English as a discourse ergative language Existential unaccusative sentences

Tricia Irwin Swarthmore College, US [email protected]

Between Existence and Location Empirical, Formal and Typological Approaches to Existential Constructions 1-2 December 2016  Universität Tübingen

Introduction

3

Main claim • Subtype of intransitive sentence in English o

“Existential unaccusative”

1. A princess waltzed in.

(existential unaccusative)

2. There was a princess who waltzed in. • Existential unaccusatives share structure and meaning with existential there BE sentences

4

Shared properties 1. A princess waltzed in.

(existential unaccusative)

2. There was a princess who waltzed in. (there BE) • Patterns with respect to syntactic and semantic diagnostics • Discourse function: Establishment of new discourse referents

5

Existential unaccusatives 1. A princess waltzed in. 2. A cab pulled up. 3. A clown came over. 4. A little boy darted out • Prototypically involve verbs of motion o o

Verbs of motion: waltz, dart, come, run Particle-like P element: in, up, over, out

6

Overview of the analysis 1.

A princess waltzed [PredP [PathP in [PlaceP TO here ] ] ]

• Subject DP (a princess) starts out VP-internally o

Semantically is of property (or higher) type

• Particle-like P element (in) heads a PathP • Path is to a silent, contextually-determined spatio-temporal location (Francez 2007) o o

Implemented as PlaceP “TO/AT here” “here” stands in for the contextually-determined location

7

Overview of the analysis 1.

A princess waltzed [PredP [PathP in [PlaceP TO here ] ] ]

• Subject DP (a princess) starts out VP-internally o

Just like an Ǝ pivot

Semantically is of property (or higher) type

• Particle-like P element (in) heads a PathP • Path is to a silent, contextually-determined spatio-temporal location (Francez 2007) o o

Implemented as PlaceP “TO/AT here” “here” stands in for the contextually-determined location

8

Overview of the analysis 1.

A princess waltzed [PredP [PathP in [PlaceP TO here ] ] ]

• Subject DP (a princess) starts out VP-internally o

Semantically is of property (or higher) type Just like an Ǝ pivot

• Particle-like P element (in) heads a PathP • Path is to a silent, contextually-determined spatio-temporal location (Francez 2007) o o

Implemented as PlaceP “TO/AT here” “here” stands in for the contextually-determined location

9

Overview of the analysis 1.

A princess waltzed [PredP [PathP in [PlaceP TO here ] ] ]

• Subject DP (a princess) starts out VP-internally o

Semantically is of property (or higher) type

• Particle-like P element (in) heads a PathP

UNLIKE AN EXISTENTIAL

• Path is to a silent, contextually-determined spatio-temporal location (Francez 2007) o o

Implemented as PlaceP “TO/AT here” “here” stands in for the contextually-determined location

10

Overview of the analysis 1.

A princess waltzed [PredP [PathP in [PlaceP TO here ] ] ]

• Subject DP (a princess) starts out VP-internally o

Semantically is of property (or higher) type

• Particle-like P element (in) heads a PathP • Path is to a silent, contextually-determined spatio-temporal location (Francez 2007) Just like an existential o o

Implemented as PlaceP “TO/AT here” “here” stands in for the contextually-determined location

11

Overview of the analysis 1.

A princess waltzed [PredP [PathP in [PlaceP TO here ] ] ]

• Subject DP (a princess) starts out VP-internally o

Semantically is of property (or higher) type

Just like an Ǝ pivot

Just like an Ǝ pivot

• Particle-like P element (in) heads a PathP • Path is to a silent, contextually-determined spatio-temporal location (Francez 2007) Just like an existential o o

Implemented as PlaceP “TO/AT here” “here” stands in for the contextually-determined location

12

Sketch of the analysis, in tree form 1. A princess waltzed [PredP [PathP in [PlaceP TO here ] ] ]

13

Roadmap 1.

Introduction 1. 2.

2.

Proposal Argument structure and discourse

Properties shared between existentials and existential unaccusatives 1. 2. 3. 4.

There BE existentials and new discourse referents Discourse coherence PP extraposition “Semantic bleaching”

3.

Conclusion

4.

Coda: Details of the analysis & additional data

14

Roadmap 1.

Introduction 1. 2.

2.

