prag and rhetoric ocr

24
Pretending to Communicate Edited by Herman Parret w DE Walter de Gruyter * Berlin - New York 1994

Upload: na-na-chriesna

Post on 18-Dec-2015

24 views

Category:

Documents


0 download

DESCRIPTION

pragmatic

TRANSCRIPT

  • Pretending toCommunicateEdited byHerman Parret

    wDE

    Walter de Gruyter * Berlin - New York1994

  • Pragmatics and Rhetoric: A CollaborativeApproach to Conversation

    Y UEGUO Gu (Beijing)

    1. Preliminaries: Two Hot Issues in Current Linguistic Pragmatics

    1.1. The Status of the Gricean CP

    More than a decade has elapsed since the publication of Grices celebratedpaper (Grice 1975). What appears to be universally recognized as valuableis Grices proposal of conversational implicature. The notion of conuer-sational implicatwe" , observes Levinson (1983, 97), is one of the singlemost important ideas in pragmatic?. While the status of conversationalimplicature seems to be well established in pragmatics, its close associates- common purpose and the Cooperative Principle (and its maxims) proveto be quite controversial.

    Grice believes that conversational implicatures are specially related tothe particular purposes that talk exchanges are adapted to serve (19?5,47).He assumes that each participant recognizes in them [i. e. talk exchanges],to some extent, a common purpose.. . (ibid, 45). He states his maxims asif this common purpose were a maximally effective exchange of infor-mation (ibid, 47). Since participants in talk exchanges share a commonpurpose, they conduct talk exchanges cooperatively in order to achievetheir common purpose, hence the Cooperative Principle.

    Recent development in pragmatics with regard to Grices notions ofcommon purpose and the CP seems to fall into three groups, which canbe called, for the sake of easy reference, denialists, redstctionists and expan-sioniJts respectively. Denialists (e.g. Kasher 1976,1977; Pratt 1981; Sampson1982) hold that it is incorrect to assume that participants share somecommon goal, linguistic or otherwise, and that the Cooperative Principletherefore ought to be abandoned. Reductionists (e.g. Wilson/Sperber 1981,1986; SperberlWilson 1982,1986; Thomas 1986), on the other hand, adoptthe position that participants share a common linguistic goal, and that therepertoire of maxims should be reduced. Expansionists (e.g. Leech 1981,1983a, 198313; Gu 1987), in contrast, reinforce the CP with other conver-

    \ sational principles in order that various conversational goals can be ac-commodated.

  • 174 Yueguo Gu

    The first hot issue therefore is: Which of the three holds true whereasthe other two are false? Or whether each has a bit of truth and a bit offalsehood.

    1.2. The expulsion of perlocution from linguistic pragmatics

    In Austins conception of speech acts, the total speech act is incompletewithout the perlocutionary act. Nevertheless, Austins devotion to illo-cution and Searles follow-up work encourage the assumption that theillocutionary act can be performed independently of the perlocutionaryact. Van Dijk (1977, 198), Bach/Harnish (1979, 16) and Leech (1983a, 203)argue that the perlocutionary act does not belong to linguistic pragmatics,since it goes beyond linguistic communication. Similarly, Sperber/Wilson(1986) expel per ocutionary phenomenon from linguistic communication,1but from a different angle. It takes some space to spell out their position.In his analysis of meaning, Grice (1957, 385) observes [S] meant some-thing by x is (roughly) equivalent to [S] intended the utterance of x toproduce some effect in an audience by means of the recognition of thisintention. Strawson ( 1 9 6 4 ) fre ormulates the Gricean intention into threesub-intentions. To mean something by x, S must intend

    (a) Ss utterance of x to produce a certain response r in a certainaudience A;

    (b) A to recognize Ss intention (a};(c) As recognition of Ss intention (a) to function as at least part

    of As reason for As response r.

    According to Schiffer (1972:10), a certain response r in (a) is equivalentto a certain propositional or affective attitude or action r. Schifferconstrues propositional or affective attitude as referring, respectively, toe.g. believing that p, or feeling distressed or humiliated and action rincluding physical behaviour, such as jumping to a drowning mans rescueat his shout Help!

    Now the connection between Austins notion of the perlocutionary actand Grices S meaning something by x becomes only too obvious. InAustinian language, one would say: S performed a perlocutionary act ofy by saying x through producing a certain perlocutionary effect in anaudience. In Gricean language, on the other hand, one would say: S meantsomething by uttering x intending to produce some effect in an audienceby means of the audiences recognition of the intention (see also 5.1below). The direction oriented by Sperber and Wilson however is reduc-tive. They argue that, in Strawsons interpretation of the Gricean intention,(a) is not the communicative intention. The true communicative intentionis (b). Communication can succeed without either (a) or (c) being fulfilled.

  • Pragmatics and Rhetoric: A Collaborative Approach to Conversation 175

    They therefore dropped out (c) and reformulate (a) into what they callinformative intention, namely, to inform the audience of something(Sperber/ Wilson 1986, 28-29). Their view of successful communicationincludes Ss ostensive behaviour which makes Ss informative intentionmanifest and Hs inferential process through which H interprets Ss os-tensive behaviour and correctly recognizes Ss information.

    Now the second issue is: Is it justified to expel perlocutionary phenom-enon and the Gricean sub-intentions (a) and (c) from linguistic commu-nication? BJhat can be gained by doing so?

    1.3. The argument of this paper

    The two hot issues appear unrelated, but a closer look will soon revealthat they are inextricably interwoven. This apparent muddle is largely dueto the failure to recognize the fact that there are three different, althoughrelated, levels of interaction involved in conversation and that there aretwo kinds of cooperation in a talk exchange, namely pragmatic cooperationand rhetorical cooperation. The former can be assumed with great confi-dence whereas the latter cannot be taken for granted and is achieved ratherthan assumed. The three-level analysis will shed illuminating light on thesetwo confused issues.

    It is argued that apart from a few limited exceptions, linguistic com-munication becomes pointless without perlocutionary pursuit. Illocution-ary acts, while making manifest the interlocutors communicative intention,are performed also in fulfilling perlocutionary (or better to be called extra-linguistic) purposes. It is to be shown that linguistic pragmatics operatesat the lower levels of interaction whereas conversational rhetoric occupiesthe higher level of interaction. The former accommodates linguistic/communicative goals while the latter accounts for extra-linguistic goals.A fresh model is proposed which will not only enable us to see the GriceanCl? in a new perspective but will also be able to accommodate perlocu-tionary phenomenon and the Gricean sub-intentions (a) and (c).

