Polysemy, Generality, and Mapping to Syntax Stephen Wechsler University of Texas at Austin Workshop on Syntax-Semantics Interface Institute of Linguistics, Academia Sinica International Taipei, Taiwan June 17, 2011, 4 p.m.

1

The big question Content words exhibit collocational restrictions: they select arguments (*John relies.), determiners (*an evidence), and so on. Some of those restrictions appear to be systematically related to the word’s meaning. Call this relation the syntaxsemantics relation. Question: What is the semantics side of the syntaxsemantics relation? Two hypotheses: Hyp 1. It is the concept of the things in the world to which the word refers. Hyp 2. It is the concept of the word meaning per se.

2

The big question Content words exhibit collocational restrictions: they select arguments (*John relies.), determiners (*an evidence), and so on. Some of those restrictions appear to be systematically related to the word’s meaning. Call this relation the syntaxsemantics relation. Question: What is the semantics side of the syntaxsemantics relation? Two hypotheses: Hyp 1. It is the concept of the things in the world to which the word refers. Hyp 2. It is the concept of the word meaning per se. 

3

Outline of the argument 1. Some property nouns (a flu, a headache) are inherent possessions; others (a coma, a good mood) are locations. 2. Possession and location are semantic types: the distribution is determined by meaning 3. World knowledge of actual flus, comas, etc. does not provide the means to distinguish these two types. 4. But linguistic knowledge does.

4

have + DP: 3 noun types Entity nouns:

car, lollipop, …

Relational nouns:

sister, nose, …

Property nouns:

headache, a flu, the blues, …

1.

Mary has a car.

2.

Mary has a sister.

3.

Mary has a headache.

5

Verbs with an implicit ‘have’

1. John wants the car. ↔ John wants to have the car. 2. John got the car. ↔ John came to have the car. 3. Sue gave Joe the car. ↔ Sue caused Joe to have the car.

6

Verbs with an implicit ‘have’ A durative adverbial can measure the ‘having’ state (McCawley 1974, Ross 1976, Dowty 1979, i.a.): 1. John wants the car for two days.

(want or have for 2 days)

2. John gave me his car for two days.

(have for 2 days)

3. I got John’s car for two days.

(have for 2 days)

Tense: 4. Yesterday John wanted your car tomorrow. (cp. *Yesterday John painted your car tomorrow.)

7

Where does ‘have’ come from? Syntactic hypothesis: An abstract syntactic formative HAVE (McCawley 1974, Ross 1976)

8

Where does ‘have’ come from? Syntactic hypothesis: An abstract syntactic formative HAVE (McCawley 1974, Ross 1976, Richards 2001, Harley 2004)

9

Lexical decomposition (i): ‘have’ in the verb Dowty (1979:269ff): transitive want introduces possession (POSS) substructure. Update: HPSG lexical entry, using Minimal Recursion Semantics:

I want2 a lollipop. ⎡want2 ⎤ ⎢SUBJ < DPi > ⎥ ⎢COMPS < DPj > ⎥ ⎢ ⎥ ⎣CONTENT s: {want(s, i, e), POSS(e, i, j)}⎦ Sublexical adverbial scope (Egg 1999; Beavers et al 2008)

10

Lexical decomposition (ii): ‘have’ in the DP Partee (1999): light verb have in have a sister + Barker (1995): car has possessor arg. in John’s car Beavers et al (2008): both, in John has a car. (Similar for get a car, give John a car, etc.)

11

have + relational DP have as a light verb; relational DP is the main predicate: have: a sister:

λPλy∃x[P(x,y)] sister’

has a sister: λy∃x[sister’(x,y)] John has a sister: ∃x[sister’(x, john)] (Partee 1999, citing Landman & Partee unpublished abstract 1987; Tham 2004, 2005)

12

Barker (1995) analysis of genitive Relational nouns:

(John’s) sister

sister: λxλy[sister(x, y)] Entity nouns: car:

(John’s) car

λy[car(y)]

Possession operator [[poss]] adds the possessor argument: [[poss]]([[car]]) = [[car’]] = λxλy[car (y) ∧ POSS(x,y)]

13

Partee + Barker (Beavers et al 2008) Barker’s possession variant of the entity noun, with Partee’s light verb have: book’: λy∃x[book(x) ∧ POSS(x,y)] • •

(John has a book)

assimilates entity nouns to relational nouns have is always light

Beavers, J., E. Ponvert, and S. Wechsler 2008. ‘Possession of a Controlled Substantive: Light have and Verbs of Possession’. SALT, Amherst.

