meaning, commitment, and prosody - esslli...
TRANSCRIPT
MEANING,COMMITMENT,ANDPROSODY
ClaireBEYSSADESFL-Paris8University,France
ElisabethDELAIS-ROUSSARIELLING,NantesUniversity,France
2
General aimTraditionalview:meaningandtruthconditions.
ØOKfordeclarativesentencesandassertionsØbuthowtoaccountforothersentencetypesandothertypesofspeechacts:questions,commands,exclamations…
3
General aimNewpropositionsinterms ofcommittments:• Gunlogson (2003,2008),Farkas &Bruce(2010),Malamud&Stephenson(2011)toaccountfordeclarativequestions,confirmationrequests,polarparticles...
• Portner&Rubinstein(2012),Portner(2015)toaccountforvariousimperatives.
• Krifka (2015, 2016) to account for informationconveyed by an utterance, but also for thecontinuationsof the conversation.
4
General aimMeaning ofintonation, inparticular,when allotheraspectsofutterance don’t vary.
Ø Falling declaratives which areassertingØ Risingdeclaratives which arequestionning
(1) Itisraining ↑Ilpleut↑
(2) Itisraining ↓Ilpleut↓
Autosegmental-metrical approach (Pierrehumbert 1980,Beckman &Pierrehumbert 1986,Selkirk, 1984,Ladd2008).
5
General aimDynamicpragmatics(à laPortner 2015)• Sentenceshavestandardstaticsemanticvalues.• Thecommunicativeeffectofutterancesindiscourseismodeledastheeffecttheyhaveondiscoursecontext.
• Theeffectofaparticularsentenceisdeterminedbypragmaticprinciplesonthebasisofthesentenceformorsemantics.
Speechacttheory(à laLevinson,orGazdar)GrammarforConversation(à laGinzburg)
6
General aimMainissues:
• Thesemanticvalueofnondeclarativesentences
• Therepresentationofthediscoursecontext
• Thepossibleassymmetry betweeninterlocutors
• Theroleofintonationininterpretation
7
Outline of lecturesLecture1.Commitment:fromHamblinto
Krifka.Lecture2. Prosodyandmeaning.Focusonquestionsandcommitments
Lecture3.Questions,commitmentsandbias.Lecture4.Alternativequestions.Lecture5.Rhetoricalquestions.
COMMITMENT:FROMHAMBLINTOKRIFKA
ClaireBEYSSADESFL-Paris8University,France
ElisabethDELAIS-ROUSSARIELLING,NantesUniversity,France
9
I. Historical overview1.1 Hamblin (1971)Commitmentisnotbelief
The speaker may commit herself to something,which she doesn’t believe.To commit oneself = to act as if one believes apropositionp.
Belief:aprivate stateofmindCommitment:apublicstance
10
I. Historical overview1.1 HamblinBy asserting, the speaker commits herself toa proposition, that she may abandon in caseincompatible new information comes in.
Ø Assertions as the speaker’scommitments to a proposition.
11
I. Historical overview1.1 Hamblin: a grammar for dialogue• Five types of locutions: assertions, retractions,inquiries, retraction demands and I don’t know.
• A dialogue is a sequenceof locutions.• A commitment slate (i.e a set of assertions) isassociatedwith each participant.
• A context is a locution and an associatedassignment of commitments slates toindividuals.
12
I. Historical overview1.1 Hamblin• Rules define the set of well-formeddialogues.1. Following an assertion, everyone’s
commitment slate includes that assertion.2. Following a retraction by p, p’s commitment
slate doesn’t include (anymore) what’sretracted, but every other participant’scommitment slate remains unaltered.
13
I. Historical overview1.2 Gazdar (1981)Distinction between sentence types, semantictypes, illocutionary force and speechacts.• The meaning of a declarative sentence is aproposition.
• The meaning of an interrogative sentence is aset of propositions.
14
I. Historical overview1.2 Gazdar (1981)• An illocutionary force is a function from contents to
update potentials.• Update potential is a function from contexts to
contexts.• A speech act assignment is a pair <f,c> consisting of
a force f and a content c.• A speech act is f(c), for any speech act assignment
<f,c>.