    2. Review of the three positions on Grices CP

    2.1. Denialists view of common purpose and the CP

    Kasher, Pratt, Sampson and others argue that it is not normally the casethat conversationalists share some common purpose. To quote Pratt (1981,14) ... only some speech situations are characterized by shared objectivesamong participants. Clearly it is at least as common for speakers to havedivergent goals and interests in a situation.. . Since it is not normal thatconversationalists share some common purpose in talk exchanges, the CP

  • 176 Yueguo Gu

    validity becomes questionable. The principle is too strong since it restson an assumption of cooperation and shared purposes rather than on theweaker assumption of coordination and independent purposes (Kasher1977, 232). Sampson even concludes that Grices view of conversationalcooperation impl[ies] a profoundly false conception of the nature of sociallife (1982, 204).

    2.2. Reductionists view of common purpose and the CP

    Sperber/Wilson (1986) define their conception of common purpose asfollows: For us, the only purpose that a genuine communicator and awilling audience necessarily have in common is to achieve stlccessful cornmu-nication: that is, to have the communicators informative intention recogni?ed by theatldience. Grice assumes that communication must have a common purposeor set of purposes, or at least a mutually accepted direction (Grice 1975,45) over and above the aim of achieving successful communication [theemphasis is mine] (1986, 161- 62). Here they have made quite explicitthat their notion of common purpose is different from that of Grices.

    One consequence as a result of Sperber and Wilsons modification ofGrices common purpose is a drastic reduction of the Gricean maxims toa single principle of relevance. We shall take up the notion of implicatureitself, and go on to argue that the maxims are not all independentlynecessary for the generation of implicatures: that they may in fact bereduced to a single principle, . . . the principle of relevance.: IWilson/Sperber 1981, 155-56). Note that it is in order to do justice to utteranceinterpretation that the reduction of the maxims to a single principle ofrelevance is made. The enterprise of using a single principle of relevanceto account for utterance interpretation is pursued in Sperber/Wilson (1986)and Wilson/ Sperber (1986).

    2.3. Expansionists view of common purpose and the CP

    In contrast to Sperber and Wilsons reductionist approach is Leechsprogramme of expanding the repertoire of conversational principles. Un-like the previous two groups, Leech takes Grices conception of commonpurpose and the CP for granted. There seem to be three sources whichcontribute to Leechs expansion programme: (1) rhetorical interpretationof the CP; (2) inadequacy of the CP; and (3) multiple and divergent goalsin conversation.

    What is the nature of Grices CP and the maxims? Levinson sees it asessentially a theory about how people use language (1983,102) efficientlyand effectively in conversation. His view can be referred to as a rhetoricalinterpretation of Grice. An explicit spokesman of rhetorical interpretationis Marpurgo-Tagliabue who argues that all the Gricean maxims are classical

  • Pragmatics and Rhetoric: A Collaborative Approach to Conversation 177

    rhetorical rules in disguise: Grice . . . proposes nothing else than the rulesof classical rhetoric adapted to the modern position of speech . . . [iIgnoringthe rather comic whim of the author, of wanting to call his principlescategories and to put them into the Kantian classes of categories. ..(1981, 505). Leech (1981, 1983a) simply calls the Gricean CP a rhetoricalprinciple and has gone much further than Levinson and Marpurgo-Tag-liabue by enriching it with more rhetorical principles. Viewing the GriceanCl? and the maxims as representing good communicative behaviour(1983a, xi), Leech puts forward a theory which incorporates grammaticalrules with the enriched repertoire of the conversational principles (and themaxims) as main constraints on the correct, efficient and effective use oflanguage. This theory is termed general pragmatic?. As is traditionallythe case that the issue of correctness in the use of language is generallyregarded as a grammatical matter, whereas the issue of efficiency andeffectiveness is generally agreed to be a rhetorical concern, so it is quitenatural for Leech to rename the Gricean CP and the maxims as rhetoricalprinciples.

    Another reason for the enrichment of Grices Cl? is that it is found tobe inadequate without being rescued by other principles such as thePoliteness Principle (Leech 1983a, 80). The third reason is that, as thedenialists point out, in conversation apart from common goals there aremany other goals. Grice himself admittedly remarks that the goal ofmaximally effective exchange of information is too narrow, with regardto talk exchanges: . . . the scheme needs to be generalized to allow forsuch general purposes as influencing or directing the actions of others(1975, 47). But Grice fails to tell us how to generalize his scheme in orderto account for these other general purposes. Leechs expansion programmeseems to be one of the ways to make Grices scheme more adequate.

    2.4. Recapitulation

    Denialists seems to take Grices CP in terms of social cooperation (seealso 4.1 below), which differs from the rhetorical construe of it as anoverarching constraint on the use of language. In Sperber and Wilsonsreduction programme, the CP is assigned an interpretive role: It providesH with premises in interpreting conversational implicatures or Ss intentionor intended meanings in general. In Leechs expansion programme, onthe other hand, conversational principles have other roles to play: theyoffer functional explanations to such phenomena as indirectness, syntacticor semantic asymmetries, which cannot be accounted for in grammaticalterms.

    Critical assessment of the three positions is postponed until we havescrutinized Grices notion of talk exchange.

  • 178 Yueguo Gu

    3. ,4n anatomy of a talk exchange

    3.1. Three levels of interaction

    Grices much neglected notion of talk exchange deserves close exami-nation. It is normally the case that conversation is initiated by oneparticipant with another responding to it (who takes the initiative is byno means trivial). It is rare that two people start talking to each othersimultaneously. Even if this does happen, one side will normally backdown and give up the right of talking so that the other can finish. Onecrucial condition for the successful initiation of conversation is the poten-tial addressees willingness to talk (willing audience in Sperber andWilsons term). So before conversation starts at all, S has a pre-conver-sational goal: H (=potential addressee) is willing to listen to S (thenotion of goal will be dealt with in detail in 5.2 below). That S wants totalk to H, and that H is willing to listen to S represent the initial stage ofstarting a conversation.

    With the initial stage being successful, S and H become engaged in thefirst level of interaction, to be called S-,endi,,iH-interpreeting interaction.That is, the speaker sends information while the addressee interprets toget the information. Without this minimum cooperation there will be nosuch thing as a talk exchange. But the interaction of S-sending/H-inter-preting information is still not a talk exchange, which requires H toprovide information in response to Ss. So a talk exchange involves anotherlevel of interaction, viz information-exchanging interaction. That is, S givesH information and H provides information in response to Ss Oneexchange of information therefore consists of two pairs of sending/inter-preting information (see Figure 1).

    Figure 1 - S/initiator: sending information-l-

    OneExchange

    !