14

Relational and Non-relational Nouns Take Light Verbs [a book] as possession: a book’:

λy∃x[book(x) ∧ POSS(x,y)]

have:

λRλw[ R(w) ] λRλw[ want(w, R(w)) ] λRλw[ BECOME(w, R(w)) ] λzλRλw[ CAUSE(w, R(z)) ]

want: get: give:

have a book: λw∃x[book(x) ∧ POSS(x,w)] want a book: λw[ want(w, ∃x[book(x) ∧ POSS(x,w)]) ] 15

Possession verbs (Beavers et al 2008) Lexical type for verbs like have, get, transitive want, etc.:

• V selects a possession-type argument. • V identifies one of its arguments with the possessor arg. • relational nouns like sister are lexically possession type

16

Same (light) have with Relational and Entity Nouns John has a sister. John has a condo.

 

Same (light) have: combinations are not zeugmatic: He has [a condo and a rich sister who pays the bills]. He has [big feet and big shoes].

or the following attested example:

17

18

Relational nouns without light have On the Partee / Beavers et al analysis, the main semantic predicate of John has a sister. is sister. The verb have is a light verb. Are there languages or constructions where the light have is not needed, so that sister becomes the syntactic head as well as main semantic predicate?

19

Indonesian middle voice (Udayana in prep)

1.

Dia men-jual mobil itu. he/she AV-sell car that ‘(S)he sold the car.’

2.

Mobil itu ber-jual. car that MV-sell ‘The car sold.’

20

Possessive use of middle (Udayana in prep) 1a. Gadis itu memakai topi girl that AV.wear hat ‘The girl wore a hat’

b. Gadis itu ber-topi girl that MV-hat ‘The girl wore a hat’

2a. Wanita itu mem-punyai suami. woman that AV-has husband ‘The woman has a husband’ b. Wanita itu ber-suami. woman that MV.husband ‘The woman has a husband (i.e. she is married)

21

Modifier stranding adjectives can be stranded: 1. Gadis itu bertopi merah. girl that MV.hat red ‘The girl wore a red hat.’ but not determiners: 2. *Gadis itu bertopi itu. girl that MV.hat that. ‘The girl wore the hat.’

22

Clothing topi baju celana sepatu sandal

‘hat’ ‘shirt’ ‘pants’ ‘shoe’ ‘sandal’

Body parts tangan ‘hand’ kaki ‘leg’ kepala ‘head’ muka ‘face’ mata ‘eye’ kuping ‘ear’

‘wear X’ (‘have X on’) bertopi berbaju bercelana bersepatu bersandal ‘have X’ bertangan berkaki berkepala bermuka bermata berkuping

23

Kinship terms bapak ibu kakak adik paman bibi

‘have an X’ or ‘call somebody X’ ‘father’ berbapak ‘mother beribu ‘older sibling berkakak ‘younger sibling’ beradik ‘uncle’ berpaman ‘aunt’ berbibi

24

English adjectival ‘passives’ from Nouns a bearded man a blue-eyed girl a three-legged stool

The man was bearded. The girl is blue-eyed. The stool is three-legged.

An Arab ambassador in Dubai has had his marriage annulled after discovering that his bride, behind her veil, was bearded and cross-eyed. • inalienable possession: a bearded man, a red-roofed house • clothing: a uniformed commissionaire, a helmeted motorcyclist and a white-coated attendant but: *a wifed man, *a carred man (Hudson 1975) 25

Relational nouns: Entity nouns:

John has a sister. John has a condo.

What about Property nouns? Eliza has…

1. 2. 3. 4.

freedom, faith, … a headache, the/a flu, a cold, … the chills, the giggles, the blues…

I have a headache. I got a headache (from that article). Formal semantics gives me a headache. I don’t want a headache.

