DOCUMENT RESUME ED 056 500 FL 002 653 hUTHOP Fraser, Bruce 7ITLE Sentences and IllocutionaryPorces. INSTITUTION Language ResearchFoundation, Cambridge, Mass. PUB DATE Jul 71 NOTE 67p.; In Language ResearchReport No. 4 EDPS PRICE MF-$0.65 HC-$3.29 DESCPIPTORS Deep Structure;Descriptive Linguistics;English; Generative Grammar; Grammar;*Linguistic Competence; Linguistic Theory; *Semantics;*Sentences; Sentence Structure; StructuralAnalysis; Surface Structure; Syntax; TransformationGenerative Grarmar; *Transformation Theory (Language);Verbs ABSTRACT This paper considers the wayin which a grammar must accuunt for the speaker'sknowledge of sentence force asopposed to sentence form or meaningand the way in which thisforce is related to a sentence. Accordingto the performativPanalysis approach, the force of each sentenceshould be stated explicitly is apart of the underlying representation ofthat sentence. Afterconsideration and rejection of performtiveanalysis, the author suggests anapproach which states that sentenceforce is a function ofsentence meaning, analogous to the work insemantics showing that sentencemeaning is a function of constituentmeaning and sentence form.Three types of relationships between sentencemeaning and sentence form(explicit, underdetermined, and idiomatic) aredescribed along with examplesof each. A list of referencesig included. (VM) Sentences and Illocutionary Forces/1 u 9 otpAs."6.4 NT OTT NI MTN IDUCATION rigyos:,0% ,) ni rut). I .. *MAPS g% RA%, TM 11, IOUCATION Bruce Fraser RI T6 .9 00.ATTPITI i HITS TIE P100v6-10 I II.AC As NW . .1, #409,6 THE PfitscoNON 0.6.**WAT014 , TING IT POINTS OT VIII" ON (.6TNNTOht AMID 00 POW NICIS SRSHI WC( kf OTTKI 06 EOu TO AYE/ ORt.Avi:AnOys Or( RATO.,e1, CATION POSITION TIN P.:, Icy I.:NO(0 AGRI( 6.5(hits ysITH Trot S OUICI 01 IDLCAT.0.4lutITH(11 RERRODUOION OwsiDt T4( 14C syst; RA RIOLNRIS 1.105.5011 ?...(COP....C..1! CAM 1. Introduction The native speaker of English knows that a sentence like (1-1) (1-1) John may leave now. standardly counts as a simple prediction of the future, a report of John's freedom of movement, or the giving of permission -- to nam: three of the most likely ways in which the r,entence can be used.1 He also knows that this sentence is not standardly taken to have the force of a plea, a request for information, or an oath, although the circumstances in which the sentence is uttered might permit these latter interpretations. I take this knowledge to be a part of the linguistic competence of the native speaker and, thus properly included within the domain of a grammar. In this paper I will be concerned first with how the pairing of a sentence with its illocutionary force(s) fits into a grammarEnid second with the principles which relate a sentence and its force(s).2 In the following discussion we will be talking about sen- tences, not utterances. A sentence, a construct with linguistic To appear in S. Anderson and P. Kiparsky (Eds.), Festschrift for Morris halle 101. 1 -2- theory, has standard ways in whichit can be used, e.g., to make a promise, make a plea,etc. We will not be talking about utte,.- ances: using sentencesto communicate. Thus, we are not con- cerned with how a sentence HASbeen usea, was INTENDED tc beused, was TAKEN by thehearer, and so forth. In addition, although the term conventional use(standard use) is a technical term, I have no adequate definitionfor it and will rely on thereader's intuitions. To pick an extreme case, sentence(1-2) (1-2) Harry is ill. is standardly used to make astatement, give a report, or make a warning, butis certainly not conventionally used as anexample of a three-word sentence. On the other hand, the sentence (1-3) I promise I'll find you. has the superficial appearanceof a promise; one might want to argue that it isconventionally used as a threat aswell. Per- haps so. What is important is not if this orother instances appear to violatelinguis4ic conventions, such as a verbused performatively denoting the illocutionaryact the sentence can count as, but rather, what we taketo be the conventionaluse(s) of the sentence and how this is tobe accounted for. I will be addressing two mainquestions in the course of the paper. The first is.the way in which thepairing between a sentence and its force(s) fits into a grammar. In section 2, we examine and reject the PerformativeAnalysis (PA), a position which holds that the conventional useof a sentence is stated explicitly as part of the underlyingrepresentation of the sen- tence, where the pairing is accounted forby the generative rules 2 - 3- illocutionary force indi- of the grammar. In this analysis the cator is always Ahighest performative verb. In section 3 we propose analternate approach inwhich conventional force istaken to be a functionof sentence meaning In discussing and linguisticconditions on illocutionaryacts. betweel meaning and this approach weshow that the relationship force is often notexplicit and requires avariety of interpre- tive principles. In section 4 we summarizethe discussion. 