Proposal Argument structure and discourse

Properties shared between existentials and existential unaccusatives 1. 2. 3. 4.

There BE existentials and new discourse referents Discourse coherence PP extraposition “Semantic bleaching”

3.

Conclusion

4.

Coda: Details of the analysis & additional data

15

Roadmap 1.

Introduction 1. 2.

2.

Proposal Argument structure and discourse

Properties shared between existentials and existential unaccusatives 1. 2. 3. 4.

There BE existentials and new discourse referents Discourse coherence PP extraposition “Semantic bleaching”

3.

Conclusion

4.

Coda: Details of the analysis & additional data

16

Existentials and discourse function • Existential sentences have special discourse properties o

Establishment of new discourse referents (McNally 1992, 1997) • Not much debate about this

• “The assertion of an existential sentence has the effect of introducing an additional discourse referent instantiating the argument of the existential predicate” (McNally 1997: 6) • Or the effect of re-establishing as a topic a previously introduced discourse referent—see Ward & Birner (1995)

17

Intransitives and discourse function • Intransitive sentences have been claimed to have special discourse properties o

o

Lambrecht (1994): Principle of the Separation of Reference and Role (PSRR) DuBois (1987): Preferred Argument Structure (PAS)

18

Intransitives and discourse • Lambrecht (1994): Principle of the Separation of Reference and Role o

PSRR is put as a “simple pragmatic maxim”: • “Do not introduce a referent and talk about it in the same clause.” (p. 185)

o

Not directly about intransitive sentences • It is about “presentation”

19

DuBois (1987): Preferred Argument Structure • Formulated PAS from work on Sakapultek, a Mayan language, which is ergative • DuBois’s finding: New discourse referents prefer the S and O roles (those marked ABS) o o

Transitive sentence: Intransitive sentence:

A (verb) O S (verb)

20

Preferred Argument Structure • DuBois’s hypothesis: The distribution of information in argument positions is skewed toward an ergative pattern o

Intransitive subjects and transitive direct objects are linked in discourse function

• Transitive sentence:

A (verb) O

• Intransitive sentence:

S (verb)

21

PAS: The perspective from English • Transitive sentence:

A (verb) O

• Intransitive sentence:

S (verb)

o

New information in the O role is not surprising • English largely obeys the given-new contract (Clark and Haviland 1977)

22

PAS: The perspective from English • Transitive sentence:

A (verb) O

• Intransitive sentence:

S (verb)

• But new information in intransitive subjects?? o

Weird! (Prince 1981, 1992; Michaelis & Francis 2007)

23

PAS in English and existential unaccusativity • My proposal: The prediction from PAS for the O role holds for a subset of intransitive sentences o

These sentences are: • Unaccusative: the subject starts out VP-internally • Existential: they make use of the same elements of structure and meaning as existential there BE sentences do

• In this way, English can be thought of as a discourse- (split) ergative language

24

Same verb: Two possible structures • Many verbs of motion (run, dance, waltz, dart) can occur in either an unergative or unaccusative structure o

As Hoekstra & Mulder (1990) showed for Dutch

1. A beautiful lady waltzed gracefully. 2. A beautiful lady waltzed in.

(unergative)

(existential unaccusative)

• (2) is a sub-type of existential sentence • Note the difference in prosody

25

Roadmap 1.

Introduction 1. 2.

2.

Proposal Argument structure and discourse

Properties shared between existentials and existential unaccusatives 1. 2. 3. 4.

There BE existentials and new discourse referents Discourse coherence PP extraposition “Semantic bleaching”

3.

Conclusion

4.

Coda: Details of the analysis & additional data

Existentials and existential unaccusatives Shared Properties

27

Shared properties • Suggestive: existential unaccusatives have there-BE counterparts 1. A little boy ran in. 2. There’s a little boy who ran in.

28

The outlaws of the ES world • Further suggestive evidence: there-insertion is possible with verbs of motion 1. A little boy ran in (to the room). 2. There ran into the room a little boy. • Milsark (1974) called sentences like (2) “the outlaws of the ES world” (246) o

Some speakers reject these (they sound literary)

29

Unergatives, by contrast • Unergatives don’t have such counterparts (Levin 1995) 1. A little boy ran quickly.

(unergative)

2. *There ran quickly a little boy. 3. A beautiful lady waltzed gracefully.