15
I. Historical overview1.2 Gazdar (1981)Issue: the speech act assignment problemThesis: the polyfunctionality of sentences.Asamesentencebutvariousspeechacts.NoconstraintupontheuptakebyAddressee
(1) A: Youwillgohometomorrow.B: a.Howdoyouknow? (assertion)
b.Yes. (question)c.Okay. (command)
16
I. Historical overview1.2 Gazdar (1981)Gazdar used the notion of commitment todefine the speech act import of utterances.He extends commitments to non-propositional contents.“ An assertion that Φ is a function thatchanges a context in which the speaker isnot committed to justifiable true belief inΦ into a context he is so committed.
17
I. Historical overview1.2 Gazdar (1981)
“A promise that Φ is a function thatchanges a context in which the speaker isnot committed to bringing Φ about intoone in which he is so committed.A permission to Φ is a function thatchanges a context in which Φ isprohibited into one in which Φ ispermissible.”
(Gazdar,1981:69)
18
I. Historical overview1.3 Beyssade&Marandin (2006)OurproposalisbasedonGazdar’sextension,thatwemakeexplicitbyusingGinzburg andSag’sontology(2000):1. Proposition (p)2. Question (?p)3. Outcome (!p)4. Fact
19
I. Historical overview1.3 Beyssade& Marandin (2006)• Commitment to a proposition: being ready tostand for the truth of that proposition.
• Commitment to a question: being interestedin the issue defined by the question.
• Commitment to an outcome: being positivelyoriented towards the actualization of apotential stateof affairs (Stefanovitch2003)
20
II. Speech acts in dialogueConsider the effect of an utterance in dialog in terms ofupdatesØ The speaker commits to a content.Ø The addressee may accept this content, but she may also
refuse it.Ø And the speaker may anticipate the addressee’s
reaction.(2) Tu ne vas pas me croire, but Marie ment.
You won’t believe me but Mary is lying.
21
II. Speech acts in dialogueDistinguish two times in dialogue and in the contextupdate:- the speech act before its acceptation- the speech act after the addressee’s reaction.
Account for the fact that the speaker always anticipatesthe addressee’s reaction:- by default, the speaker anticipates an acceptation by
interlocutors (non defective context à la Stalnaker)- but there are utterances which explicitly convey the idea
that interlocutors disagree.
22
II. Speech acts in dialogue2.1 StalnakerHe accounts for accepted assertions only.
« Once the context is adjusted to accommodate theinformation that the particular utterance was produced,how does the content of an assertion alter the context ?[…] The essential effect of an assertion is to change thepresuppositions of the participants in the conversationby adding the content of what is asserted to what ispresupposed. This effect is avoided only if the assertionis rejected » (Stalnaker, 1978: 86)
23
II. Speech acts in dialogue2.2 GinzburgHe wants to account for the assertionbefore its acceptation.He associates an assertive speech actASSERT(p) with to updates:ü add p to Factsü add ?p to QUD
24
II. Speech acts in dialogue2.2 GinzburgHighlights the asymmetry between Speaker and Addressee:"Both the view of context incrementation deriving fromStalnaker and the discourse-structure tree-based view facecertain problems. The crux of the matter is that when anew assertoric contribution is encountered, it cannot, as itis the case in the various standard approaches to discoursesemantics, be attached simpliciter or added to FACTS. [...]A cannot update FACTS before receiving acceptance fromB" (Ginzburg, 1997: 10).
25
II. Speech acts in dialogue2.2 GinzburgAny assertion can be accepted, rejected, discussed…(3) A: Jean est venu hier.
Jean came yesterday.B: a. C’est noté. / I get it.
acceptationb. C’est faux. Il est en congé./ It is false. He ‘s in
vacation.reject
c. Qu’est-ce qui te fais croire ça ? / How do you knowthat?discussion
26
II. Speech acts in dialogue2.2 GinzburgBy default, an assertion is accepted, and backchannels are away, among others, to indicate this acceptation. They can beverbal or non verbal (head movements, brief vocalizations,glances, and facial expressions, often in combination).