    L H/responder: interpreting to get the information- H/responder: providing information in return

    2-L S/initiator: interpreting to get the informationThe two levels of interaction together with the initial preparatory stage

    can occur independently in actual face-to-face situations. It is possible forthe willing-to-listen stage to be successful without the S-sending/H-inter-preting interaction. For instance, the addressee is willing to listen to whatis said but is unable to interpret. It is also possible to have the S-sending/H-interpreting interaction without the information-exchanging interaction.

  • Pragmatics and Rhetoric: A Collaborative Approach to Conversation 179

    The speakers talk is interpreted, but the addressee is unwilling to providefurther information in return; or the speakers talk is overheard or eaves-dropped, thus being interpreted by an intruder or an eavesdropper whohas no such goal as exchanging information.

    The third level of interaction is an uha-linguihc interaction. This can beillustrated with the following example (which actually took place in astudent kitchen).

    Paul: Do you have any milk?Roger: Yes (gives Paul some milk).

    Pauls utterance is used not just to elicit information from Roger, but alsoto get Roger to actually give him some milk. So what is involved in thetalk exchange is more than an exchange of information. This shows thaton top of the information-exchanging interaction there is yet another levelof interaction, viz an ultra-linguistic level of interaction. At this level, astill higher degree of cooperation is required: For Roger to provide trueinformation (a verbal response-act) to Pauls question is one thing; for

    Figure 2

    I1st level

    one talkexchange -

    - 2nd level

    merging withsocial action

    ( 3 r d l e v e l :

    Initiator Stage

    S-sending/H-interpreting interaction

    S-initiator: sending a message 4H-responder: interpreting the message

    Information-exchanging interaction

    H-responder: sending a return message-S-initiator: interpreting the message

    Ililtra-Linguistic interaction

    H-responder: producing a certain ultra-linguistic response

    (Note: It is possible to have the first and the third levels of interactionwithout the second.)

  • 180

    him to actually offer Paul some milk (a physical response-act at his owncost) is another.

    One may argue that the ultra-linguistic interaction ought not to beincluded as a level of analysis of a talk exchange. Our quick defence of itis this. It is a fact that talk exchanges are normally infiltrated by ultra-linguistic motives and considerations. Pauls utterance Do you have anymilk was not merely aimed at eliciting a verbal response. Rogers verbalresponse-act, on the other hand, was also constrained by his physicalresponse-act. If Roger did not want to give Paul any milk, although hehad some, this might lead him to give Paul a different verbal response:He might tell Paul that he had run out of milk as well.

    The three levels of interaction discussed so far is graphically representedin Figure 2 (Figure 2 see p. 179).

    3.2. Degrees and kinds of cooperation i

    It becomes apparent that the tripartite levels of interaction require differentdegrees of cooperation. The initial willing-to-listen stage requires thelowest but basic degree of cooperation between S and H (to be referredto as the 1st degree of cooperation). The S-sending/H-interpreting inter-

    i

    action requires a higher degree of cooperation (i. e. the 2nd degree). Theinformation-exchanging interaction, in comparison, demands an evenhigher degree of cooperation (i. e. the 3rd degree). The ultra-linguisticinteraction terminates at the highest degree of cooperation (i. e. the 4thdegree).

    I

    Conversational cooperation can be defined in terms of participantscontributions to goal attainment. If a participant makes more contributionto goal attainment, s/he is said to be more cooperative. For instance, if Hjust interprets Ss utterance without telling S the required information, Hmakes less contribution to goal attainment than s/he does if s/he not onlyinterprets Ss utterance, but also provides S with some appropriate infor-mation. In the latter case, H is more cooperative. Thus we can differentiate

    t

    1 There is a difference, however, between Ss cooperation with H and Hs cooperationwith S. S (= S/initiator) cooperation with H is secured by his rational pursuit of hisown goals (i. e. if one pursues a goal, one will normally by rationality do things conduciveto the attainment of the goal.) Hs (=H/ res on er cooperation with S, in contrast,p d )needs a motivational basis, for in principle H can refuse to interpret Ss utterance andignore Ss goals, if he wishes to. So Ss conversational cooperation can be assumed,whereas Hs conversational cooperation cannot. Whether H is to be cooperative with Sdepends on many factors, including situational contingencies. For instance, suppose thatS want to initiate conversation with H. It just happens that H wants to talk to S as well.This coincidence makes Ss initiation of a conversation a great deal easier and less risky.For another instance, a pedestrian who is asked to show a travellers way to the VictoriaStation cherishes an altruistic belief that he should help anyone who is in need ofassistance. This belief can be a decisive motivational force for the pedestrian to becooperative with the traveller.

  • Pragmatics and Rhetoric: A Collaborative Approach to Conversation 181

    two kinds of conversational cooperation: pragmatic cooperation and rhe-torical cooperation. Pragmatic cooperation deals with attainment of in-formative goals. It operates at the S-sending/H-interpreting interaction.Rhetorical cooperation, on the other hand, is concerned with the attain-ment of rhetorical and extra-linguistic goals. It operates at the levels ofthe information-exchanging and the ultra-linguistic interaction. Table 1summarizes the relations among levels of interaction, goals, Hs contri-bution and cooperation:

    Table 1

    Levels ofInteraction

    ContributionDegrees ofCooperation

    Kinds ofCooperation

    preparatorystage

    willing to listen toeach other

    1st pragmatic

    S-sending/H-interpretinginteraction

    S says something to Hwhile H interprets it 2nd cooperation

    information-exchanginginteraction

    S and H exchange in-~ formation with eachother by adopting

    3rd rhetorical

    each others goals I I

    ultra-linguistic S appeals to H to per-interaction / form a nonverbal act

    and H does it4th cooperation

    The differences between pragmatic cooperation and rhetorical cooper-ation lies (1) in the contributions made by the participants to goal-attainment; (2) in the sensitivity of goals with regard to e.g. cost andbenefit, interpersonal relations; and (3) in the degree of confidence assumedby the participants. Since pragmatic cooperation involves participantswillingness to listen to each other and their interpreting each othersmessage, it in general enjoys stronger confidence than rhetorical cooper-ation, and is usually goal insensitive and situation-insensitive. In themajority of social situations we can assume that participants are pragmat-ically cooperative. This is because the participants contributions cost littleeffort and energy, and can also keep the participants interests and goalsintact. Even in confrontational situations such as a detective interrogatinga suspect, pragmatic cooperation can be taken for granted: We can beassured that both the detective and the suspect will try their best tointerpret each others message.

  • 182 Yueguo Gu

    Rhetorical cooperation, on the other hand, requires goal adoption andis realized in Hs emotive, cognitive, verbal, and physical response-acts(see Gu, 1993, for detailed analysis). These categories of response-acts arehighly sensitive to goal relations, intrapersonal and interpersonal factors,and other situational contingencies. The goal sensitivity of rhetoricalcooperation can be illustrated in Table 2.