(unergative)

4. *There waltzed gracefully a beautiful lady.

30

Discourse coherence • Establishment of new discourse referents o

Modulo factors discussed in Karttunen (1976)

• There BE sentences 1. There was a unicorn at the renaissance fair. … It was sparkling white. • Transitive direct objects 2. I saw a unicorn at the renaissance fair. … It was sparkling white.

31

Subjects and discourse coherence • Transitive subjects o

1.

Can establish discourse referents, but not very coherently

A frog kissed the princess. a. … ??It was green and covered with warts. b. … It was a frightening sight.

• Unergative subjects o

1.

Most natural subsequent reference is to the event of the unergative

A unicorn whinnied a. … ??It was sparkling white. b. … It was a beautiful sound.

32

Existential unaccusatives • Existential unaccusative subjects pattern with pivots of existential there BE sentences o

And with transitive direct objects

1. A unicorn galloped over. … It was sparkling white. 2. A unicorn walked in. … It was sparkling white. • Tempting to call this ‘presentation’ o

BUT: canonical word order

33

Not all unaccusatives pattern like this • Change-of-state unaccusatives pattern differently o

These verbs typically allow causative/inchoative alternation (e.g., break, freeze, melt)

1. A unicorn froze. … ?? It was sparkling white. 2. A unicorn fell. … ?? It was sparkling white.

34

Roadmap 1.

Introduction 1. 2.

2.

Proposal Argument structure and discourse

Properties shared between existentials and existential unaccusatives 1. 2. 3. 4.

There BE existentials and new discourse referents Discourse coherence PP extraposition “Semantic bleaching”

3.

Conclusion

4.

Coda: Details of the analysis & additional data

35

Guéron (1980): PP extraposition • Asymmetries in PP extraposition: Transitive subjects 1. A lady with auburn hair read a book on cats. o

Two interpretations of with auburn hair

2. A lady read a book on cats with auburn hair. o

(extraposed)

Only one of the interpretations is now available • The strange one: the lady is holding some auburn hair

36

Intransitives and PP extraposition • Asymmetries in PP extraposition: Intransitive subjects 1. A lady with auburn hair danced gracefully. o

Two interpretations

2. A lady danced gracefully with auburn hair. o

(extraposed)

Only one interpretation • Again, the strange interpretation

o

(1) and (2) are unergatives, not existential unaccusatives

37

Intransitives and PP extraposition • Existential unaccusatives allow both interpretations when the PP is extraposed 1. A lady with auburn hair walked in. o

Two interpretations

2. A lady walked in with auburn hair. o

Both interpretations available!

(extraposed)

38

Guéron’s (1980) explanation • When does PP extraposition (from subjects) allow both interpretations? o

Only when sentence can be interpreted with a Presentation LF

• Guéron posits a Presentation LF and a Predication LF

39

Predication and presentation • Guéron’s Presentation LF: o

o

Verb meaning is bleached or “pragmatically emptied” of semantic content (Guéron 1980: 653-4) “The VP denotes, essentially, the appearance of the subject in the world of the discourse.” (p. 653)

• Guéron’s Predication LF: subject-predicate o

Topic-comment, theme-rheme, etc.

40

Guéron (1980): The presentation LF • Meaning: o

“The VP denotes, essentially, the appearance of the subject in the world of the discourse.” (Guéron, 1980: 653)

• Structure: o

Note the similarity to the proposed (underlying) structure for existential unaccusatives:

1. (s VERBi (s (NP) (. . . vi . . . ) ) ) • Verb moves at LF: remember, this was 1980…

41

Returning to existential unaccusatives • Guéron’s diagnostic for a Presentation LF holds for our existential unaccusatives 1. A lady waltzed in with auburn hair. 2. A cab pulled up with tinted windows. 3. A man came in with green eyes. • When a subject PP is extraposed, both interpretations are available

42

What we get from Guéron • Intransitive sentences show an asymmetry with respect to interpretation of PP extraposition 1. A lady danced gracefully with auburn hair. o

Only one interpretation

2. A lady walked in with auburn hair. o

(unergative)

Two interpretations

(existential unaccusative)

43

There BE sentences and PP extraposition 1. There’s a lady with auburn hair at the door. o

Not extraposed: two interpretations

2. There’s a lady at the door with auburn hair. o

Two interpretations (?)