Is a headache (etc.) a possession? 26

An English resultative alternation (Grône 2011) He charmed her out of her scruples. He charmed

her scruples out of her. (no charm θ-role)

He charmed her. *He charmed her scruples.

27

He charmed her scruples out of her. ⇒ He charmed her. He washed the soap out of his eyes. ⇒ He washed his eyes. He rubbed the tiredness from his eyes. ⇒ He rubbed his eyes. Levin and Rappaport-Hovav (1995, 66-67): meaning shift from V of contact through motion (wash, rub, etc.) to V of removal. Grône (2011): No. It is more general; and there are unexplained constraints. 28

flu versus coma (Grône 2011) The fever is trying to burn the flu out of him. His mother’s homemade medicine knocked the flu out of him. no hits for: ‘him out of the flu’, ‘him into the flu’, etc. I petted her into a coma. (picture of a very relaxed cat) Father-of-two punches woman into a coma after flying into a rage over a parking space no hits for: ‘a coma out of him’ or ‘a coma into him’.

29

The puzzle 1. a. The fever is trying to burn the flu out of him. b. *The fever is trying to burn him out of the flu. 2. a. *I petted a coma into her. b. I petted her into a coma. (relaxed cat)

30

Grône’s solution a flu is an inherent possession: He has the flu. *He is in the flu. (*He is fluish / *He is influenzal.) The fever is trying to burn the flu out of him. a coma is a location: She is in a coma. (She is comatose.) *She has a coma. I petted her into a coma. 31

Our Hope-coma But, for the rest of the Obama Left wavering in and out of states of consciousness and clarity, if this doesn’t shock them out of their Hope-coma for good, I’m not sure how much more has to happen for them to understand that Obama is a politician and a president.

32

Hope-coma

X • flu

•X

‘X has the flu’

‘X is in a Hope-coma’

33

Hope-coma X • flu

•X

‘burn the flu out of X’

‘shock X out of his Hope-coma’

34

hope-coma versus hope Unlike a (hope-)coma, hope is a possession: She has hope(s). *She is in hope. My successfully married friends innocently squeeze what remains of my marriage hopes out of me by saying things like, "When you know, you just know."

35

scruples: ambiguous? possession:

She has scruples. ⇒ He charmed her scruples out of her.

location:

She is scrupulous. ⇒ He charmed her out of her scruples.

BUT: *She is in scruples. Maybe scruples is not a location; above could be related to: He cheated her out of her money. (n.b. money: not a location) We are out of money. *The money is all out. (The money is all gone.)

36

Summary so far: possessions Entity / Relational / Property nouns as possessions: 1. 2. 3. 4.

I have a car / a sister / a flu. I got a car / a sister / a flu. Luke gave me a car / a bloody nose / the flu. I don’t want a car / a sister / a flu.

Relational & (some) property nouns are inherent possessions. Entity nouns can be type-shifted into possessions.

37

Summary so far: locations Other property nouns are inherent locations: a coma, a (good) mood, a funk 1. 2. 3.

I am in a good mood. Wait for him to get out of that bad mood. A martini should put you into a better mood.

4.

??I have / got / want a good mood.

5.

??Formal semantics gives me a bad mood.

38

Some Possessions & Locations in Mandarin 1a. Wo you ganmao. I have a.cold ‘I have a cold.

b. Wo de le ganmao . I get ASP a.cold. ‘I have/caught a cold.

2a. Mary xianzai xianru hunmi. Mary now sink-into coma ‘Mary is now in a coma / falling into a coma.’ b. Mary chuyu hunmi {zhuangtai / zhong}. Mary be.in coma {condition / within} ‘Mary is in coma.’ (Thanks to Po-Ting Chen and Yahui Huang for data & discussion)

39

Metaphors we live by (Lakoff 1980) Geeraerts (2009, 207) on cognitive semantics and embodiment: Metaphoric uses of the containment image schema occur when someone enters a depression, to take an example: the abstract emotional condition is seen as a container restricting the person’s behavior.

40

These ‘metaphors’ are: • extended: for a given noun, the ‘metaphor’ is semantically regular across different structures and collocations • lexical: the semantic category varies from noun to noun I entered a depression %??I have a depression.

(location)

I get the blues most every night. *I enter the blues most every night.