    Table 2

    I Levels of Interaction H is appealed to Iinformation-exchanginginteraction

    (a) to tell S the time(a) to tell S Hs secret

    ultra-linguisticinteraction

    (b) to give S a piece of paper(b) to give S $100

    Ss confidence in Hs rhetorical cooperation in (a) can be as strong ashis/her confidence in Hs pragmatic cooperation, whereas his/her confi-dence in Hs rhetorical cooperation in (a) will be very weak. Similarpoints can be made about (b) and (b). It must be emphasized that rhetoricalcooperation can only be assumed to be achievable. When it does occur, itis not due to the participants observance of the Gricean CP, but achievedand this is something to be explained. In general, linguistic pragmaticsrequires lower degrees of cooperation with strong confidence in partici-pants cooperation, while conversational rhetoric demands higher degreesof cooperation with weaker confidence in participants cooperation.

    4. The Status of the Gricean CP revisited

    4.1. Locating denialists, reductionists and expansionists

    At this point we are ready to locate the two hot issues in our analyticscheme of talk exchanges. Let us take up the first issue first. As pointedout in 1.1 above, the common purpose Grice had in mind was maximallyeffective exchange of information. Let us call it common goal (abbre-viated as G-corn), which can be reformulated as follows:

    G-corn: when S and H talk to each other, they exchange information ina maximally effective manner.

    Notice that G-corn is common in the sense that there are two goalspursued by S and H which are identical: S has a goal of giving H someinformation in an effective manner, while H has a goal of giving S some

  • Pragmatics and Rhetoric: h Collaborative Approach to Conversation 183

    information in return in the same manner. This is the essence of exchanginginformation.

    Grices common purpose and the Cl? are formulated at the level of theinformation-exchanging interaction. Let us recall Grices CP: Make yourconversational contribution such as is required... (1975, 45). Grices Cl?is normally applied to the speaker in the second S-sending/H-interpretinginteraction, as shown in Figure 3:

    Figure 3FA: Can you tell me the time? - to ___

    B: (interprets to get the information)(S-sending/H-interpreting Interaction I)

    8: Well, the milkman has come. response -A: (interprets to get the information)

    (S-sending/H-interpreting Interaction 2)

    Grices G-corn of effective exchange of information lies in the linkbetween the utterances by A and B qua speakers (as shown by the arrowabove). A in the first S-sendingjH-interpreting interaction sets up a con-versational demand (Dascals words, 1977) by asking B: Can you tell methe time? B, by the Gricean CP, makes his/her conversational contributionsuch as is required by A. That is, B not only interprets As message, butalso gives A true, informative and relevant information in a perspicuousmanner.

    In contrast, the reductionist programme operates at the level of S-sending/H-interpreting interaction. To continue with the previous exam-ple. The reductionists concern lies between As (qua speaker) utteranceand Bs (qua addressee) interpretation in the first S-sendingjH-interpretinginteraction; and between Bs (qua speaker) utterance and As (qua ad-dressee) interpretation in the second S-sending/H-interpreting interaction.One may point out that the difference noted above is merely a matter ofthe alteration of discourse roles -B changes his/her discourse role asaddressee in the first S-sending/H-interpreting interaction to the role ofspeaker in the second S-sending/H-interpreting interaction (with As rolealso being reversed). I would like to argue that there is much moreinvolved. Bs change of his/her discourse role from addressee to speakerinvolves goal-adoption and a change in the degree of cooperation. Themove from Bs (i. e. qua addressee) interpreting As utterance to Bs (i. e.qua speaker) providing A with the required information is by no meansa matter of course.

    Let us now turn to the denialists. None of them have clearly definedwhat they mean by common purpose. They seem to have understood

  • 184 Yueguo Gu

    Grices common purpose as referring to situations in which S shares withH some common aim or purpose be_rond that of getting H to understandwhich proposition is being expressed or implied (Thomas 1986, 27-28).This interpretation is characterized by Thomas as social goal-sharing.Evidence from denialist writings does support Thomas understanding.Denialist rejection of the Gricean common purpose seems to includesocial goals (in our terminology, ultra-linguistic goals) as well as GricesG-corn redefined above. So the denialists location in our tripartite schemeoccupies two levels of interaction: the information-exchanging and theultra-linguistic interactions.

    Expansionists, in comparison, are all-embracing. They are concernedwith all three levels of interaction. Expansionists position will be examinedin 4.2 below. Table 3 summarizes the various positions discussed so far.

    Table 3:

    information-exchanging + + tinteraction

    ultra-linguisticinteraction

    + ?

    (? = to be clarified later)

    4.2. Critique

    Now we are ready to evaluate the criticisms of Grice levied by denialists,reductionists and expansionists. The denialists are correct in pointing outthat conversationalists pursue divergent goals. But they have failed to seethat Grice is very careful not to confuse G-corn with other goals such asa goal of influencing or directing the actions of others (see Grice 1975,47). Grice states, in unambiguous terms, that his scheme needs to begeneralized to allow for such general purposes (ibid, 47). Their criticismof Grices CP seems to be based on the assumption that a common goalis a necessary condition for conversational cooperation. This assumptionis to some extent defective: Participants may be cooperative with one

  • Pragmatics and Rhetoric: A Collaborative Approach to Conversation 185

    another in order to attain collaborative goals rather than common goals.For instance, when S talks to H, S may have goal of getting H tounderstand the message; and when H is talked to, H may have a goal ofinterpreting Ss utterance in order to obtain the message. Ss and Hs goalshold a collaborative relation; that is, the attainment of Ss goal helps theattainment of Hs goal and vice versa. This collaborative goal relation alsoprovides a motivational basis for the cooperation between S and II. Sothe challenge of the generality of common goals in talk exchanges doesnot necessarily weaken the position of the Cl?

