Epistemic)Parenthe/cal)Verbs)and) Associaon)with)Focus)hedberg/Paris_presentation3.pdf ·...
Transcript of Epistemic)Parenthe/cal)Verbs)and) Associaon)with)Focus)hedberg/Paris_presentation3.pdf ·...
Epistemic Parenthe/cal Verbs and Associa/on with Focus
Nancy Hedberg and Noureddine Elouazizi Simon Fraser University
Burnaby, BC Canada
Parenthe/cal verbs
• Urmson (1952) – Parenthe/cal verbs “guide the hearer to an apprecia/on
of the matrix statement in its social, logical, or eviden/al context.”
(1) It's true i mean <laughter> acceptance i guess is the key. – The proposi/on is asserted as a guess, not a fact.
(2) So he took um English-‐Comp which he got a an A minus in i believe because he really he's a really good writer. – The proposi/on expressed in the rela/ve clause is
asserted as a belief, not a fact. 2 Parenthé/ques 2012
Outline of Talk
1. Corpus Study 2. Associa/on with Focus
3. Interim Conclusions
4. Implica/ons for Syntax
1. Corpus Study
Parenthé/ques 2012 4
Corpus Study: Methodology
• We searched two spoken telephone corpora of American English (Callhome and Fisher) for epistemic parenthe/cal verbs men/oned by Urmson (1952) and Ross (1973).
• We exhaus/vely searched for sentence-‐medial parenthe/cal verbs and found 47 instances. I think (19), I guess (11), I believe (8), I suppose (6), I know (1), I suspect (1), I presume (1)
5 Parenthé/ques 2012
Corpus Study: Methodology (cont.)
• We omiced instances where the verb occurred at the beginning of a clause and thus was not clearly parenthe/cal.
• We also did a par/al search for sentence final parenthe/cal verbs in the Fisher Corpus so that we could compare them with medial parenthe/cal verbs: I think (10+), I guess (10+), I believe (10+), I suppose (10+), I know (0), I suspect (1), I presume (1)
Parenthé/ques 2012 6
Corpus Study: Prosody
• Using PRAAT, we are analyzing the prosody of each token. – We are aiming at a full ToBI analysis, including break indices, but are only par/ally there (Beckman & Ayers-‐Elam 1997)
• Confirmed for American English the findings of Peters (2006) for Hamburg German, and Kaltenböck (2008) and Dehé (2009) for Bri/sh English, that a parenthe/cal verb in clause-‐medial posi/on is typically phrased prosodically with preceding or following material.
7 Parenthé/ques 2012
Prosodic Nota/on
(1) It's true i mean <laughter> [(acceptance#i guess is the KEY)]. (2) [(So he TOOK um)(English-‐ COMP)] [(which he GOT a) (an A
MINUS in#i believe)] [(because he REALLY) (he's a really good WRITER)].
– [Intona/onal phrase] – (intermediate phrase) – NUCLEAR ACCENT – #: links parenthe/cal to cons/tuent it is phrased with. – FOCUS EXPONENT OF PHRASE WITH PARENTHETICAL. – seman/c scope of parenthe/cal verb.
8 Parenthé/ques 2012
Corpus Study: Seman/cs
• We are also analyzing the seman/cs of each example.
• Confirmed the finding of Kaltenböck (2008) that the scope of the epistemic mi/ga/on of the parenthe/cal verb can be an en/re clause or only a subcons/tuent.
9 Parenthé/ques 2012
Medial Posi/on: Clausal Scope
• When the parenthe/cal occurs between subject and predicate as in (1), parenthe/cal verbs tend to take the whole clause in their seman/c scope (excluding topics), and in this case they tend to be prosodically acached to the subject.
(1)It's true i mean <laughter> [(acceptance#i guess is the KEY)].
10 Parenthé/ques 2012
Medial Posi/on: Narrow Scope
• When they occur lower in the clause, they take only the phrase that they are prosodically acached to in their seman/c scope.
(4) [(He's LIKE) (well now i'm ACTIVELY looking for a job)] [(a-‐ as opposed to PASSIVELY#i guess) (looking for a JOB)].
(5) [There's a TRIP coming up] [(i think#like December the EIGHTH)].
(6) [(The airport TERRORISTS that) (you KNOW) (were in control during nine ELEVEN)] [(they ALL had) (FAKE ah) (DRIVERS licence from) (i believe#FLORIDA)].