3. A lady walked in with auburn hair. o

Two interpretations

(existential unaccusative)

44

Existential verb meaning • Existential verbs: “semantically empty” or “bleached” o

(Borschev & Partee 1998: 87,90)

• Partee et al. (2011): “the verbs that may occur in existential sentences are an open class; some are independently characterizable as existential or perceptual, and others may undergo `semantic bleaching''' (p. 138) 1. VBE (THING, LOC)

(ParteeEtAl2011, 142)

45

Existential unaccusatives and verb meaning • The verb meaning is still present—there is still a waltzing event in (1) 1. A lady waltzed in with auburn hair. 2. A cab pulled up with tinted windows. 3. A man came in with green eyes. • But the VP comes to mean “coming on the scene”

46

Guéron on bleaching • Guéron points out in her examples that none of the verbs are straightforwardly synonymous with “appear” (p. 653) • Guéron’s examples (653; ex. 56): 1. A man walked in from India. 2. A train chugged past with many passengers. 3. A bird darted by with golden wings.

47

Roadmap 1.

Introduction 1. 2.

2.

Proposal Argument structure and discourse

Properties shared between existentials and existential unaccusatives 1. 2. 3. 4.

There BE existentials and new discourse referents Discourse coherence PP extraposition “Semantic bleaching”

3.

Conclusion

4.

Coda: Details of the analysis & additional data

48

Roadmap 1.

Introduction 1. 2.

2.

Proposal Argument structure and discourse

Properties shared between existentials and existential unaccusatives 1. 2. 3. 4.

There BE existentials and new discourse referents Discourse coherence PP extraposition “Semantic bleaching”

3.

Conclusion

4.

Coda: Details of the analysis & additional data

49

Roadmap 1.

Introduction 1. 2.

2.

Proposal Argument structure and discourse

Properties shared between existentials and existential unaccusatives 1. 2. 3. 4.

There BE existentials and new discourse referents Discourse coherence PP extraposition “Semantic bleaching”

3.

Conclusion

4.

Coda: Details of the analysis & additional data

50

Roadmap 1.

Introduction 1. 2.

2.

Proposal Argument structure and discourse

Properties shared between existentials and existential unaccusatives 1. 2. 3. 4.

There BE existentials and new discourse referents Discourse coherence PP extraposition “Semantic bleaching”

3.

Conclusion

4.

Coda: Details of the analysis & additional data

51

Roadmap 1.

Introduction 1. 2.

2.

Proposal Argument structure and discourse

Properties shared between existentials and existential unaccusatives 1. 2. 3. 4.

There BE existentials and new discourse referents Discourse coherence PP extraposition “Semantic bleaching”

3.

Conclusion

4.

Coda: Details of the analysis & additional data

52

Roadmap 1.

Introduction 1. 2.

2.

Proposal Argument structure and discourse

Properties shared between existentials and existential unaccusatives 1. 2. 3. 4.

There BE existentials and new discourse referents Discourse coherence PP extraposition “Semantic bleaching”

3.

Conclusion

4.

Coda: Details of the analysis & additional data

53

Recap: Main claim • I have presented arguments that sentences like (1)-(4) are best analyzed as a subspecies of existential sentence 1. A princess waltzed in. 2. A cab pulled up. 3. A clown came over. 4. A little boy darted out

54

Shared properties • Availability of there-insertion • Discourse function: Establishment of new discourse referents o

Judgments on discourse coherence

• PP extraposition • Verb meaning and “semantic bleaching”

55

Consequence of the analysis • Support for a generalization about the “information flow of discourse” that otherwise seems not to hold for English • DuBois (1984): Preferred Argument Structure o

o

The distribution of information in argument positions is skewed toward an ergative pattern. New information prefers direct objects and intransitive subjects.

56

PAS: The perspective from English • Transitive sentence:

A (verb) O

• Intransitive sentence:

S (verb)

• Intransitive subjects and transitive objects are linked in discourse function • But in English, only a subset of intransitive subjects follow the ergative pattern of discourse o

These are the existential unaccusatives

57

PAS in English and existential unaccusativity • The prediction from PAS for the S role holds for a subset of intransitive sentences o

These sentences are: • Unaccusative: the subject starts out VP-internally • Existential: they have same elements of structure and meaning as existential there BE sentences

• In this way, English can be thought of as a discourse- (split) ergative language

Thank you! …time for chocolate!