2. The PerformativeAnalysis 2.1 Statement of thePerformative Analysis The most widely knownand detrAled attempt topair a sen- tence with its forcc isthe effort which Iwill refer to as the Performative Analysis(PA). Based primarily on someideas pre- sented in Austin(1962) and elaborated ingreatest detail and most carefully byRoss (1970), the majorthrust of the PA is that the force of eachsentence should be statedexplicitly as The a part of theunderlying representation ofthat sentence. PA asserts that sentenceforce should be carriedby a single per- formative verb present inthe highest clause ofthe sentence, and that this highestclause can, under certainsyntactically- statable conditions, bedeleted. It is argued, forexample, that (2-1) has(2-2) underlying it and that aPerformative Deletion Rule deletesall but the embeddedS. 3 (2-1) Prices slumped. (2-2) +Performative s[ I- +Communication - you -[sprices slump_Ed]s]s +Linguistic +Declarative Aside from the problems of thePA statement and its justifi- cation, points to be taken upin §2.2, there are a couple of important implications whichfollow from the adoption of thePA. First, sentences like (2-3) i) I promise you that I will behome at 5 o'clock. ii) I will be home at 5 o'clock. both have the force of apromise: the first only that force, standardly, and the second,that force among others. I maintain that the sentences differin meaning: (2-31) means that the speaker undertakes anobligation to be home, at 5 o'clock;(2-31i), on the otherhand, lacks the sense ofobligation as a part of its meaning. In using the sentence(2-3ii) to make a promise (a pos- sible use), the speaker assumesthe obligation; but thisobliga- tion is not part of the'meaning' of the sentence(2-311). The PA precludes this position. Since the PA claims that(roughly) (2-31) underlies (2-3ii), andaccepts that transformations are meaning preserving, derivingthe abbreviated sentencefrom the fuller form requires meaningpreservation. To support the PA, one must maintainthat the examples in (2-3) andsimilar cases have identical meaning. Moreover, one is forced to maintainthat transformations deletemeaning-carrying lexical items, forthere is certainly nothingin I or will, etc., that carriesthe sense of speaker obligationto carry out the actionspecified by the -5- following verb phrase. Second, since the underlyingrepresentation is defined as the level at which theforce of the sentence isdetermined, it follows that force ambiguityof a particular sentence isdisam- biguated at the underlyinglevel. For example, in terms ofthe PA, (2-311) has threeunderlying representaticns: (2-4) 1) I predict ii) I warn you that I will be there on time. iii) I promise you 2.2 Critique of the PerformativeAnalysis In this section I want toexamine the PA as presented by Ross (1970) in some detail,to see just how well itstands up when pushed. At issue is both the accuracyof the overall state- ment of the PA and thesyntactic evidence which hasbeen adduced to support it. 2.2.1 Problems of the General Approach The PA maintains that (2-5) every deepstructure contains one and only one per- formative sentence as itshighest clause (Ross,Ibid., p. 261) Thepe are two differentclaims: first, that everysentence has only a single performativeverb associated with itwhich (verb) specifies the force of thesentence. And second, that this verb is in the topmost clauseof the sentence's underlyingrepresenta- tion. I will examine the secondof these claims first. Note that the sentences (2-6) i) I regret that I must inform you of your dismissal. ii) I am pleased to be able to offer you the job. iii) Let me point out that I admit you're right. iv) I would like to congratulate you. can be taken as a statement of a regret, anoffer, an admission and a congratulation, respectively, even though theperformative verb is embedded and is not the highest verb as requiredby Ross (2-7) "All declarative sentences occurring in contexts where first person pronouns can appear derive from deep structures containing one and only one super- ordinate performative clause whose main verb is a verb of saying." (p. 259) Regret and point out may be verbs of saying; pleaseand like are not. In fact, examples such as those in (2-6) run counter to the claim that "there is an independently necessaryconstraint that prohibits any verb from having a performativeinterpretation when it is embedded as the complement of anotherverb." (p. 251) I doubt that evidence can be found to ergue thatin (2-6) the sentence-initial material is not at the highestlevel. Sentences with -1.z. adverbs form one class ofcounterexamples to the highest verb claim. (2-8) i) Obviously I concede that I've lost the election, ii) It is obvious that I concede that I've lostthe election.