speak simply to distract your attention from relevant information elsewhere. In this case, you will be unable to find an interpretation which satisfies the guarantee.
Let us say that an utterance on a given interpretation is consistent with the principle of relevance if a rational communicator might have expected it to be optimally relevant to the hearer, i.e., to achieve an adequate range of contextual effects as economically as possible. Then it is easy to show that every utterance has at most one interpretation which is consistent with the principle of relevance.
We will show this using our example of disambiguation, (8b), with possible interpretations (15a) and (15b):
  (8) b. Mary: I'll bring a bat.
  (15) a. Mary: I’ll bring a living rodent.
         b. Mary will bring a hitting instrument.
Logically speaking, there are two routes that the disambiguation process might follow: one interpretation may be more accessible than the other, and be tested first for consistency with the principle of relevance: or both interpretations may be equally accessible, and be tested in parallel. We consider each possibility in turn.
Suppose that interpretation (15a) is more accessible to Peter than (15b), and is therefore the first to be tested for consistency with the principle of relevance. Suppose, moreover, that there is an easily accessible context in which this interpretation would have a manifestly adequate range of contextual effects, and that there would has e been no obviously cheaper way of obtaining them. Then as long as Mary could have foreseen this situation, interpretation (15a) is consistent with the principle of relevance, and is the only interpretation consistent with the principle of relevance, as the following argument shows.
Imagine that Mary had wanted to convey interpretation (15b), but had forseen that interpretation ( 15a) would be both more accessible and consistent with the principle of relevance. By reformulating her utterance to eliminate this unwanted interpretation, for example, saying I'll bring a baseball bat', thus eliminating interpretation ( 15a) entirely - she could have spared her hearer the effort of first accessing and processing interpretation (15a), then accessing and processing interpretation ( 15b), and then engaging in some form of inference process to choose between them. In other words, she could have achieved the intended range of contextual effects at a much reduced processing cost, and at a much smaller risk of misunderstanding, by rephrasing her utterance. On this interpretation, although Mary's utterance (8b) may achieve an adequate range of contextual effects, it would put her hearer to some unjustifiable processing effort in recovering them, and is not consistent with the principle of relevance.
What would happen if interpretations (15a) and (15b) were equally accessible, and were thus simultaneously tested for consistency with the principle of relevance? Suppose that Peter has easy access to a context in which interpretation (15a) has an adequate range of contextual effects, while a comparable context for ( I 5b) is much less accessible or not accessible at all. As long as Mary could rationally have foreseen this situation. (15a) is consistent with the principle of relevance, and is the only interpretation consistent with the principle of relevance. If Mary had intended to convey interpretation (15b), she could manifestly have spared Peter some processing effort by rephrasing her utterance to eliminate the unwanted interpretation (15a). For example, by saying 'I'll bring a baseball bat,' she could have spared Peter the effort of accessing and processing both (15a) and (15b), and then engaging in some inference process to choose between them. On this interpretation, Mary's utterance (8b) would 1put] Peter to some unjustifiable processing effort, and is not consistent with the principle of relevance.
Finally what would happen if interpretations (15a) and (15b) were equally accessible, and, moreover, yielded comparable contextual effects at comparable processing costs? Then there would be no way of choosing between them, the ambiguity would remain unresolved, and neither interpretation would be consistent with the principle of relevance, since each could only be preferred, if at all, after an effort of comparison which Mary could easily have spared Peter. Thus the principle of relevance provides an account, not just of successes, but also of failures of disambiguation.
This example shows that, whatever the procedures used in disambiguation, the first interpretation - if any - tested and found consistent with the principle of relevance is the only interpretation consistent with the principle of relevance. A speaker who does not intend this interpretation should rephrase her utterance to eliminate it. This general principle applies to every aspect of utterance interpretation, shedding light on a number of long-standing pragmatic problems, as our next section is designed to show.
 