(4) A: Jean est venu hier.B: Mhmh. /Ouais. (backchannels)
(4’) A: Jean came yesterday.B: Uh huh. (backchannels)
27
II. Speech acts in dialogue2.3 Generalisation to other speech actsDoubleimpactofspeechacts oncontextØ Thespeakertakes apublicstance:she
showssomething from her privatementalstate(Belief,Desire,Intention)
Ø The speaker expects a reaction from theaddressee. Every expressed attitude,except exclamatives, is volitional(Zaeferrer, 2001)
28
II. Speech acts in dialogue2.3 Generalisation to other speech actsAssertion« Constativesexpressthespeaker's belief andhis intentionthat thehearer haveorform alike belief »(BachandHarnich,1979:41).(5) It’s raining.Ø Ibelieve it’s raining.Ø Iwant that you share this belief /Iask
you tohavealike belief.
29
II. Speech acts in dialogue2.3 Generalisation to other speech actsQuestion (6) Is it raining?
Ø I’m wonderingwhether it’s raining.Ø Iwant you toanswer this question/I
ask you tohavealike question.Order (7)Take anumbrella!
Ø Isuggest you totake anumbrella.Ø Iwant you accept this suggestionand
perform thecorrespondingaction.
30
II. Speech acts in dialogue2.3 Generalisation to other speech actsConfirmation request(8) Jean est venu hier, (hein / n’est-ce pas) ?
Jean came yesterday, didn’t he?
Particles or tags may be used to make explicitthis inquisite part of each utterance.Not just a question, but an utterrance oftenanalyzed as conveying both an assertion (Ibelieve that…) and a question (Could you confirmthat…)
31
II. Speech acts in dialogue2.3 Generalisation to other speech actsIndirect speech acts(9) Peux-tu me passer le sel?
Could you pass me the salt?An utterance in which one speech act isperformed indirectly by performing another.A question used to perform a command.
32
II. Speech acts in dialogue2.3 Generalisation to other speech actsindirect speech acts• The Speaker makes as if she was askingabout the hearer’s ability. She’s committingto a question.
• But she’s expecting that the addresseeinterprets her utterance as a command.She wants that the addressee accepts thesuggestion and performs the correspondingaction.
33
II. Speech acts in dialogue2.3 Generalisation to other speech actsBy distinguishing two aspects in each speech act,we can easily account for :- Simple speech acts (declaratives conveyingassertions, interrogatives conveying questions,imperatives conveying command)
- Complex speech acts like biased questions orconfirmation requests
- Indirect speechacts.
34
II. Speech acts in dialogue2.4 Proposal• Associate each speech act with twoupdates, analyzed in terms ofcommitments.
• Most approaches focus only on theupdate of common ground or of thespeaker’s commitments.
35
II. Speech acts in dialogue2.4 ProposalAnalyze the Addressee-oriented aspect ofutterances in terms of commitment: Speakerwants Addressee to get committed.
The commitment that the speaker wants theaddressee to endorse may be a commitment toa proposition, or a question,or an outcome.
Dialogues as a negociation betweenSpeaker and Addressee.
36
II. Speech acts in dialogue2.4 Proposal• Speaker commits herself either to aproposition, or to a question or to anoutcome.
• Simultaneously, she calls on Addressee tocommit himself to a proposition, to aquestion, or to an outcome.
• By default, these two commitments sharethe same content.
37
II Speech acts in dialogueSentence Speech
actSpeakerCommit--ment
Call-onAddressee
Jeanavuquelqu’un.Johnsaw somebody.
Simpleassertion p p
Est-cequeJeanavuquelqu’un?DidJohnseeanybody?
Simplequestion ?p ?p
38
II. Speech acts in dialogue2.4 ProposalBut the content of the call-on-addresseemay be different from the content of thespeaker’s commitment.
Various constructional devices (particles,tags…) specify the type of content that thespeaker expects the addressee to getcommitted to.
39
II. Speech acts in dialogue2.4 ProposalConfirmation requests: a speech actcombining:Ø The speaker’s commitment to p (like in
assertion)Ø Call-on-Addressee to?p (like in question)
The speaker calls for the addressee to confirmher belief. The speaker is not stronglycommitted. She’s waiting for a commitmentfrom the addressee.
40
II. Speech acts in dialogue(10)Jeanavuquelqu’un,n’est-cepas?
Johnsaw somebody,didn’the?
Sentence Speechact
SpeakerCommitment
Call-on-Addressee
(10) Confirmationrequest p ?p
41
III. Other proposals3.1 Asher andReeseNegative bias inpolarquestion,SuB 9,2005
• Focusonquestionswhich convey anexpectationonthepartofthespeakerofanegativeanswer(e.g.questionswithNPI)
(11) DidJohnliftafingertohelpMary?
