semantic structures 2011 henriëtte de swart revisions by joost zwarts

53
Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

Post on 19-Dec-2015

228 views

Category:

Documents


11 download

TRANSCRIPT

Page 1: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

Semantic Structures 2011

Henriëtte de Swartrevisions by Joost Zwarts

Page 2: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

0 Table of contents

0 Introduction

Page 3: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

0 Who is this course for?

Students in the research master in linguistics

Students in the MA CAI. Students in the one-year MA in

linguistics (linguistics, modern languages)

Page 4: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

0 What is this course about? Semantics: empirical knowledge,

theories, research skills, integration in ongoing research

Focus: ongoing NWO programme “Weak referentiality: bare nominals at the interface of lexicon, syntax and semantics” (2008-12).

http://www.hum.uu.nl/medewerkers/b.s.w.lebruyn/weakreferentiality/

Page 5: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

0 What is this course about?

Group project collective teaching different perspectives

General intro (today) (Joost) What are bare nominals? What is

weak referentiality? What are the research questions? Why do we worry about them? What is the approach? What are the results so far?

Page 6: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

0 What are bare nominals? I

Bare nominals are nominal structures that do not have an article or a quantifier.

In English we find lots of bare plurals and bare mass nouns: I read books, I drank milk.

‘Totally’ bare nominals do not have any functional morphology (plurality). Mass nouns are different from count nouns.

Page 7: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

0 What are bare nominals? II

In English, we cannot use bare, singular count nominals in regular argument position: *I read book, I ate apple.

But we find them elsewhere: at school, in hospital, the way to use knife and fork, door after door. WHY?

Page 8: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

0 What are bare nominals? III In other languages, the use of bare

count singular is much more free. WHY?

Wò kànjiàn xióng le. [Chinese] I see bear ASP ‘I saw a bear/some bears.’

dan ra’a namer. [Hebrew] Dan saw tiger ‘Dan saw a tiger.’

Page 9: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

0 Weak referentiality We find bare nominals in English/Dutch

in contexts in which the referential force of the nominal is ‘weak’.

John is in prison. #It is a brick building. Ik weet dat Peter viool speelt. #Kan hii

‘m meenemen? [Dutch] I know that Peter plays violin. #Can he bring it?

Page 10: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

0 Lexical restrictions

John is major of NY/is a lawyer. In prison/at school/at the office. Why does English permit bare

predication only with nouns that somehow have a uniqueness feature?

Why does English permit bare PPs with prison, school, etc. but not office?

Page 11: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

0 Cross-linguistic differences

In prison (E)/en prison (F)/in de gevangenis (D).

In hospital (Br.E.)/in the hospital (Am.E.)/ in het ziekenhuis (D).

At school (E)/ op school (D)/ à l’école (F).

There is overlap in nominal domains, but also differences: where? why?

Page 12: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

0 Weak definites/indefinites

We also find weakly referential nominals that are not bare.

John is a lawyer (cf. Jan is advokaat --Dutch) Mary is listening to the radio (cf. Mary is watching television) How do we understand the def/indef

article in weakly referring contexts?

Page 13: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

0 Back to organization General intro: issues, approach,

organization. Part I: cross-linguistic semantics of bare

nominals (corpus research, offline experiments) (Bert, week 2-3).

Part II: bare prepositional phrases, corpus research and the syntax-semantics interface (Bert & Joost, week 4-5).

Part III: processing weakly referential definites (Ana, week 6-7).

Page 14: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

0 Website

http://www.hum.uu.nl/medewerkers/b.s.w.lebruyn/semstruct2011/

Links to papers, other sources, exercises, results.

Please consult regularly for updates!

Page 15: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

0 Participation

Discussion of reading materials. Workshop on bare nominals. Presentations of research on theme. Final paper: more or less elaborate

research paper (depending on credit).

Page 16: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

0 Languages

What languages do we speak?

Page 17: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

0 Nominal structure: data

Does your language use definite articles?

Does your language use indefinite articles?

Bare plurals? Bare singulars? Please give examples!

Page 18: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

1 Articles: Indefinite

A book, a student: existential quantification. GQ definition:

||a || = PQx[P(x) & Q(x)]

Page 19: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

1 Articles: Indefinite

A child was playing in the park. The funny little creature wore a green hat, and purple socks.

New (in discourse perspective): a P introduces a new discourse referent u and the condition P(u).

Page 20: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

1 Articles: Definite

What is the semantic contribution of a definite article? The sun, the queen of the Netherlands. GQ definition:

||the || = PQx[y[P(y) x=y] & Q(x)]

Uniqueness part is taken to be asserted (Russell) or presupposed (Strawson).

Page 21: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

1 Articles: Definite

A child was playing in the park. The funny little creature wore a green hat, and purple socks.

Familiarity (in discourse perspective): the P introduces a discourse referent v and the condition P(v), and v = u, where u is an accessible discourse referent in the DRS.

Page 22: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

1 Articles: Bare plurals

Existential reading: I bought flowers, unicorns appeared on the horizon.

Generic reading: Cats hate dogs, Cats have four legs. (special semantics needed)

Semantics of existential reading: existential quantification + plurality (sums, sets)/new discourse referent (over sums).

Page 23: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

1 Articles: Bare plurals

Farkas and de Swart (2003): plural morphology presupposes discourse referent accommodation takes care of discourse referential force.