5. PRAGMATICS AND RELEVANCE THEORY
 
Grice has a maxim 'Be brief.' There are a number of problems attaching to the maxim of brevity. One is that brevity itself is left undefined: should it be measured in terms of phoneme counts, syllable counts, word counts, or what? Another is that there are clear counterexamples. Compare (16a) and (16b):
(16) a. I have no brothers or sisters.
  b. I have no siblings
By any intuitive measure of brevity, (16b) is shorter than (16a), which is thus predicted as stylistically inappropriate on Grice's account. The fact is, though, that it is (16b) rather than (16a) which is the stylistically inappropriate member of this pair.
Relevance theory offers a solution to both these problems. There is no maxim of brevity. The intuitions Grice wanted to explain are intuitions about processing effort, and in particular, about the fact that a speaker aiming at optimal relevance should spare her hearer any unnecessary processing effort. Whereas 'brevity' is an ad hoc linguistic category, processing effort is a psychological category whose empirical causes and consequences are at least to some extent known. In particular, it is known that word frequency affects processing effort: in general, the rarer the word, the greater the processing effort. Now 'sibling' is a very rare word indeed. The differences between (16a) and (16b) are straightforwardly explained on the assumption that the relative brevity of the word ‘sibling' is not enough to offset the increase in processing cost resulting from its infrequency, so that (16a) is more economical overall.
Relevance theory also offers solutions to a variety of problems raised by traditional accounts of implicature. In our framework, implicatures are the contextual assumptions and implications which form part of the intended interpretation of an utterance. Consider (3) above:
  (3) a. Peter: Do you want some coffee?
        b. Mary: Coffee would keep me awake.
On what grounds might Mary have thought her utterance (3b) would be optimally relevant to Peter? In the circumstances, Peter can reasonably assume that Mary intended him either to supply the contextual assumption ( 17) and derive ( 18) as a contextual implication, or to supply the contextual assumptions in ( 19a-b) and derive (20) as a contextual implication:
  (17) Mary does not want to stay awake.
  (18) rotary does not want any coffee.
  (19) a Mary wants to stay awake.
         b. Mary wants anything that will keep her awake.
  (20) Mary wants some coffee.
But which of these possible interpretations did Mary have in mind?
The answer follows from the principle of relevance. Suppose that in the circumstances, assumption (17) is more accessible to Peter than assumptions (19a-b), and that Mary could have foreseen this. Then the interpretation based on (17) and (18) is the only one consistent with the principle of relevance. Suppose that Mary did not intend this interpretation: then she could have spared Peter the unnecessary effort of first accessing the interpretation based on (17) and (18), then accessing the interpretation based on (19) and (20), and then engaging in some inference process to choose between them. For example, by adding 'And I want to stay awake,' she could have eliminated the interpretation based on (17) and (18). In this situation, the interpretation based on ( 17) and ( 18) is the only one consistent with the principle of relevance. It follows that in the recovery of implicatures, as in disambiguation, the first interpretation tested and found consistent with the principle of relevance is the only interpretation consistent with the principle of relevance.
In fact, if Mary had merely intended to implicate that she didn't want any coffee, she could have conveyed this information more economically by simply saying 'No'. A speaker aiming at optimal relevance, who wanted to spare her hearer any unnecessary processing effort, must have intended to achieve some additional contextual effects, not derivable from the direct answer No.' Here, the most natural assumption for Peter to make is that she is refusing the coffee because it keeps her awake rather than, say, because he makes horrible coffee, or because she doesn’t want to spend a moment longer with him. Thus, (3b), unlike the direct answer ‘No,' simultaneously conveys a refusal and an explanation of that refusal. The principle of relevance explains both when an implicature is needed, and what the best hypothesis about its content will be.