• Biased questionsarecomplex speechacts associated withthetypeassertion•question(asdottypesingenerativelexicon)
42
III. Other proposals3.1 Asher andReeseThey accountüneither forthedynamicity ofdialogues(assertionbefore its acceptation),
ünor forthedifference between confirmationrequests andbiased questions.
üWhy thetype assertion•question rather thanquestion•assertion?
43
III. Other proposals3.2 Farkas et BruceOnReacting toAssertionsandPolarQuestions,2009.- Focusonreactions toassertionsandpolar
questions,tocapturethesimilarities andthedifferencesbetween these speechacts.
- Characterize twotypesofresponses:confirmingandreversing reactions.
- Examinethedistributionofasetof‘polarity’particles inRomanian andacross languages.
44
III. Other proposals3.2 Farkas et Bruce(12)Anne:Samishome.
Ben:Yes/Yeah,he’s home./No,he isn’t home.(13)Anne:IsSamhome?
Ben:Yes/Yeah,he’s home./No,he isn’t home.
Ø SamereactionsØNo in(12)createsaconversationalcrisis,notin(13)ØAnne’sutterancein(12)andin(13)raisestheissueonSam’swhereabouts
45
III. Other proposals3.2 Farkas et BruceAmodelofcontextstructurewith therepresentationofthediscoursecommitments ofparticipants.They separate:ü thediscourse commitment setofeachparticipantü theCG(propositionsshared byallparticipants)ü a tablewhich registers QUDsThey account fordefaultconversationalmoves(assertionacceptance)andfordisagreements inordinary discourses.
46
III. Other proposals3.2 Farkas et Bruce
• Anassertionprojectsconfirmation(12)Samishome.(assertedinacontextwheres1isshared)
Anassertionyieldsanoutputcontextthatiscategoricallybiasedinfavorofconfirmationoftheassertedproposition.
A Table B
p <Samishome[D];{p}>
Common Grounds1
Projected Set{s1∪ {p}}
47
III. Other proposals3.2 Farkas et Bruce
• A polar question projects two possible answers.(13)IsSamhome?(askedinacontextwheres1isshared)
Defaultpolarquestionsarenon-biasing: theydon’tcommittheirauthortoeitherpropositionintheirdenotationandprojectaninquisitive contextwithrespecttotheirsentenceradical.
A Table B
<Samishome[I];{p,¬p}>
Com. Grounds1
Projected Set{s1∪ {p}, s1∪ {¬p}}
48
III. Other proposals3.2 Farkas et Bruce
• There are non default polar questions that are notimpartial.ü Polar interrogatives involving external negation(14) Isn’t Sam home? (asked in a context
where s1 is shared)
ü Polar interrogatives involving NPI (cf. Asher etReese)(15) Does Fred do a damn thing around the
house?
49
III. Other proposals3.2 Farkas et Bruce
To summarizeü discourse commitment set for each participantü a table with the utterance and its denotationü a common groundü a projection of future common groundDone:ü the difference between assertions, positive and negative
polar questionsTo be done:ü distinguish confirmation requests from questionü Account for different kinds of bias
50
III. Other proposals3.2 Farkas et Bruce
Similarity: emphasizingtheproposalnatureofassertionsAssertion: proposing additions to the common ground,rather than actually changing it.
Differences:- focus on reactions to assertions, on agreement and
disagrement between interlocutors- don’t account for the diversity of questions (biased,
rhetorical…)- don’t consider other types of speech acts (command...)
51
III. Other proposals3.3 Malamud and StephensonThreeWaystoAvoidCommitments:DeclarativeForceModifiersintheConversationalScoreboard,2011.InthelineofFarkas andBruce,butdiscussEnglishmarkersthat modify theforceofdeclarativeutterances:• reverse-polarity tags(Tom’s here,isn’t he?)• same-polarity tags(Tom’s here,ishe?)• rising intonation(Tom’s here?).
52
III. Other proposals3.3 Malamud and Stephenson
• Thethree markersallseem toindicate somekind ofuncertainty ofthespeaker,and/oradesire toseekconfirmationfrom theaddressee.