(possession)

I am in a funk. I entered a deep funk. *I have a funk.

(location)

41

Proposed conceptual structures Proposal: property nouns denote concepts with some internal structure (i.e. they are typed): a flu, N

λx∃y[POSS(x,y) ∧ flu(y)]

(a possession; x is the possessor of the flu) a coma, N

λLλx∃y[L(x, y) ∧ coma(y)]

(a location; x is the locatum, L the locative relation) in a coma

λx∃y[in(x, y) ∧ coma(y)]

42

Alternative: Contructional encoding of POSS? Harley (2004): • An abstract (silent) preposition PHAVE. • (Silent) BE adds durative aspect. • BECOME yields punctual aspect. spell-outs: have (the word) = BE + PHAVE get (the word) = BECOME + PHAVE

43

Alternative: Contructional encoding of POSS? Syntactic hypothesis: An abstract syntactic formative HAVE (McCawley 1974, Ross 1976, Richards 2001, Harley 2004)

44

Alternative: phrasal idioms with abstract HAVE?

PHAVE + the blues = ‘be sad’ PHAVE + the flu = ‘be sick with influenza’ …etc.

That would account for the parallel across have, get, give, etc. (cp. Richards 2001, Harley 2004)

spell-outs: BECOME + PHAVE + the blues = get the blues BECOME + PHAVE + the flu = get the flu But the lexical decomposition account allows a stronger theory…

45

Real words versus sublexical semantic formatives On the lexical decomposition account: • Real words (e.g. have) can select other words and idiosyncratic types, with special meanings (idioms, multiword expressions, ‘special selection’, contextual polysemy)

• Sublexical formatives (e.g. POSS) can have simple

polysemy, but cannot for idioms, etc. with neighboring words

46

Simple polysemy of POSS: have (Tham 2005) 1. Relational DP: John has a sister. 2. Alienable Possession: John has a car. 3. Discourse-Determined (“Focus”) have: Q: What will Eliza be polishing? A: She has the mirrors. (= ‘She will be polishing the mirrors.’) 4. Control have (with definite object; sentient control): Q: Where are the mirrors? A: Eliza has them. (cp. #The bathroom has them.) 47

want, give, get: parallel range of meanings 1. John wanted / got / has a sister. 2. John wants / got / has a car. 3. Q: What will Eliza be polishing? A: She wants / gets / has the mirrors. 4. Q: What will happen to the mirrors? A: Eliza wants / will get / has them. Beavers et al (2008): POSS is 4-ways ‘polysemous’ (multi-functional): simple polysemy

48

Prediction of the lexical decomposition account If POSS is sublexical, we predict that, while it can exhibit simple polysemy, it cannot form a phrasal idiom (multiword expression; idiosyncratic collocation; etc.) with nearby words. The POSS in Mary wants / had / got a baby—

baby, N

λx∃y[POSS(x,y) ∧ baby(y)]

—is encapsulated in the word and thus cannot form an idiom with other words in its environment.

49

having a baby Natalie is having a baby.

(‘giving birth to a baby’)

have a baby: ‘give birth to a baby’ / ‘possess (etc.) a baby’ Lexicalist prediction: want a baby ≠ ‘want to give birth’: Natalie wants a/the baby: • relational: ‘wants to be the mother of a baby’ • alienable: ‘wants possession of some baby’ • discourse: ‘wants [some contextual relation to] a baby’ • control: ‘wants to (e.g.) hold the baby’ But crucially NOT ‘want to give birth to a baby’ 50

having a baby have a baby: ‘give birth to a baby’ / ‘possess (etc.) a baby’ want a baby: ‘want to possess (etc.) a baby’ (NOT ‘want to give birth to a baby’) 1. Natalie doesn’t want to have a baby, so she is going to adopt one. 2. #Natalie doesn’t want a baby, so she is going to adopt one. => have a baby is ambiguous => want a baby is general

51

More have-idioms 1. I had a ball. (‘enjoyed myself’) *I wanted a ball.

(McCawley 1974)

2. I had it out with Fred. (‘argued angrily’) *I want it out with Fred. (McCawley 1974) 3. C’mon, have a heart and give my kid an A. (‘be compassionate’) *I don’t want a heart, and besides, he flunked the exam. 4. The okra is ready. Go ahead, have at it! (‘do something heartily’) *But I don’t want at it! Yuck!