Often, as in (4)-(6) above, the implicatures of an utterance are less determinate than this. The problem of providing an explicit account of the indeterminacy of implicatures has defeated many pragmatists. Relevance theory offers a solution. The implicatures of an utterance are the contextual assumptions and implications that a speaker aiming at optimal relevance must have expected the hearer to supply. Sometimes, as with ( 17) and ( 18) above, there is only a single possible hvpothesis about what these implicatures might be. Let us call ( 17) and ( 18) strong implicatures of (3b). Sometimes, however, there is a range of alternative hypotheses, all compatible with each other and roughly comparable in accessibility and contextual effects. Here, the utterance provides a certain amount of evidence for each alternative hypothesis, and the hearer is free to adopt whichever he chooses, and regard it as weakly confirmed by the speaker. The more alternative hypotheses there are, the weaker the implicatures will be, up to the point where the utterance provides no evidence at all for a certain assumption or conclusion, and all the responsibility for supplying it rests with the hearer. Although we hare no space here for an analysis of stylistic and figurative effects, the notion of weak implicature plays a decisive role in our account, as it does in our account of the vaguer aspects of both verbal and non-verbal communication.
A speaker aiming at optimal relevance should leave implicit everything that the hearer can be trusted to access for himself with less effort than would be required to process an explicit prompt. In (1) above, for example, the speaker does not explicitly specify what the food is too hot for, trusting the hearer to access this information for himself and use it to enrich the explicit content of the utterance to the point where it is consistent with the principle of relevance. This aspect of utterance interpretation has been largely ignored in the pragmatic literature, which tends to equate the domain of pragmatics with the domain of what we have been calling strong implicatures. We believe that a number of pragmatic problems have arisen directly from the fact that a variety of phenomena which have been standardly analysed as typical strong implicatures are not implicatures at all, but pragmatically determined aspects of explicit content.
Here is an example. Grice (1989) argued that the connectives 'end.' 'or' and 'if. . . then' were not ambiguous between truth-functional and non-truth-functional senses. According to him, an utterance of the form UP and Q' would be true if and only if both its conjuncts were true. The occasional temporal and causal connotations of conjoined utterances did not result from extra senses of the word 'and.' but were implicatures arising from the co-operative principle and maxims. A major problem for this account is that the alleged 'implicatures' fall within the scope of logical operators such as negation, disjunction, comparison and conditionals, contributing to the truth conditions of complex utterances in which they occur. Thus, (21a) and (21b) are not, respectively, tautologous and contradictory, as they should be if the temporal connotations of conjoined utterances are not part of truth-conditional content, but merely implicatures:
(21)a. It s always the same at parties: either you get drunk and no one will talk to you, or no one
        will talk to you and you get drunk.
b. If the old king died of a heart attack and a republic was declared Sam will be happy, but if a
  republic was declared and the old king died of a heart attack Sam will be unhappy.
[Adapted from Cohen 1971]
Grice himself seems to have regarded this objection as decisive. After discussing a parallel objection to his analysis of 'if,' he concludes:
    I'm afraid I do not see what defence (if any) can be put up against this objection.
                                                                                                                        [Lecture v. p. 10]
For Grice, as for most pragmatists, there seemed to be only two possibilities: either 'and' is ambiguous between truth-functional, temporal and causal senses, or land' has a single, truth-functional sense, and the temporal and causal connotations of conjoined utterances arise at the level of implicature. However, as Carston ( 1984, 1985) has pointed out, relevance theory offers a third possibility: it may be that 'and' has a single, truth-functional sense, but that the temporal and causal connotations of conjoined utterances are pragmatically determined aspects of explicit propositional content, and not implicatures at all.