Ø Howtodistinguish them?Ø IsF&B’s framework fine-grained enough to
capturethedifferencebetween them?
53
III. Other proposals3.3 Malamud and Stephenson• reverse-polarity tags(Tom’s here,isn’t he?)Thespeaker isnotdirectly committing top,butisindicating thatifpisconfirmed,she will share responsibility forit.
• same-polarity tags(Tom’s here,ishe?)Thespeaker ismaking aguess astoB’s belief.IfBaccepts thismove,pisadded toB’s commitments.
• rising intonation (Tom’s here ↑)Risingdeclaratives arepossiblewhenever thespeaker isn’t sureifaplainassertionisappropriate.Theuncertainty licenses thespeaker inputtingametalinguistic issueaboutsuch anassertionontheTable.
54
III. Other proposals3.3 Malamud and Stephenson• TheyenrichFarkas &Brucemodel.Theypositpresentandprojectedversionsofü participants’commitments,ü theTable,ü andtheCommonGround.
• Theyintroduceprojectedcommitmentsforboththespeakerandthehearer,whichpermitstoaccountforthedyssymmetry betweenthem.
• TheprojectedTableallowsspeakerstotentativelyraiseissues.
55
III. Other proposals3.3 Malamud and StephensonSimilaritiesDistinguish actual andprojected CG
NewnessButalsoü actual andprojected commitments foreachparticipantsü Actual andprojected table
Ø Commitments may be dependent orindependent. Acanproposetoattribute acommitment ctoB.Inthis case,Bisthesourceofc.
56
III. Other proposals3.3 Malamud and StephensonAvantages• Account forsubtle distinctionsinlanguage• Thesystemincludes twoways forinformationtomake ittotheCommonGround.ü Thefirstway isviatheprojected CG.ü Thesecondisthat when both(all)participantsarepubliclycommitted toaproposition,this propositionisadded totheCG
ØAllpartsinthis systemareindependent.
57
III. Other proposals3.3 Malamud and StephensonAasserts p
Current ProjectedCG{…} CG*{{…,p},...,{…,p}}
Proposestoadd ptotheCGDC(A){…,p} DC(A)*{{…,p},...,{…,p}}
adds ptoA’s current andprojected commitmentsDC(B){…} DC(B)*{{…},...,{…}}DC(C){…} DC(C)*{{…},...,{…}}
No changeto B orC’scommitmentsTable<p,…> Table*{<…>,..., <…>}
Adds ptothetopofthetable;proposesthat itbe resolved
58
III. Other proposals3.3 Malamud and StephensonAutters pwith aRP-tag
Current ProjectedCG{…} CG*{{…,p},...,{…,p}}
Proposestoadd ptotheCGDC(A){…} DC(A)*{{…,p},...,{…,p}}
adds ptoA’sprojected commitmentsDC(B){…} DC(B)*{{…},...,{…}}DC(C){…} DC(C)*{{…},...,{…}}
No changeto B orC’scommitmentsTable<p,…> Table*{<…>,..., <…>}
Adds ptothetopofthetable;proposesthat itbe resolved
59
III. Other proposals3.3 Malamud and StephensonAutters pwith aSP-tag
Current ProjectedCG{…} CG*{{...},...,{…}}
NochangetotheCGDC(A){…} DC(A)*{{…},...,{…}}
nochangetoA’s commitmentsDC(B){…} DC(B)*{{…,p},...,{…,p}}
Adds pto B’s projected commitmentsTable<p,…> Table*{<…>,..., <…>}
NochangetotheTable
60
III. Other proposals3.3 Malamud and StephensonAutters pwith arising intonation
Current ProjectedCG{…} CG*{{...},...,{…}}
nochangetotheCGDC(A){…} DC(A)*{{…,p},...,{…,p}}
adds ptoA’s projected commitmentsDC(B){…} DC(B)*{{…},...,{…}}
nochangeto B’commitmentsTable<MLI(p),…> Table*{<p,…>,..., <p,…>}
adds ptotheprojected Table- pisexpected tobecome anissue:adds ametalinguistic issue(MLI(p)totheTable
61
III. Other proposals3.4 KrifkaBias incommitment space semantics:declarative questions,negated questionsandquestiontags,2015• Aframework forillocutionary acts• Captureinformationshared byinterlocutors,butalso possible
continuationsü Assertions:commitments forthetruth ofpropositionsü Questions:movesthat restrict thecontinuationto
assertionsbyother participants• Applied todifferent typesofquestions:biased,negated,
questionstags…
Moredetails inLecture3onbiased questions.