52

5. I’ve been had! (‘cheated’) (*)I’ve been wanted! (‘someone wanted to cheat me’) 6. He had it away with her. (had casual sex with’; Brit.) *He wanted it away with her. 7. I’ll ‘ave you! (‘exact revenge on you’; Brit.) (*)I want you! 8. Don’t have a cow, man! (Bart Simpson) (*)What if I want a cow? Mini-summary: • there are some idioms containing have (the word) • there are no idioms containing POSS

53

Another problem for the constructional hypothesis:

VP V

|

XP

want PRO

Xʹ′

XHAVE

DP

a lollipop No evidence for this syntactic structure

54

DP (not XP) structure is supported by: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

History of want cognate language passive adjective formation ECM coordination of complements infinitival relatives adverb intervention modification by right particle shift coordination of verbs

Wechsler, Stephen (2008). ‘Dualist Syntax’. Proceedings of the 15th International Conference on HPSG, CSLI On-line Publications.

55

Conclusion so far Property nouns belong to semantic categories that vary from noun to noun. Suppose word meanings are concepts (Murphy 2002, inter alia). Then: [a flu]:

the ‘flu’ concept is of the type possession

[a coma]:

the ‘coma’ concept is of the type location

Question: Are these typed concepts part of the ‘grammar’? Proposed answer: Yes. 56

Borer’s Dichotomy

• Content words (‘listemes’) such as nouns, verbs, and adjectives have no grammatical features. • Functional morphemes (‘f-morphs’) such as plural inflection, determiners, numerals, and classifiers do have features. Hagit Borer (2005). Structuring Sense, Oxford University Press.

57

Borer (2005, vol. I, p. 11): [There is] no direct interface between the conceptual system and the grammar, in that properties of concepts do not feed directly into any determination of grammatical properties. A substantive listeme is a unit of the conceptual system, however organized and conceived, and its meaning, part of an intricate web of layers, never directly interfaces with the computational system.

58

Borer (2005:10-11) distinguishes between • ‘what is grammatically real—structures and formal properties of functional items’ • ‘what may be very real, but not grammatically so— properties of substantive vocabulary. … creatures born of perception and conceptualization, representing an intricate web of layers upon layers of a complex perceptual structure and emerging world knowledge…’

59

Suppose we accept Borer’s distinction between ‘f-morph grammar’ and the conceptual system. Then (*) Mary is in a flu. is not ungrammatical; rather, it fails to make sense because in requires a location, but flu is not one. TO SHOW: Whether or not we choose to call it part of ‘grammar’, that conceptual system for combining listemedenoted concepts is decidedly linguistic; it must be distinguished from (the rest of) the ‘intricate web of layers upon layers of a complex perceptual structure and emerging world knowledge.’

60

Claim: Paula’s knowledge of the meaning of the word dog is NOT the same as Paula’s world knowledge of dogs

‘dog’

≠

61

What is left • Evidence for the dissociation between linguistic meaning and concepts: o reasons for expecting it, in principle o experimental evidence. • Apply this to 1st and 2nd person pronouns. • Apply this to property nouns like flu.

62

Norms of Communication ‘All specific linguistic norms are justified relative to the highest norm of communication, which is: ‘Express yourself in such a way that what you say is recognizable and interpretable by your partner in agreement with what you intend him to understand.’ And, correspondingly, for the hearer it is: ‘Interpret such that the interpretation will be in agreement with what the speaker intends.’ Bartsch (1987:212) Cp. Grice’s Cooperative Principle. Bartsch, Renate 1987. Norms of language: theoretical and practical aspects. London: Longman.

63

Functional pressure to conventionalize Strategy speakers and addressees should adopt, when using the word dog: • focus on mutual knowledge of dogs (Tomasello 2008) • disregard private knowledge, including perceptual knowledge and history.

64

Experimental evidence from Referential Pacts lexical entrainment: when people in conversation refer repeatedly to the same object, they come to use the same terms. They are said to have implicitly agreed to a referential pact.