Within relevance theory, the recovery of explicit content is not necessarily a simple matter of decoding, disambiguation and reference assignment. To obtain an interpretation consistent with the principle of relevance, the explicit content of the utterance may have to be enriched in certain Ways, As always, this enrichment is severely constrained: the first accessible enrichment consistent with the principle of relevance is the only enrichment consistent with the principle of relevance. By assuming that the temporal and causal connotations of conjoined utterances am pragmatically determined aspects of explicit content, we cm preserve Grice's arguments against the ambiguity of 'and', while still maintaining that the temporal and causal connotations of conjoined utterances contribute to truth conditions, and thus fall within the scope of logical operators.
We will end by showing that relevance theory eliminates the need for a maxim of truthfulness, the maxim that Grice himself regarded as the most important of the maxims, and without which Horn (1985:17) believes that 'the entire conversational and implicatural apparatus collapses.' We will argue that the assumption that there is a maxim, or norm, or principle, or convention of literal truthfulness creates unnecessary and insoluble problems for pragmatic theory.
A speaker observing the maxim of truthfulness should express only propositions she believes to be true, or at least propositions which she has some reason to believe are true. There are glaring counterexamples to the maxim of truthfulness, some familiar from the literature, others rarely considered: they include fictions, fantasies, idle speculations, guesses, rough approximations, metaphors, ironies and free indirect speech. Griceans have a set of labels for classifying these counterexamples: 'deliberate violation', 'opting out', 'clash of maxims', 'making as if to say', and so on. Thus, metaphor and irony are analysed as 'deliberate violations' of the maxim of truthfulness, from which the hearer is supposed to recover some logically related proposition as in implicature. However, nothing approaching an explanatory theory is proposed; and indeed, there are quite ordinary counterexamples to the maxim of truthfulness which the Gricean framework cannot deal with at all.
Consider the exchange in (72):
  (22) a. Peter: What does the election pamphlet say?
         b. Mary: We’ll all be rich and happy if we vote for them.
In a framework with a maxim of truthfulness, in saying (22b), Mary must be understood as expressing a belief that we will all be rich and happy if we vote for that particular party. However, it would often be quite natural to understand her not as expressing her own belief, but as representing or reporting the belief expressed in the manifesto. All we need to do to account for this interpretation is to drop the maxim of truthfulness and assume that an utterance may achieve optimal relevance by representing either the beliefs of the speaker or those of someone other than the speaker which it would be relevant to the hearer to know. The correct assumption about whose views are being represented is, as always, the first one tested and found consistent with the principle of relevance.
Irony fits quite straightforwardly into this framework. There is nothing to stop Mary, in representing the beliefs expressed in the manifesto, from communicating her own attitude to those beliefs. For example, she may indicate by facial expression or tone of voice, or simply trust Peter to realize from his knowledge of her, that she dissociates herself from the views she is representing, and indeed regards them as ridiculous or contemptible. The result is irony. Irony, on this account, does not involve 'a deliberate violation of any pragmatic principle or maxim: it is merely one means among others of communicating relevant information.
Another range of counterexamples to the maxim of truthfulness is the class of utterances intended and understood as rough approximations. Compare the alternative answers (23b) and (73c) to the question in (23a):
  (23) a. Peter: How far is Nottingham from London?
         b. Mary: 120 miles.
         c. Mary: 118 miles.
According to the maxim of truthfulness. Mary should not say that Nottingham is 120 miles, from London unless she believes that it is exactly 120 miles Tom London. If she believes that the true distance is in fact 1 18 miles, then she would violate the maxim of truthfulness by answering as in (23b). However, there are many situations in which a speaker aiming at optimal relevance should prefer the rough approximation (23b) to the strictly truthful (23c). Suppose Peter, who normally drives at about 60 miles an hour, is trying to decide when he should leave London for dinner in Nottingham. From both (23b) and (23c) he can recover the contextual implication that it will take him about two hours to drive to Nottingham, and that he should plan his journey accordingly. However, given that mental calculation is easier to do in round numbers, it will cost him less effort to recover these implications from (23b) than from (23c), and a speaker aiming at optimal relevance should prefer (23b) to (23c), as long as Peter has some way of knowing that the contextual effects she