11 Parenthé/ques 2012
Final Posi/on • Sentence-‐final parenthe/cal verbs can take the whole sentence in their scope:
(7) [(WELL) (it could go the other way AROUND) (TOO)(i SUPPOSE)]
• Or they can take just the final cons/tuent in their scope:
(8) [(well ACTUALLY) (in washington STATE) (it's seven twenty FIVE#i believe)]
Parenthé/ques 2012 12
2. Associa/on with Focus
Associa/on with Focus: Focus Adverbs
• Rooth (1985) shows that ‘only’ in pre-‐predicate posi/on has scope only over the cons/tuent that is focused in the predicate.
• Hence (9a-‐b) have different truth condi/ons:
(9) a. Fred only introduced Bill to SUEF. b. Fred only introduced BILLF to Sue.
• If Fred introduced Bill to Sue, and John to Sue and Mary, then (9a) is true but (9b) is false.
Parenthé/ques 2012 14
Associa/on with Focus: Focus Adverbs
• ‘Only’ can also be placed in posi/on con/guous to the focus in its scope, as shown in (10).
(10)a. Fred introduced Bill only to SUEF.
b. Fred introduced only BILLF to Sue.
Parenthé/ques 2012 15
Associa/on with Focus: Parenthe/cal Verbs
• We suggest that parenthe/cal verbs also associate with focus, in the sense that only the focused material in the clause or phrase containing the parenthe/cal has its epistemic force mi/gated by the parenthe/cal verb.
• (11) shows a variant of (6), where the parenthe/cal verb appears in matrix posi/on.
(11) The 9/11 terrorists, I believe that they all had fake drivers licenses from Florida.
Parenthé/ques 2012 16
Associa/on with Focus: Parenthe/cal Verbs (cont.)
• Several focus structures are possible, due to the possibility of percola/on of F-‐marking (Selkirk 1995), but one op/on is to have F-‐marking (and hence focus) narrowly restricted to just the DP ‘Florida’, as in (c).
(11’) The 9/11 terrorists, I believe that [they all had [fake drivers licenses from [FLORIDA]F ]F ]F. a. What about the 9/11 terrorists? b. What did the 9/11 terrorists have? c. Where did the 9/11 terrorists have fake drivers licenses from?
Parenthé/ques 2012 17
Associa/on with Focus: Parenthe/cal Verbs (cont.)
• When the parenthe/cal is placed between subject and predicate the same focus possibili/es obtain for the material excluding the parenthe/cal:
(12) The 9/11 terrorists, [they all, I believe, had [fake drivers licenses from [FLORIDA]F ]F ]F. a. What about the 9/11 terrorists?
b. What did the 9/11 terrorists have? c. Where did the 9/11 terrorists have fake drivers licenses from?
Parenthé/ques 2012 18
Associa/on with Focus: Parenthe/cal Verbs (cont.)
• The pre-‐focus material in the narrow-‐focus case in (a) and (b) would have to be “given” in the discourse, as explained in Schwarzschild (1999).
• But in (c), the proposi/on that the terrorists all had fake drivers licenses can be asserted as new informa/on.
(13) a. The 9/11 terrorists, I believe that [they all had fake drivers licenses from]G [FLORIDA]F. b. The 9/11 terrorists, [they all]G, I believe, [had fake drivers licenses from]G [FLORIDA]F. c. The 9/11 terrorists, they all had fake drivers licenses from I believe [FLORIDA]F.
Parenthé/ques 2012 19
Associa/on with Focus: Parenthe/cal Verbs (cont.)
• The possibility of allowing preceding material to be asserted as new informa/on perhaps provides mo/va/on to the grammar to permit low-‐acachment of parenthe/cal verbs.
Parenthé/ques 2012 20
Exclusion of Topics
• If parenthe/cal verbs associate with focus, then topics and presupposed material should be excluded from their scope.
• And indeed in the it-‐clew in (14), the presupposed material in the that-‐clause, which func/ons as the topic of the sentence, is excluded from the scope of the parenthe/cal verb.
(14) (NOW it's) (it's the BLACKS#i think)(more than the WHITES) (that are keeping it ALIVE).
Parenthé/ques 2012 21
Exclusion of Topics (cont.)
• And when the subject of the sentence is definite and topical, as in (15) where the interlocutors have been discussing San Francisco, it seems to be excluded.
(15) [But uh (san FRANCISCO)#(i BELIEVE) (has a) (eight) (or nine LIVING wage) (UM) (on the BOOKS) (WITH) (i THINK)#(eleven PROPOSED).
Parenthé/ques 2012 22
3. Interim Conclusions
Conclusions from the corpus study
• We conclude that parenthe/cal verbs take the focus of the ucerance in which they occur as their seman/c scope; topics are excluded.
• When they adjoin low but internal to a clause, their scope is narrow.
• When they occur finally, between subject and predicate, or in matrix superordinate posi/on, they can take the whole clause as their scope.