59

Selected references Borschev, Vladimir, and Barbara H. Partee. 1998. Formal and lexical semantics and the genitive in negated existential sentences in Russian. In Formal Appproaches to Slavic Linguistics: The Connecticut Meeting 1997 , ed. Željko Bošković, Steven Franks, and William Snyder, 75–96. Ann Arbor: Michigan Slavic Publications. Bowers, John. 1993. The syntax of predication. Linguistic Inquiry 24:591–656. Clark, Herbert H., and Susan E. Haviland. 1977. Comprehension and the given-new contract. In Discourse Production and Comprehension, ed. Roy O. Freedle, 1–40. Norwood, NJ: Ablex Publishing Corporation. Davies, Mark. (2008-) The Corpus of Contemporary American English: 520 million words, 1990-present. Available online at http://corpus.byu.edu/coca/. Francez, Itamar. 2007. Existential propositions. Ph.D. dissertation, Stanford University. Guéron, Jacqueline. 1980. On the syntax and semantics of PP extraposition. Linguistic Inquiry 11:637–678. Hoekstra, Teun, and René Mulder. 1990. Unergatives as copular verbs: Locational and existential predication. The Linguistic Review 7:l–79. Karttunen, Lauri. 1976. Discourse referents. In Notes from the Linguistic Underground (Syntax and Semantics, vol. 7), ed. James D. McCawley. New York: Academic Press. McCloskey, James. 2014. Irish existentials in context. Syntax 343–384. McNally, Louise. 1997. A semantics for the English existential construction. New York: Garland. Partee, Barbara H., Vladimir Borschev, Elena V. Paducheva, Yakov Testelets, and Igor Yanovich. 2011. Russian genitive of negation alternations: The role of verb semantics. Scando-Slavica 57:135–159. Stowell, Tim. 1978. Stylistic movement rules (ms). MIT. Szekely, Rachel. 2015. Truth without predication: The role of placing in the existential there-sentence. London: Palgrave Macmillan.

Coda Additional sections: Details of the analysis & additional data

Details of the analysis

62

Two existential structures • The VP in both structures has a SC headed by Predexist there BE structure

Existential unaccusative structure

63

The existential Pred head: Predexist • Predexist takes three arguments: o o

o

THING: Predexist requires a property (type-shifted) LOC/”here”: a contextually-determined location (Francez 2007) An event: either stative or eventive

• Predexist executes INSTANTIATE (McNally 1992, 1997) o

This establishes a discourse referent

64

Building on McCloskey (2014) • This analysis reformulates McCloskey’s (2014) analysis of Irish existentials • Irish has an existential predicate: ann ‘in it’ 1. Tá fion ann. (McCloskey 2014: 374) is wine in-it There’s wine 2. [[ann]] =

Context dependence

65

Irish existentials (McCloskey 2014) • McCloskey’s analysis 1. Tá fion ann. (p. 374) is wine in-it There’s wine

3.

2. Irish existential SC (p. 349)

66

English adaptation of McCloskey (2014) • Adaptation for there BE sentences o

The denotation of Irish ann ‘in it’ is distributed in the SC (implemented as a PredP)

1. VP for English there be existential

67

VP for existential unaccusatives • English existential unaccusatives require a Path to the contextually-determined Place (=“here”)

68

Two existential structures • The VP in both structures has a SC headed by Predexist there BE structure

Existential unaccusative structure

69

Semantics of Predexist • Existential predicate like Irish ann ‘in it’ o

But in English, it is silent

• Takes a property (type-shifted), a contextuallydetermined location (LOC), and an event 1.

70

Semantics of Predexist • Predexist has the same building blocks as ann ‘in it’ on McCloskey’s analysis 1. • (McCloskey, 2014: 36)

• But Predexist adds an event variable 2.