    Reductionists have contributed a great deal to the tendency in pragmaticsof shifting focus from the study of talk exchanges to the study of linguisticcommunication in general (in this connection see also Schiffer, 1972, 1989;Bach/Harnish, 1979). This focus-shifting is made through modifyingGrices G-corn and through filtering out extra-communicative goals fromthe domain of linguistic pragmatics. They are all concerned with interpre-tation, that is, with what we have called the S-sending/H-interpretinginteraction. Note that it is in order to do justice to utterance interpretationthat the reduction of the maxims to a single principle of relevance is made.The reductionist programme lead to at least two possibly misleadingconsequences. The first is on the ensuing line. Since the concern with theS-sending/H-interpreting interaction requires a lower degree of conver-sational cooperation than the concern with the information-exchanginginteraction, not surprisingly at all, Thomas, following Sperber and Wilson,points out (1986, 63-64) that Implicature-interpretation is possible onthe basis of a very weak notion of conversational cooperation. In factit is rendered so weak that the Gricean CP is in danger of being trivialized.Thomas (ibid, p. 53) admittedly observes: I think that if one rejects thesocial goal-sharing interpretation of Grices theory (as I do), then theconcept of conversational cooperation does become trivially true. To mymind, we can reject the social goal-sharing interpretation of Grices theory(as Thomas and I do) without trivializing the concept of conversationalcooperation. Thomass version of CP does trivialize the concept of con-versational cooperation, but not because of the rejection of the social goal-sharerss interpretation, but because of her formulation of the CP fromthe point of view of the speaker (Use your language in such a way thatyour interlocutor can understand what you are saying and implying (1986,47). If S has something to say to H (e.g. if A wants to ask B to tell himthe time), S will by hisjher rationality be cooperative with B. This is likesomeone who wants to go to London by coach being cooperative with

    Thomass very weak notion of conversational cooperation is likely to be w-hat we havereferred to as the 2nd degree of cooperation at the level of the S-sending, H-interpretinginteraction.

  • 186 Yueguo Gu

    the coach driver. However, if we follow Grice by formulating the CP withthe participants contribution to talk exchanges in mind, conversationalcooperation at each of the four degrees above distinguished is by no meanstrivial. The willingness of participants to listen to each other, interpreteach others message, exchanging information and adopt each others goalsare genuine cooperations.

    The second consequence is Thomass argument that ...it is perfectlypossible to interpret implicatures withoz/t having any faith in your interloctltorsVera+, [the emphasis is mine]. In other words, the maxim of quality,etc. is not necessary for implicature-interpretation. This shows that Thomashas gone even further than Wilson/Sperber (1981,155- 156), and Sperber/Wilson (1986), who have not abandoned the maxim of quality, but mergedit into the Principle of Relevance. To establish the relevance of anutterance, the hearer has to make valid inferences from a set of premiseswhich are true, or assumed to be true (Wilson/Sperber 1981, 171). If Ssostension makes manifest an assumption that is false and that H too knowsthat it is false, Ss ostension will yield no contextual effect, hence beingirrelevant (see Sperber and Wilson 1986, 128). If Ss ostension makesmanifest an assumption which is contradictory to a contextual assumption,no contextual effect will result, hence no relevance (Sperber/Wilson 1986,128). This actually forces Sperber and Wilson to make a stronger assump-tion than Grice does that the speaker speaks truthfully. In the cases ofuttering falsehood, Grice will say that the speaker breaches the maxim ofquality, but Sperber and Wilson will have to say that s/he is irrelevant,because s/he has made manifest a false assumption which fails to yield anycontextual effect. The merging of the maxim of quality into the principleof relevance therefore does not gain anything.

    Finally we come to the expansionists programme. Of the three, theexpansionists are the most sympathetic to Grices theory. The wholeprogramme aims at generalizing Grices scheme so as to accommodatemore goals. Leechs expansion programme, however, vacillates betweenthe 1st and the 2nd levels of interaction, and equally between the 2nd andthe 3rd levels of interaction. Leech wants to know how utterances havemeanings in situations (Leech, 1983a, x). In this aspect, Leech is closerto the implicature-interpretation enterprise pursued by reductionists. ButLeech also talks about dynamic goals and regulative goals (e.g. goals ofbeing polite). Among dynamic goals is the perlocutionary goal, which isincluded in the domain of general pragmatics (ibid, 201, 202). However,Leech seems to distinguish perlocutionary goals from perlocutionary ef-fects, and moves to expel perlocutionary effects from linguistic pragmatics.Furthermore, Leechs expansion scheme encounters some serious difficul-ties. Dillon et al (1985) correctly point out that Leechs rhetorical principles

  • Pragmatics and Rhetoric: A Collaborative Approach to Conversation 187

    are formulated in a rather ad hoc manner and that he fails to give areasonable criterion according to which a certain rhetorical principle canbe justified.

    4.3. Answer to the first hot issue

    Now it is time to state our answer to the first hot issue. The answer is:Denialists , reductionists and expansionists all have a bit of truth, and abit of falsehood. Denialists are correct in attacking Grices CP for beingtoo strong, but they are wrong in rejecting the CP all together. GricesCP is still valid at the level of the S-sending/H-interpreting interaction.Reductionists are right in narrowing down linguistic pragmatics to lin-guistic communication and calling for a weaker notion of Grices Cl?, butgo off the track in trivializing the CP. The expansionist programme a laLeech is praiseworthy for generalizing Grices original scheme to accom-modate more goals, but gets lost in the complexity of the matter.

    The stance this paper advocates is: Use the tripartite distinction of levelsof interaction to allocate a division of labour. Linguistic pragmatics takescare of the S-sending/H-interpreting interaction, leaving the information-exchanging and ultra-linguistic interaction to conversational rhetoric. Inthis new perspective, none of the arguments, valid or otherwise, maderespectively by the denialists, the reductionists, and the expansionists willbe an argument against the other two. For instance, on the face of it, thereductionist programme appears to be a direct denial of the expansionistprogramme. With the anatomy of talk exchanges of ours, they are closeassociates rather than life-or-death enemies.

    5. Linguistic communication vs purposeful interaction

    5.1 Answer to the second hot issue

    Now we are ready to come back to the question of whether it is justifiedto expel perlocution and the Gricean sub-intentions (a) and (c) fromlinguistic pragmatics. As pointed out in 1.2 above, van Dijk, Sach/Harnishand Leech expel perlocution out of linguistic pragmatics on the groundthat perlocutionary effects go beyond linguistic communication. It is onsimilar grounds that Sperber/lVilson erase the Gricean sub-intentions (a)and (c) from their framework of linguistic communication. I have noquarrel with the previous authors over their ousting perlocution fromlinguistic pragmatics. The point I would like to drive home is that theexpulsion is justifiable on methodological ground only, but indefensibleagainst empirical evidence. Perlocutionary phenomenon is by no meansperipheral; in real life situations the performance of illocutionary acts is

  • 188 Yueguo Gu

    infiltrated with perlocutionary considerations. In theory the two can beabstracted for the sake of linguistic analysis and for methodological con-venience; in practice they are not separate and inseparable. Austin isbasically correct when he observes: The total speech act in the totalspeech situation is the on4 actual phenomenon which, in the last resort,we are engaged in elucidating (Austin, 1980 [1962], 147). The total speechact is incomplete without the perlocutionary act. In his extension of Gricesaccount of S-meaning to the Austinian notion of illocutionary force,Strawson (1964) actually takes the position that one knows the particularillocutionary force of an utterance just in case one knows the particularperlocutionary effect S intends to achieve by means of the audiencesrecognition of Ss intention. 3 The three sub-intentions Strawson refor-mulates are steps towards working out illocutionary force (see also Schiffer1972, 92-94).