Parenthé/ques 2012 24
4. Implica/ons for Syntax
Implica/ons for syntax • We hope to integrate parenthe/cal verbs into a cogni/on-‐oriented minimalist syntac/c framework. – A sentence is built in ‘narrow syntax’ by hierarchical merging of lexical items or phrases that had been selected by the conceptual sentence planning system to be put into the numera/on.
– The deriva/on as a whole is spelled out in a single phase or perhaps phase by phase into a syntac/c buffer in the deriva/onal work space.
– Then from lew-‐to-‐right, chunks of the deriva/on are sent into the phonological component of the grammar to become prosodic phrases.
Parenthépiques 2012 26
Implica/ons for syntax (cont.) – Just before a chunk is sent into PF, a seman/c, truth scanning process guided by the conceptual sentence planning system checks it for accuracy.
– If it is deemed necessary, an epistemic parenthe/cal verb is merged as an adjunct in the buffer. • It is either merged at the lew edge of the current segment or tacked on to the right edge.
– Thus, parenthe/cal verbs are at least some/mes inserted post-‐syntac/cally—within a stage of speech produc/on which is subsequent to the ini/al planning+execu/on stage that governs the original numera/on and deriva/on in narrow syntax.
Parenthé/ques 2012 27
Implica/ons for syntax (cont.)
– However, they s/ll are generated within the linguis/c computa/onal system that outputs the completed deriva/on to the ar/culatory-‐perceptual and conceptual-‐inten/onal systems.
– We hope to thus account for the apparent hybrid nature of parenthe/cal verbs as both an integrated part of the sentence produced but also somehow external to the syntac/c core of the sentence, i.e. parenthe/cal.
Parenthé/ques 2012 28
Implica/ons for syntax (cont.) • We excluded clause-‐ini/al instances of epistemic predicates from
our corpus study because we couldn’t tell from the transcripts if they were parenthe/cal or not.
• If parenthe/cal verbs are inserted post-‐syntac/cally for purposes of on-‐line repair, then perhaps such inser/on should not be possible clause-‐ini/ally unless only the chunk of informa/on corresponding to the first phrase is modified.
• And indeed, (16) shows a clause ini/al epistemic verb that does not sound like it is parenthe/cal, compared with the clause-‐final epistemic verb, which does sound parenthe/cal.
(16) [(UH)( i think it was COMEDY]<laughter>[(and WHERE ah) (and HOW ah) (where you draw the LINE between) (what's funny and GOOD comedy) (and what's in bad TASTE)#(i THINK)].
Parenthé/ques 2012 29
References • Beckman, Mary, and Gayle Ayers-‐Elam. 1997. Guidelines for ToBI Labelling. Ohio State University. • Boersma, Paul, and David Weenink. 2006. Praat: doing phone/cs by computer [computer program]
version 4.4.04. hcp://www.praat.org/. • Canavan, Alexandra; David Graff, and George Zipperlen. 1997. CALLHOME American English Speech.
Linguis/c Data Consor/um. Philadelphia. • Cieri, Christopher; David Graff; Owen Kimball; Dave Miller; and Kevin Walker. 2004. Fisher English
Training Speech Part 1 Speech; Transcripts. Linguis/c Data Consor/um. Philadelphia • Dehé, Nicole. 2009. Clausal parenthe/cals, intona/onal phrasing, and prosodic theory. Journal of
Linguis/cs 45. 569-‐615. • Kaltenböck, Gunther. 2008. Prosody and func/on of English comment clauses. Folia Linguis/ca 42
(1). 83-‐134. • Peters, Jörg. 2006. Syntac/c and prosodic parenthesis. Proceedings of the 3rd Interna/onal
Conference on Speech Prosody. • Rooth, Mats. 1985. Associa/on with Focus. Doctoral Disserta/on, University of Massachusecs at
Amherst. • Ross, John Robert. 1973. Sliwing. In Maurice Gross, Morris Halle & Marcel-‐Paul Schü̈tzenberger
(eds.), The Formal Analysis of Natural Languages: 1st Interna/onal Conference, 133–169. The Hague & Paris: Mouton.
• Schwarzschild, Roger. 1999. Givenness, AvoidF and other constraints on the placement of accent. Natural Language Seman/cs 7. 141-‐177.
• Selkirk, Elisabeth. 1995. Sentence prosody: Intona/on, stress, and phrasing. In The Handbook of Phonological Theory, ed. by John Goldsmith, chapter 16, 550–569. London: Blackwell.
• Urmson, J. O. 1952. Parenthe/cal verbs. Mind 61. 480-‐496.
Parenthé/ques 2012 30