62
Conclusions• Speech acts rather than sentences• Commitments rather than belief• Distinguish between what is (effective) CGand projected CG
• Distinguish between private beliefs andpublic commitments
• Commitments to propositions and to nonpropositional contents.
63
ConclusionsØUse commitments for accounting for thedouble aspect of speech acts.
ØAccount for the division of labour,disentangling contributions of syntax(clause type), lexical semantics(particles), and prosody to theseupdates.
64
Conclusions
65
ReferencesN. Asher and B. Reese. 2005. 'Negative bias in polar questions'. In E. Maier, C.
Bary, and J. Huitink (eds), Proceedings of SuB9, 30–43.C.Beyssade&J.-M.Marandin.2006.’FrenchIntonationandAttitude
Attribution'.InDenisP.etal. (eds) Issuesatthesemantics-pragmaticsinterface,Selected papers from TLS8.Somerville:Cascadilla Press.
C.Beyssade&J.-M.Marandin.2006.‘Thespeechact assignment problemrevisited:Disentangling Speaker’scommitment from Speaker’scallonAddressee’ CSSP’sproceedings. http://www.cssp.cnrs.fr/.
C.Beyssade &J.-M.Marandin.2006.From Complex to Simple SpeechActs:aBidimensionalAnalysis of Illocutionary Forces.InD.Schlangen &R.Fernandez(eds.),Proceedings of BRANDIAL06,Potsdam,42-49.
D.Farkas &K.Bruce.2010.OnReacting toAssertionsandPolarQuestions,JournalofSemantics 27,81-118.
G.Gazdar.1981.Speechactassignment.InJoshi,WebberandSag(eds.),ElementsofDiscourseUnderstanding.CambridgeUniversityPress.64-83.
J.Ginzburg,1997.QueryingandAssertion inDialogueI,HebrewUniversityofJerusalem,ms.
66
ReferencesJ. Ginzburg. 2012. The interactive Stance: Meaning in conversation, Oxford
University Press.J. Ginzburg and I. A. Sag. 2000. Interrogative investigations. Stanford: CSLI.C.Gunlogson.2001.TruetoForm:RisingandFallingDeclarativesas
QuestionsinEnglish.PHDThesis,publishedin2003,Routledge.C.L.Hamblin.1970.Fallacies,Methuen,London.C.L.Hamblin.1971.Mathematical models ofdialogue,Theoria 37:130-155.F.Hamlaoui.2010.OntheroleofphonologyanddiscourseinFrancilian
Frenchwh-questions. JournalofLinguistics47:1–34M.Krifka.2015.BiasinCommitmentSpaceSemantics:Declarativequestions,
negatedquestions,andquestiontags.InProceedingsofSALT25,328-345.M.Krifka.2017.Negativepolarityquestionsasdenegationsofassertions. In
L.Chungmin,F.Kiefer&M.Krifka,Contrastiveness ininformationstructure,alternativesandscalarimplicatures,359-398.
67
ReferencesS.A.Malamud&T.Stephenson. 2011.ThreeWaystoAvoidCommitments:
DeclarativeForceModifiersintheConversationalScoreboard. SemDial.C.Portes,C.Beyssade, A.Michelas,J.-M.Marandin,M.Champagne-Lavau.
2014. TheDialogical DimensionofIntonational Meaning:EvidencefromFrench.JournalofPragmatics, 74,15-29.
P.Portner.2015.CommitmenttoPriorities. InD.Fogel,D.Haris &M.Moss(eds.),NewWorkonSpeechacts,OxfordUniv.Press,296-316.
A.Stefanowitsch.2003. TheEnglishimperative:Aconstruction-basedapproach.WorkshopForm andFunction ofSentenceTypes,Munchen,ms.
D.Walton.2000.Theplaceofdialoguetheory inlogic,computerscienceandcommunicationstudies, Synthese 123:327-346.
D.Zaefferer. 2001.Deconstructing aclassical classification:Atypological lookatSearle's conceptofillocutiontype.RevueInternationaledePhilosophie2:209-225.