    Our position that in real life face-to-face situations illocutionary acts areperformed to fulfil perlocutionary purposes invites immediate challenge.Searle in his taxonomy of illocutionary acts treats the purpose or pointof the illocutionary act as the most important of his total 12 dimensions.However, he loses no time in drawing our attention to the fact that theterminology of point or purpose is not meant to imply, nor is it basedon the view, that every illocutionary act has a definitionally associatedperlocutionary intent. For many, perhaps most, of the most importantillocutionary acts, there is no essential perlocutionary intent associated bydefinition with the corresponding verb, e.g. statements and promises arenot by definition attempts to produce perlocutionary effects in hearers(Searle 1979, 3). Searle (1991) in his response to Alstons criticism (Alston,1991) further clarifies his notion of illocutionary point, observing that itis the point that the act has in virtue of its being an act of that type(Searle 1991, 101). Our notion of perlocutionary purpose is certainlydifferent from Searles illocutionary point or purpose. Perlocutionarypurpose is the purpose(s) of the speaker s/he attempts to reach in per-forming an illocutionary act. It is contingent and dependent on the totalspeech situation. When S makes a promise, Searle would say that theillocutionary point or purpose will be an undertaking of an obligation bythe speaker to do something. However, we would like to know whetherS by doing so meant to please H or not (see 5.3 below for furtherdiscussion). When Searle disassociates illocutionary point with perlocu-tionary purpose, he is thinking of illocutionary type and by definition,not of the actual performance of an illocutionary act, nor of the contingentpurpose(s) of the speaker.

    This scheme is not meant to account for highly institutionalized illocutionary acts.

  • Pragmatics and Rhetoric: ,4 Collaborative A4pproach to Conversation 189

    Where does the ousted perlocution belong? In the remainder of thispaper I would like to outline a general framework in which perlocutionaryphenomenon can be fruitfully investigated. This brings us back to theforesaid tripartite distinction of the three levels of interaction in talkexchanges. Illocutionary phenomenon is primarily located at the level ofS-sending/H-interpreting interaction. Perlocutionary phenomenon residesprimarily at the levels of the information-exchanging and ultra-linguisticinteraction. Illocution is part of pragmatic concern whereas perlocution ispart of rhetorical concern. Linguistic pragmatics and conversational rhet-oric work in collaboration to account for the complexity of talk exchanges.

    5.2 Transactional nature of perlocutionary phenomenon

    Elsewhere I have argued (see Gu, 1993) that the perlocutionary act hasbeen fundamentally misconceived since its very inception, and that thereis no such thing as perlocutionary act to be performed by S alone.Perlocution is in fact a transaction involving at least one speech actperformed by S and one response-act performed by H. The relationbetween the two acts are not causal, but rhetorical in the sense that Sinfluences H in adopting Ss goals. Successful perlocution therefore pre-supposes rhetorical cooperation between S and H. This understanding ofthe transactional nature of perlocutionary phenomenon demands a freshapproach to it, to be called rhetorical. The essential concepts of this newapproach include extra-linguistic goal, rhetorical goal and goal adop-tion, to which we now turn.

    Some ground preparation is needed here. So far we have been talkingabout intention, purpose and goal, relying on ordinary understanding ofthe terms. I have redefined Grices common purpose in terms of two goalswhich are identical. I assume, as Grice and many others do, that conver-sation is a purposeful activity, hence it is goal-directed. The relationbetween purpose and goal is neatly expressed by Taylor (1980 [1966],225): It is fairly natural to suppose that purposeful behavior is directedtoward some goal. Van Dijk (1977, 174) captures the tripartite relationsof intention, purpose and goal as follows: Whereas an intention has theaction itself as its scope, a purpose will be taken as a mental event inwhich an agent represents the GOAL(s) of the action. To put it bluntly,purpose goes beyond the action itself, hence beyond intention. To quoteTaylor again: Any descriptions of purposeful behavior must represent itas the means to some end or goal; no description or analysis of thatrelationship in terms of the concepts of physical science will work (ibid,p. 225). In this paper, goal is defined narrowly as a state of affairs whichis consciously or unconsciously desired and pursued in an individuals

  • 190 Yueguo Gu

    behaviour. 4 The individual is said to possess a goal when s/he has a mentalrepresentation of it. S/he achieves the goal when the desired state isreached or realized. S/he suffers from goal frustration if the desired statefails to be attained. The notion of goal presupposes an initial state atwhich the individual possesses the goal, and an end-state at which s/heattains or fails to attain the goal. There is always a time span betweenthe possession of a goal and the realization of a goal. A goal is to bedescribed in two ways. One way is to describe it in terms of possession.When we say that A has a goal A feels warm, that A feels warm doesnot refer to the end-state wherein A feels warm, but describes the mentalrepresentation of the goal that A feels warm. That A feels warm andthat A feels warm therefore capture two different concepts: the formerdepicts a goal, whereas the latter the end-state, viz the state of goalattainment. The other way to describe a goal is teleological: A does X inorder to G. For example,These two ways are both

    A switches on the heater in order to feel warm.adopted in this paper.5

    Figure 4

    (G-possession) (G-attainment)

    (initialstateat tl)

    i As speech act 1------ B interprets to knowi

    What is said, meantand aimed atWould you please

    switch on theheater?(at t3)

    (at t4)

    i

    4 This is different from that defined by Parisi/Castelfranchi (1981252) and by Leech(1983x40), but close to that by Argyle et al. (1981:68-69).

    5 For a general and philosophical treatment of teleological description of action and goalsee Woodfield (1976). The last decade or so has witnessed a promising revival of moreor less discredited teleology, e.g. Wright (1976)and Rescher (1986). Philosophers ofmind also begin to show strong interest in it, see Lycan (1990).

  • Pragmatics and Rhetoric: A Collaborative Approach to Conversation 191

    Now let us suppose that Individual A has a goal A feels warm andthat s/he wants to attain hisjher goal by performing a speech act, say, byuttering Would you please switch on the heater? to Individual B. Thuswe obtain the ensuing flowchart (Figure 4 see p. 190).

    At the initial state, A feels cold, which activates A to possess (withmental representation) the goal which is the desired state that A feelswarm. This goal motivates A to take a certain action to attain the goal,here a speech act. After processing As speech act, B realizes As goal byswitching on the heater. Notice that stages at t3 and t4 represent linguisticcommunication. The remaining stages represent the merging of the lin-guistic communication into ultra-linguistic purposeful interaction.