intends her utterance to achieve arc one a subset of those he could in principle derive from her utterance (23b). We will argue that this is a possibility to which hearers are quite generally alert.
An utterance, we have seen, can be used to represent the thought of the speaker, or of someone other than the speaker. But who should we assume that the proposition expressed by an utterance must be identical to the thought it represents' Representation involves the exploitation of resemblances, and two objects resemble each other to the extent that they have properties in common. However, a representation can achieve its aim without sharing all its properties with the original it represents. For example, a bust of Napoleon may be made of white plaster, have no arms and legs, be found in a certain museum, and have been bought for a certain price. No rational addressee would attribute these properties to Napoleon himself. Nor should he. In this, as in every other aspect of interpretation, the minimal assumption - that is, the first accessible assumption-consistent with the principle of relevance is the only assumption consistent with the principle of relevance. If a communicator aiming at optimal relevance could have intended to convey an adequate idea of Napoleon without intending to suggest that Napoleon was made of white plaster, lacked arms and legs, etc., then he must be credited with this minimal intention: it is the only intention which a communicator aiming at optimal relevance could have hoped to achieve.
The proposition expressed by an utterance resembles the thought it is used to represent to the extent that they share logical properties, i.e. logical and contextual implications. Let us say that if they share all their logical and contextual implications, the utterance is a literal representation of the speaker s thought. Then it follows from the maxim of truthfulness that an utterance should always be a literal representation of the speaker's thought. By contrast, it follows from the principle of relevance that the hearer should take the utterance to be a literal representation of the speaker s thought only if this is the minimal assumption, i.e. the first accessible assumption, consistent with the principle of relevance. In the case of rough approximations such as (23b), since, as we have seen, there is an easily accessible less-than-literal interpretation of (23b) which is consistent with the principle of relevance, there is no need for the hearer to consider the literal interpretation at all.
Metaphorical utterances such as (4) fit straightforwardly into this pattern:
  (4) Their friendship blossomed.
By processing (4) in the context of his encyclopedic knowledge of blossoming, the hearer might derive a number of contextual implications. Some, for instance the implication that their friendship belonged to the plant kingdom, carry no plausible information and could hardly have been intended by the speaker to contribute to the relevance of her utterance. Other contextual implications, by contrast, do contribute to the relevance of the utterance and can therefore be assumed to have been, at least weakly implicated by the speaker, in the sense that the speaker intended the hearer to derive some such implications, if not exactly these. Thus, the hearer might conclude that the friendship in question grew from small beginnings, in a favorable environment, by a natural process, into something beautiful, that was perhaps destined to fade. , As with most metaphors, there is a substantial element of indeterminacy in the interpretation of (4), and its associated implicatures will be relatively weak. For a speaker who wanted to achieve a range of effects along these lines, (4) would be the most economical way of achieving them. Since (4) has an easily accessible non-literal interpretation which is consistent with the principle of relevance, there is no need for the hearer to consider the literal interpretation at all. By assuming that speakers aim not at literal truthfulness but at optimal
relevance, an explanatory insight into metaphor, irony and a variety of related phenomena can be achieved.
 