71

Informal meaning of Predexist 1. A fancy lady waltzed in. • When it has all of its arguments, the PredexistP can be paraphrased informally as follows: • There’s an event of movement along a path, in which a fancy is a participant and such that the lady ends up at a contextually-determined location (‘here’) o

The verb adds: the event is a waltzing event

72

The whole VP, in context • A waltzing event in which a princess is a participant…

73

Final details: Path to Place • In an existential unaccusative, the specifier of Predexist is interpreted as a Path

74

Path to Place • PathP is standardly analyzed as containing a PlaceP (possibly other functional heads) o

Observations going back to Jackendoff (1973, 1983); see Gerke (2008) for discussion

75

Place and Path denotations • We can think of PlaceP as simply the state of having the “at here” property • This is the property that the princess will have

76

Place and Path denotations • Path is a function that requires an individual and an event o

It makes the individual (which will be our princess) a participant in the event of movement along a path

77

Place and Path denotations • It combines with Place through Predicate Conjunction • It needs an individual (the princess) and the name of an event (e.g. waltz)

78

What about ‘here’? 1. A princess waltzed. 2. A princess waltzed in. • (1) is interpreted as happening in a location o

Because every waltzing event happens somewhere

• But only (2) can be interpreted as motion along a path to a contextually-determined location.

79

Contextually-determined location 1. A princess waltzed. 2. A princess waltzed in. • Why is it that (2) can only mean something like in(to) here? • For the same reason that there BE sentences always have a contextually-determined spatiotemporal meaning

80

Contextually-determined location • here ≈ center of discourse o

Reminiscent of Partee & Borschev’s observations about existentials and perspective structure

1. There ran towards me a group of noisy children. 2. #There ran away from me a group of noisy children. Sentences from Stowell (1978), who credits J. Guéron with the observation

81

The Stowell-Guéron Generalization • “There-insertion is possible with verbs of motion only if the motion is directed towards what is perceived as the center of the discourse” (Stowell 1978: 5) 1. There ran towards me a group of noisy children. 2. #There ran away from me a group of noisy children. Sentences from Stowell (1978), who credits J. Guéron with the observation

Additional data Diagnostics for presentation from Guéron (1980)

83

Diagnostics for “presentation” 1. A fancy lady waltzed in. • Further diagnostics: o o o

Negation Tag questions Adverbial modification

84

Negation: Existentials • Can block the establishment of a discourse referent (Karttunen 1974) 1. There was a lady at the door. She was asking for donations. 2. There wasn’t a lady at the door. #She was asking for donations.

85

Sidenote on negated existentials • In Guéron’s discussion of negation, she marks negated existentials as ungrammatical, as in (1) below 1. *There wasn’t a lady at the door. • We would not do so today, but Guéron’s observation holds: (1) sounds acceptable in a very particular context, e.g., with contrastive focus: 2. No, there ISN’T a lady at the door!

86

Negation: Existential unaccusatives • Negation blocks the establishment of a discourse referent in existential unaccusatives 1. A lady didn’t dance in. #She was wearing white. • Negation does not block discourse referents in superficially similar sentences (most likely unergative): 2. A lady didn’t dance gracefully. She was wearing white.

87

Tag questions: Existentials • The new discourse referent established by an existential sentence is not available for pronominal use until after the sentence 1. There’s a salesman at the door, isn’t there? 2. There’s a salesman at the door, *isn’t he/one? 3. There’s a salesman with his shoes untied at the door.

88

Tag questions: existential unaccusatives • Guéron argues that tag questions are unacceptable on all presentational sentences • The following are marginal, in my judgment 1. A cab pulled up, *didn’t it/one? 2. A clown just came over, *didn’t she/it? 3. A little boy darted in, *didn’t he/it?

89

Presentation and adverbial modification • Guéron observes that adverbial modification can make PP extraposition unacceptable (663) o

It blocks a presentational interpretation

1. A man walked in with green eyes. 2. ??A man walked in purposefully with green eyes.

90

Adverbial modification and existential unaccusatives • Not able to test this with there BE sentences • Judgments with existential unaccusatives are not clear-cut 1. A mouse darted in 2. ??A mouse darted quickly in.

91

Adverbial placement • Additional problem: It is not clear where the adverbial should go 1. A mouse darted in. 2. ?A mouse quickly darted in. 3. ?A mouse darted quickly in. 4. ?A mouse darted in quickly.

92

Adverbial placement and PP extraposition 1. A mouse darted in with cute ears. o

PP extraposition allows both readings

2. ?A mouse quickly darted in with cute ears. 3. ??A mouse darted quickly in with cute ears. 4. ??A mouse darted in quickly with cute ears.