    5.3. Goal adoption and goal commitment

    It is important to notice that Bs action of switching on the heater has Asgoal A feels warm. We say that B has adopted As goal in switching onthe heater. So in performing the speech act, A not only has a goal A feelswarm, but has another goal, viz B adopts As goal. We call the first goalan extra-linguistic goal (G-extra for short), and the second rhetoricalgoal (G-rhet for short). G-extra thus defined is another way of sayingperlocutionary goal. However, a perlocutionary effect is not a G-extra.It is an end-state at which G-extra is realized. Suppose that Individual Asays to Individual B: The earth is flat. Assuming that in saying this Ahas a G-extra, B believes that the earth is flat. This G-extra is not aneffect. That B actually believes that the earth is flat is an effect. The actualoccurrence of an effect is already the attainment of G-extra. (Once a goalis achieved, it ceases to be a goal.) Hence a goal in our sense of the termcan never be an effect in the Austinian sense of the term.

    S has a G-extra (or more) in saying something to H in C (= the totalspeech context). But the attainment of G-extra requires a rhetorical trans-action, because it ultimately depends on the addressees cooperation byadopting a goal identical with G-extra. In essence the direct goal of Ssaying something to H in C is the G-rhet: H adopts G-extra by doingsuch and such. S saying something to H in C therefore is always anindirect means to G-extra. The direct means to G-extra is the addresseesresponse-act performed to realize G-extra.

    One important difference between the notion of the rhetorical transac-tion and that of perlocution is that the achieving of perlocutionary effectsbecomes the issue of goal-attainment. The attainment of G-rhet and G-extra is a joint endeavour of both speaker and addressee. This jointendeavour takes place through rhetorical force on the part of the speaker,and adoption of an identical goal on the part of the addressee. Theaddressees role in the rhetorical transaction is just as active and important

  • 192 Yueguo Gu

    as the speakers. The relation between illocution and perlocution becomesnot by performing x I did y, but performing x in order to reach goal(s),the description of which may involve y.

    One objection to our view of the relation between illocution andperlocution is the claim: I warn, I order, I request, etc., but I dont carewhether or not my addressee pays any heed to my warning, or carries outmy order, or fulfills my request. All in all, I dont have any perlocutionarypurpose to fulfil. At the first hearing, such a claim appears upholding,but collapse to scrutiny. In real life social situations, it would be pointlessor even irrational to warn you, order you, etc., without any purpose.It makes perfect sense to ask: Why do you (qua speaker) perform suchand such illocutionary act(s) in the first place? To look for an answer islooking for G-rhet and G-extra defined above. But Ss attitude towardhis/her own G-rhet and G-extra varies from goal to goal, and fromsituation to situation. We shall use the term goal commitment to referto Ss different attitudes toward G-rhets and G-extras. There are degreesof goal commitment. Some G-rhets and G-extras are strongly desired tobe attained, and S invests more effort in reaching them. Consequently s/he suffers from severe frustration if the goals are not attained (goal failure).G-rhets and G-extras like these are goals with strong goal commitment.Strong goal commitment exerts pressure on H for cooperative rather thanuncooperative responses. S may have very weak goal commitment (I-dont-care attitude) to his G-rhets and G-extras. For instance, a librariansays to a student: I just want to remind you that one of your books isoverdue. We can infer that the librarian has two extra-linguistic goals:

    (1) The student is aware that his book is overdue

    (2) The student returns the book

    The adjunct just indicates that the librarian has weak goal commitmentto the second extra-linguistic goal: She cares little about the realization ofthe goal, i. e. the student returns the book in fear of being interpreted asimposing. 6

    Some explanation of the notion of goal adoption is needed here, sincesome counter-examples against this notion are not difficult to find. TakeDont wake up! to awaken the sleeping H (Sadock 1974, 153; Bach/Harnish 1979, 82) and Theres a spider on your lap (see Davis, 1980).Suppose that in saying these S has the G-extras: H is awakened and His frightened. By our goal-attainment analysis, the achievement of theseG-extras will depend on Hs adopting them. One may wonder whether Hactually adopts them before being awakened or frightened. Instances like

    6 The first extra-linguistic goal, here G-extra, is somehow unique. Its realization requiresHs heuristic effort only. The adoption of the goal becomes automatic.

  • Pragmatics and Rhetoric: A Collaborative Approach to Conversation 193

    these can be multiplied, and seem to have some force against the notionof goal adoption, hence against the rhetorical transaction in general. Goaladoption is meant to account for the inter-relatedness of one individualsact to anothers act. One popular explanation is to evoke the notion ofconvention. Would you please switch on the heater? counts as an act ofrequest. 14n act of request requires by social convention a reply. So Hsswitching on the heater is an act in conformity with social convention(see Austin 1980 [1962], 117). Conversation analysts with ethnomethodol-ogical connections and advocates of speech act models of dialogue havegenerally adopted this explanation (see Levinson 1981). The explanationthat we attempt to advance is in terms of motivation. If we say that Bsswitching on the heater is an act performed to fulfil As goal A feelswarm without the notion of goal adoption, it will cause considerableconfusion to the theory of action; for it would mean to say that oneagents goal motivates another agents action. In general a correct andprofitable explanation of Bs switching on the heater is to say that s/hehas adopted a goal identical to As goal. Bs adoption of an identical goalis the link that relates Bs switching on the heater to As speech act.

    In general Hs adoption of a goal identical to Ss is a form of socialcooperation. It may be due to some altruistic motive or other motives,and it is quite normal for H to adopt goals the realization of which is athis/her cost. Sjhe does so for some other reasons or in order to fulfil othergoals. In general Hs adoption of a goal identical to Ss can be the killingof two or many birds with one stone. Hs switching on the heater realizesSs G-extra, and meanwhile it may also fulfil Hs own goal, viz H feelswarm, the goal H merely possesses without bothering to take any actionto realize it, or it may achieve some social goals such as S is pleased, andso on.

    Let us come back to the counter-examples above. In these instances,goal adoption is unnecessary. But speech acts like these make a very smallsubclass, which are even negligible. Hs response of being awakened orfrightened may very well take place before linguistic communication issuccessfully completed (see Gu, 1993, for detailed criticism of the twocounter-examples). Another subclass of illocutionary acts which appear topose problems are those highly ritualized conventional acts like wedding,christening, and sentencing at court. In these cases, goal adoption isachieved before the actual ceremonies or procedures.

    6. ,4 look at the future

    Limited space does not allow us to elaborate the notion of rhetoricaltransaction any further. To conclude this paper I would like to give acontour map of future research. The pragmatics/ rhetoric collaborative

  • 194 Yueguo Gu

    approach to conversation (or talk exchanges) focuses on the developingof speech acts, goal structure and goal attainment in the process ofconversationists rhetorical management of conversation. The developmentof speech acts include propositional development and illocutionary forcedevelopment (cf Widdowson 1978). Goal structure refers to goal relationsbetween Ss goals and Hs goals, and goal relations among Ss own goalsand among Hs own goals. Goal attainment is to assess how successfulthe conversation is being managed by the participants. (For preliminarydiscussion of these notions interested readers may consult Gu 1987, 1989.)