6. CONCLUDING REMARK
 
In this paper, we have tried to show that considerations of relevance play a decisive both in communication and cognition. Notice, though, that the principle of relevance, and the criterion of consistency with the principle of relevance, apply only to the recognition of communicators' intentions: other non-demonstrative inference processes are under no such constraint. The consequences are considerable.
As we have shown, every utterance has at most a single interpretation whisk is consistent with the principle of relevance. Moreover, it is in the speaker's interests to make sure that this interpretation, and the contextual assumptions needed to obtain it, are instantly accessible to the hearer: otherwise she is wasting her breath. It is an interesting question whether any comparable criterion applies to other non-demonstrative inference processes. Certainly, there is no criterion which is powerful enough to eliminate all but a single scientific hypothesis on the basis of immediately accessible contextual information. For this reason, pragmatic processes are more amenable than scientific theorizing to scientific study. For this reason too, the search for an adequate pragmatic theory is doubly worthwhile: not just for its own sake, but for the light it may shed on other non-modular processes, about which, as Fodor has rightly emphasized, so little is so far known.
 

REFERENCES

Carston, R. (1984). "Semantic and Pragmatic Analyses of 'And'." Paper delivered to the Linguistics  Association of Great Britain, April 1984.
Carston, R. (1985). "Saying and Implicating." Paper delivered to the Cumberland Lodge Conterence   on Logical Form, April 1985.
Cohen, L. J. (1971). Tithe Logical Particles of Natural Language." In Y. Bar-Hillel (ed.). Pragmatics   of Natural Language. Dordrecht: Reidel.
Fodon J. (1983). The Modularity of Mind Cambridge. Mass.: MIT Press.
Grice, H. P. (1989). Studies in the way of words. Cambridge. Mass.: Harvard University Press.
Horn, L. R. (1985). “Metalinguistics Negation and Pragmatic Ambiguity.” Language 61:121 - 174.
Sperber, D.,and D. Wilson (1986). Relevance: Communication and Cognition. Cambridge. Mass.:   Harvard University Press: Oxford: Basil Blackwell.
 
 
-- From Deirdre Wilson and Dan Sperber, "Pragmatics and Modularity". In The Chicago Linguistic Society Parassrssion on Pragmatics and Grammatical Theory, edited by Anne M. Farley, Peta T. Farley, and Karl-Erilc McCullough. Chicago Linguistic Society, 1986.
 
, from London unless she believes that it is exactly 120 miles Tom London. If she believes that the true distance is in fact 1 18 miles, then she would violate the maxim of truthfulness by answering as in (23b). However, there are many situations in which a speaker aiming at optimal relevance should prefer the rough approximation (23b) to the strictly truthful (23c). Suppose Peter, who normally drives at about 60 miles an hour, is trying to decide when he should leave London for dinner in Nottingham. From both (23b) and (23c) he can recover the contextual implication that it will take him about two hours to drive to Nottingham, and that he should plan his journey accordingly. However, given that mental calculation is easier to do in round numbers, it will cost him less effort to recover these implications from (23b) than from (23c), and a speaker aiming at optimal relevance should prefer (23b) to (23c), as long as Peter has some way of knowing that the contextual effects she
intends her utterance to achieve arc one a subset of those he could in principle derive from her utterance (23b). We will argue that this is a possibility to which hearers are quite generally alert.
An utterance, we have seen, can be used to represent the thought of the speaker, or of someone other than the speaker. But who should we assume that the proposition expressed by an utterance must be identical to the thought it represents' Representation involves the exploitation of resemblances, and two objects resemble each other to the extent that they have properties in common. However, a representation can achieve its aim without sharing all its properties with the original it represents. For example, a bust of Napoleon may be made of white plaster, have no arms and legs, be found in a certain museum, and have been bought for a certain price. No rational addressee would attribute these properties to Napoleon himself. Nor should he. In this, as in every other aspect of interpretation, the minimal assumption - that is, the first accessible assumption-consistent with the principle of relevance is the only assumption consistent with the principle of relevance. If a communicator aiming at optimal relevance could have intended to convey an adequate idea of Napoleon without intending to suggest that Napoleon was made of white plaster, lacked arms and legs, etc., then he must be credited with this minimal intention: it is the only intention which a communicator aiming at optimal relevance could have hoped to achieve.
The proposition expressed by an utterance resembles the thought it is used to represent to the extent that they share logical properties, i.e. logical and contextual implications. Let us say that if they share all their logical and contextual implications, the utterance is a literal representation of the speaker s thought. Then it follows from the maxim of truthfulness that an utterance should always be a literal representation of the speaker's thought. By contrast, it follows from the principle of relevance that the hearer should take the utterance to be a literal representation of the speaker s thought only if this is the minimal assumption, i.e. the first accessible assumption, consistent with the principle of relevance. In the case of rough approximations such as (23b), since, as we have seen, there is an easily accessible less-than-literal interpretation of (23b) which is consistent with the principle of relevance, there is no need for the hearer to consider the literal interpretation at all.
Metaphorical utterances such as (4) fit straightforwardly into this pattern:
  (4) Their friendship blossomed.
By processing (4) in the context of his encyclopedic knowledge of blossoming, the hearer might derive a number of contextual implications. Some, for instance the implication that their friendship belonged to the plant kingdom, carry no plausible information and could hardly have been intended by the speaker to contribute to the relevance of her utterance. Other contextual implications, by contrast, do contribute to the relevance of the utterance and can therefore be assumed to have been, at least weakly implicated by the speaker, in the sense that the speaker intended the hearer to derive some such implications, if not exactly these. Thus, the hearer might conclude that the friendship in question grew from small beginnings, in a favorable environment, by a natural process, into something beautiful, that was perhaps destined to fade. , As with most metaphors, there is a substantial element of indeterminacy in the interpretation of (4), and its associated implicatures will be relatively weak. For a speaker who wanted to achieve a range of effects along these lines, (4) would be the most economical way of achieving them. Since (4) has an easily accessible non-literal interpretation which is consistent with the principle of relevance, there is no need for the hearer to consider the literal interpretation at all. By assuming that speakers aim not at literal truthfulness but at optimal
relevance, an explanatory insight into metaphor, irony and a variety of related phenomena can be achieved.
 