Brennan, S. E, and H. H Clark. 1996. Conceptual pacts and lexical choice in conversation. Journal of Experimental Psychology: Learning, Memory, and Cognition 22, no. 6: 1482.

65

Referring to objects Procedure: Confederate speaker makes repeated references to objects. Then either the original speaker or a new speaker used either the original expression or a new one for the object. Result: Hearers take longer to interpret a word when the same speaker changes to a new word; but not so if a new speaker uses a new word. Metzing, C. & Brennan, S. E. (2003). When conceptual pacts are broken: Partner-specific effects in the comprehension of referring expressions. Journal of Memory and Language, 49, 201–213.

66

Original partner: 698 msec longer (with new word) New partner: 70 msec longer (with new word) Metzing & Brennan (2003:208) Metzing, C. & Brennan, S. E. (2003). When conceptual pacts are broken: Partner-specific effects in the comprehension of referring expressions. Journal of Memory and Language, 49, 201–213.

67

Studies of categorization by children • Children form categories on the basis of similarities they detect themselves (for a review, see Rakison & Oakes, 2003). • Children also do so from verbal testimony (Jaswal and Markman 2007) • Categories formed from these two types of source can be

dissociated (Jaswal and Markman 2007).

68

Conclusions: Language users pay close attention to each other’s usage and adjust their own usage and interpretation accordingly. Hypothesis: word meanings are established and conventionalized by this process. Such word meanings, being subject to normative pressures, are not equivalent to our private concepts of the things to which the word refers.

69

Incorrect view of interpretation

Dog!

Mary

The word evokes Paula’s dog-concept

70

Incorrect view of interpretation

Dog!

Mary

The word evokes Paula’s dog-concept

71

Correct view of interpretation what Mary means by ‘dog’

Dog!

Mary

Paula figures out what Mary intends

72

Example 1: 1st and 2nd person pronouns I= ‘current speaker’ I like you

Mary

Paula

I = ‘current speaker’ ?? you = ‘current hearer’ ??

73

Example: 1st and 2nd person pronouns I= ‘current speaker’ I like you

Mary I = ‘current speaker’ ?? you = ‘current hearer’ ??

Paula No. No.

74

myself Mary is selfrefering

…I…

Mary Rules:

Paula I is used by speakers to self-refer you is used by hearers to self-refer

The De Se Theory (Wechsler, S. 2010. ‘What you and I mean to each other’, Language 86.2.)

75

Paula is selfreferring myself

…you…

Mary Rules:

Paula I is used by speakers to self-refer you is used by hearers to self-refer

76

Evidence for the De Se Theory 1. The Associative Plural Generalization 2. 2nd person with multiple addressees

77

1st and 2nd person plural: associative we = ‘speaker + speaker’ ‘speaker + other’

We are the champions! We want you to join us.

‘speaker + addr.’ ‘speaker + addr. + other’

Shall we go? Can’t we all get along?

you.PL = ‘addr. + addr.’ ‘addr. + other’

You should behave yourselves. When did y’all move here?

78

Regular plural versus associative plural 1st/2nd pers. pronouns are associative: 1PL: ‘Some member of reference set is the speaker’ 2PL: ‘Some member of reference set is the hearer’

regular plural NP: ‘Every member of reference set is…’ the dogs: ‘Every member of the reference set is a dog.’

79

The Associative Plural Generalization A universal: 1pl and 2pl forms have associative semantics, not regular plural semantics.

These are ‘absolute universals rather than strong trends’ (Bobaljik 2008, 209) (Bobaljik, 2008; Cysouw, 2003; Greenberg, 1988; McGregor, 1989; Moravcsik, 1978; Noyer, 1992; Silverstein 1976) 80

The Associative Plural Generalization A universal: 1pl and 2pl forms have associative semantics, not regular plural semantics.

This follows from the de se theory: 1st and 2nd person pronouns are for self-reference by speaker/hearer, not reference to speaker/hearer. So they cannot be restricted to referring to ‘only speakers’ or ‘only hearers’.

81

2nd sg, multiple addressees Teacher to her class: Write your name on the page. Each addressee x knows to write x’s name: Tom writes Tom’s name; Mary writes Mary’s name. Predicted by de se theory: 2nd Person indicates self-reference by each addressee, not reference to ‘the current addressee’!