    References

    Alston, William P. 1991. Searle on illocutionary acts. In Lepore and Gulick, (eds.) pp. 57-80Argyle, M., Adrian Furnham, and Jean Ann Graham, 1981. Social Situ&on. Cambridge

    University Press.Austin, J. L., 1980. Horn to Do Things with Wordr. Oxford University Press. [1962]Bach, K. and R. M. Harnish, 1979. Linguistic Commutsicution and Speech Acts. The MIT Press.Cole, P., and Morgan, J. L. (eds), 1975. Sjztax and Semantics, vol. 3: Speech Acts. New York:

    Academic Press.Dascal, M. 1977. Conversational relevance. In ~bheJournaf of Pragmotics, ~01.1, pp. 309-328Davis, S., 1980. Perlocution. In: John Searle et al. (eds.) pp. 37-55.van Dijk, Teun A., 1977. Text and Context. London: Longman.van Dijk, Teun A., 1981. Towards an empirical pragmatics: some social psychological

    conditions of speech acts. Phifo~opbica 27:(1):127- 138.Dillon, G. L., Coleman, L., Fahnestock, J. and Agar, M., 1985. Review article. In Language,

    vol. 61, pp. 446-64Grice, H. P., 1975. Logic and conversation. In Cole and Morgan (eds.) pp. 41-58Grice, H. P., 1989. Studies in the wuy of Words. Cambridge: Harvard University PressGu, Y. G., 1987. Towards a model of conversational rhetoric: an investigation of the

    perlocutionary phenomenon in conversation. Ph D thesis (typ.) Lancaster UniversityGu, Y. G. 1989. An outline of conversational rhetoric. In the Journal of Foreign Lungwag

    Teaching and Research, ~01.2, pp. 8-15 (written in Chinese)Gu, Y. G. 1993. Impasse of perlocution. In the JournaE of Pragmatics.Kasher, A., 1976. Language in Focus: Foundations, Methods and Systems. D. Reidel Publishing

    CompanyKasher, A., 1977. What is a theory of use? In the Journal of Pragmatics, vol. 1, pp. 69-84Kiefer, F., 1979. What do conversational maxims explain.2 In Linguisticoe Inuestigationes, vol.

    3, pp. 57-74Leech, Geoffrey N., 1981. Pragmatics and conversational rhetoric. In: H. Parret, M. Sbisa

    and J. Verschueren, (eds.) pp. 413-439Leech, Geoffrey N., 1983a. Principles of Pragmatics. London: Longman.

    Leech, Geoffrey N., 198313. Pragmatics, discourse analysis, stylistics and the CelebratedLetter. In Herman, V. and Philip Dodd, (eds.), Prose Studies, vol. 6, pp. 142-157

    Lepore, Ernest and Robert Van Gulick, (eds.) 1991. John Se&e and His Critics. London: BasilBlackwell.

    Levinson, S. C., 1983. Pragmatic_r. Cambridge: Cambridge University Press

  • Pragmatics and Rhetoric: A Collaborative i\pproach to Conversation 195

    Levinson, S. C., 1981. The essential inadequacies of speech act models of dialogue. InParret, H. et al. (eds.), pp. 473-492

    Lycan, William G. (ed.) 1990. i2find and Cognition: a Reader. London: Basil BlackwellMyers, T., K. Brown and B. McGonigle, (eds.) 1986. Reasoning and Dirsoum Processes. London:

    Academic PressMorpurgo-Tagliabue, Guido, 1981. Grammar, logic and rhetoric in a pragmatic perspective.

    In Parret, H. et al. (eds.), pp. 493-508Parisi, Domenico and Cristiano Castelfranchi, 1981. A goal analysis of some pragmatic

    aspects of language. In Parret, H. et al. (eds.), pp. 551-567Parret, H., M. Sbisa and J. Verschueren, (eds.), 1981. Possibilities and Limitations ofPragmatirs.

    Amsterdam: John Benjamin.Pratt, MI. L., 1981, The ideology of speech-act theory. In Centrzim, l:l, pp. 5-18Rescher, Nicholas, (ea.) 1986. Currmt Zssues in TeLofogy. University Press of America, Inc.Sampson, G., 1982. The economics of conversation. In Smith, (ed.) pp. 200-210Schank, R. and R. Abelson, 1977. Scripts, Plans, Goals, and Understanding. Hillsdale: Lawrence

    Erlbaum AssociatesSchiffer, S.R., 1972. Neaning. Oxford: Clarendon PressSchiffer, S. R. 1989. Remnants of Meaning. The MIT PressSea&, John R., 1979. Expression and Meaning: Studies in the Tbeoq of Speech Arts. Cambridge:

    Cambridge University PressSea&, John R., 1969. Speech Acts. Cambridge University PressSearle, John R. and E Kiefer and Manfred Bierwisch, (eds.) 1980. @eecb Act Theory and

    Pragmatics. D Reidel Publishing CompanySea&, John R. 1991. Response: meaning, intentionality, and speech acts. In Lepore and

    Van Gulick, (eds.) pp. 81- 102Smith, N., (ed.) 1982. iMu&ul &row/edge. London: Academic PressSperber, D. and Wilson, D., 1982. Mutual knowledge and relevance in theories of compre-

    henslon. In Smith, (cd.) pp. 61-87Sperber, D. and Wilson, D. 1986. Releoance. London: Basil BlackwellStrawson, Peter F., 1964. Intention and convention in speech acts. In Philosophical Review,

    vol. 73, pp. 439-460Taylor, R., 1980. z4ction and Pmpose. N.J.: Humanities Press [1966]Thomas, Jenny, 1986. The dynamics of discourse: a pragmatic analysis of confrontational

    interaction. Ph D thesis (tvp.) Lancaster UniversityWerth, P., (ed.) 1981. Conuers&n and Dimurse. Groom Helm LtdWiddowson, H., 1978. Teaching Lang.wge as Communication. Oxford University PressWilson, D. and D. Sperber, 1981. On Grices theory of conversation. In Werth, (ed.)

    pp. 155-78Wilson, D. and D. Sperber, 1986. Inference and implicature in utterance interpretation. In

    Myers er al (eds.), pp. 241-64Woodfield, A., 1976. Teleology. Cambridge University PressWright, Larry, 1976. Teleological Explanations: An Etiological Am&is OJ Goals and Functions.

    Berkeley: University of California Press