6. CONCLUDING REMARK
 
In this paper, we have tried to show that considerations of relevance play a decisive both in communication and cognition. Notice, though, that the principle of relevance, and the criterion of consistency with the principle of relevance, apply only to the recognition of communicators' intentions: other non-demonstrative inference processes are under no such constraint. The consequences are considerable.
As we have shown, every utterance has at most a single interpretation whisk is consistent with the principle of relevance. Moreover, it is in the speaker's interests to make sure that this interpretation, and the contextual assumptions needed to obtain it, are instantly accessible to the hearer: otherwise she is wasting her breath. It is an interesting question whether any comparable criterion applies to other non-demonstrative inference processes. Certainly, there is no criterion which is powerful enough to eliminate all but a single scientific hypothesis on the basis of immediately accessible contextual information. For this reason, pragmatic processes are more amenable than scientific theorizing to scientific study. For this reason too, the search for an adequate pragmatic theory is doubly worthwhile: not just for its own sake, but for the light it may shed on other non-modular processes, about which, as Fodor has rightly emphasized, so little is so far known.
 

REFERENCES

Carston, R. (1984). "Semantic and Pragmatic Analyses of 'And'." Paper delivered to the Linguistics  Association of Great Britain, April 1984.
Carston, R. (1985). "Saying and Implicating." Paper delivered to the Cumberland Lodge Conterence   on Logical Form, April 1985.
Cohen, L. J. (1971). Tithe Logical Particles of Natural Language." In Y. Bar-Hillel (ed.). Pragmatics   of Natural Language. Dordrecht: Reidel.
Fodon J. (1983). The Modularity of Mind Cambridge. Mass.: MIT Press.
Grice, H. P. (1989). Studies in the way of words. Cambridge. Mass.: Harvard University Press.
Horn, L. R. (1985). “Metalinguistics Negation and Pragmatic Ambiguity.” Language 61:121 - 174.
Sperber, D.,and D. Wilson (1986). Relevance: Communication and Cognition. Cambridge. Mass.:   Harvard University Press: Oxford: Basil Blackwell.
 
 
-- From Deirdre Wilson and Dan Sperber, "Pragmatics and Modularity". In The Chicago Linguistic Society Parassrssion on Pragmatics and Grammatical Theory, edited by Anne M. Farley, Peta T. Farley, and Karl-Erilc McCullough. Chicago Linguistic Society, 1986.
 
 
 
 

语言学之家版权所有
《中华人民共和国信息产业部ICP/IP备案》 备案许可证号:苏ICP备09089076号