82

Example 2: possession vs. location nouns 1. Sue can’t come to your talk. She has a flu. ⇒ flus are possessions, so ‘she’ is interpreted as the patient with the flu. 2. Sue can’t come to your talk. (#)She has a coma. ⇒ comas are not possessions; have must be Discourse-have; e.g. Sue is a doctor and must observe a coma, etc. How do language users classify those two concepts? • via world knowledge of actual flus and comas? • by considering the normative concepts associated with the words flu and coma?

83

Categorize ‘flu’ as possession: by world knowledge?

…flu…

Mary Paula

84

Categorize ‘coma’ as location: by world knowledge?

…coma…

Mary

85

World knowledge fails

type = possession? = location?

?

type = possession? = location? Paula

86

When world knowledge really fails ‘the blues’

‘a funk’

a possession

a location

87

Categorize ‘flu’ as possession: by linguistic knowledge

what Mary means by ‘flu’

…flu…

Mary Paula

88

Categorize ‘coma’ as location: by linguistic knowledge

what Mary means by ‘coma’

…coma…

Mary Paula 89

Linguistic knowledge works …have a flu…got a flu…

aha!

flu: type = possession …in a coma…out of a coma…

coma: type = location

90

How do language users classify those two concepts as possessions versus locations? • Via world knowledge of actual flus and comas? No, these lack the requisite structure to distinguish possessions from locations. • By considering the normative concepts associated with the words flu and coma, as evidenced by previous exposure to those words in context? Yes.

91

Conclusion: the big question Content words exhibit collocational restrictions: they select arguments (*John relies.), determiners (*an evidence), and so on. Some of those restrictions appear to be systematically related to the word’s meaning. Call this relation the syntaxsemantics relation. Question: What is the semantics side of the syntaxsemantics relation? Two hypotheses: Hyp 1. It is the concept of the things in the world to which the word refers. Hyp 2. It is the concept of the word meaning per se. 

92

Thank-you!

93

References Barker, Chris. 1995. Possessive Descriptions. Stanford, CA: CSLI Publications. Beavers, John, Elias Ponvert, and Stephen Wechsler. 2008. Possession of a Controlled Substantive: Light have and Verbs of Possession. In SALT, Amherst. Borer, Hagit. 2005. Structuring Sense, Volume I: In Name Only. Oxford: Oxford University Press. Geeraerts, D. 2009. Theories of lexical semantics. Oxford University Press, USA. Grône, Maryse. 2011. La productivité des tournures résultatives du type he charmed the scruples out of her / I’ll scare some sense into you en anglais. presented at the Du discours au système : variation et changements linguistiques autour du verbe, Institut d’Etudes Anglophones de Paris 3 Sorbonne Nouvelle, June 20, 2011.

94

Harley, H. 2004. Wanting, Having, and Getting: A Note on Fodor and Lepore 1998. Linguistic Inquiry 35, no. 2: 255-267. Hudson, R. A. 1975. Problems in the analysis of ed-adjectives. Journal of Linguistics 11, no. 01: 69-72. Jaswal, V. K, and E. M Markman. 2007. Looks aren't everything: 24month-olds' willingness to accept unexpected labels. Journal of Cognition and Development 8, no. 1: 93-111. Levin, Beth, and Malka Rappaport-Hovav. 1995. Unaccusativity: At the Syntax-Lexical Semantics Interface. Cambridge, Massachusetts: MIT Press. Metzing, C., and S. E Brennan. 2003. When conceptual pacts are broken: Partner-specific effects on the comprehension of referring expressions. Journal of Memory and Language 49, no. 2: 201-213. Murphy, Gregory L. 2002. The Big Book of Concepts. Cambridge: MIT Press. Partee, B. H. 1999. Weak NP’s in HAVE sentences. In Essays Dedicated to Johan van Benthem on the Occasion of his 50th

95

Birthday, ed. J. Gerbrandy, M. Marx, M. de Rijke, and Y. Venema. Amsterdam: Amsterdam University Press. Wechsler, Stephen. 2008. Dualist Syntax. In Proceedings of the 15th International Conference on Head-Driven Phrase Structure Grammar. Keihanna, Japan: CSLI On-line publications.

96