Bare plural with existential reading: similar to singular indefinite, but no article.

Lack of article: where does the existential semantics of bare plurals come from?

Page 24: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

2 Cross-linguistic variation

Puzzle: semantics of definite/indefinite article alike across languages that have such an article.

But not all languages have a definite/indefinite article. Why?

Semantics of bare nominals in a language depends on presence/absence of plural morphology, definite/indefinite article. Why?

Page 25: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

2 Form-meaning mapping

Assume: all humans make the same conceptual disctinctions (atoms vs. sums, old vs. new, uniqueness, …).

Language variation resides in mapping of meanings unto forms.

Approaches: ‘covert’ projections, lexical variation, optimality theory.

Page 26: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

2 Speaker and hearer economy

Languages can choose economy of form (‘bare’ nominals, less elaborate functional morphology). Easy to produce, hard to interpret (ambiguities)

Language can choose elaborate functional morphology to convey uniqueness, newness, etc. Easy to interpret (semantics hardwired into form), hard to produce (formal complexity).

Page 27: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

2 Markedness: economy

Basic markedness constraint: *FunctN.

*FunctN: avoid functional morphology in the nominal domain.

Markedness constraint bars formal complexity preference for bare nominals.

Page 28: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

2 Faithfulness: plurality

Faithfulness constraints encode form-meaning correspondence.

FPl: Plural predication on a discourse referent maps to expression in Num.

Conceptual distinction between atom/sum triggers syntactic reflex (English –s).

Page 29: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

2 Faithfulness: definiteness

Fdef: Uniqueness/familiarity of a discourse referent corresponds with a definite article in D.

Conceptual notion of uniqueness/ familiarity triggers reflex in D (English the).

Page 30: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

2 Faithfulness: reference

Fdr: the presence of a discourse referent in the semantics corresponds with a strong functional layer above NP.

English: plural morphology (-s) or article/quantifier in D (last resort: a).

Page 31: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

2 Ranking constraints

All constraints are universal; ranking is language specific.

Contraints are soft, violable. Ranking determines ‘weight’. Lower ranked constraints can be violated in order to satisfy higher ranked constraints.

Reranking constraints = language typology.

Page 32: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

2 Mandarin Chinese *FunctN >> {FPl, Fdef, Fdr} Wò kànjiàn xióng le.

I see bear ASP ‘I saw a bear/some bears.’

No plural morphology, no definite/ indefinite article: bare nominals are number neutral, but can introduce discourse referents.

Page 33: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

2 Hindi, Georgian, Russian, ..

FPl >> *FunctN >> {Fdef, Fdr} burtebi goravs.

[Georgian] balls.pl.nom roll.3sg ‘Balls/the balls are rolling.’

Plural morphology on the noun, no definite/indefinite article.

Page 34: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

2 Hebrew {FPl, Fdef} >> *FunctN >> Fdr dan ra’a namer.

Dan saw tiger ‘Dan saw a tiger.’

ha-yam-im ‘avru maher. The day.pl pass.past.3pl quickly ‘The days passed quickly.’

Sg/pl morphology, def./bare contrast.

Page 35: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

2 St’átimcets (Salish)

{Fpl, Fdr} >> *FunctN >> FDef Tecwm-mín-lhkan ti púkw-a lhkúnsa.

Buy.appl.1sg.sub det book.det today ‘I bought a/the book today.

Singular/plural morphology on noun, circumfixed determiner for discourse referentiality, but neutral for def/indef.

Page 36: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

2 English, Dutch, Italian, …

{Fdr, Fdef, FPl} >> *FunctN I bought a book/the

book/books/the books. Def/indef contrast, no bare

singulars in regular argument position, bare plurals OK (strong pl).

Page 37: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

2 French

{Fdr, Fpl, Fdef} >> *FunctN J’ai acheté un livre/le livre/des

livres/les livres. I bought a book/the book/indef_pl books/the books.

Def/indef contrast in sg and pl (weak pl morphology).

Page 38: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

2 OT typologyranking features example*FunctN >> {Fpl, Fdef, Fdr}

No number, no articles

Chinese, Japanese

Fpl >> *FunctN >> {Fdr, Fdef}

Sg/pl contrast, no articles

Hindi, Georgian, Russian

{Fpl,Fdef} >> *FunctN >> Fdr

Sg/pl contrast, def/bare contrast

Hebrew

{Fpl, Fdr} >> *FunctN >> Fdef

Sg/pl contrast, no bare nominals (weak Num)

St’átimcets

{Fpl, Fdr, Fdef} >> *FunctN

Def/indef contrast, bare plurals OK

English, Dutch, Italian

{Fpl, Fdr, Fdef} >> *FunctN

Def/indef contrast, no bare nominals

French

Page 39: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

2 Semantics of bare nominals

The semantics of the bare nominal: complement of the marked expression under strong bidirectional optimization.

Hindi/Mandarin bare sg: def/indef Hebrew bare sg/pl: indef (for def is

marked) English bare plural: indef (for def is

marked).

Page 40: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

2 English bare plurals

non-det det

bare (pl)

def (pl)

Page 41: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

2 Emergence of the unmarked

Bare nominal: satisfies *FunctN. Minimal form unmarked. Even in languages in which several

faithfulness constraints outrank *FunctN, we find bare nominal wherever we can.

Emergence of the unmarked

Page 42: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

3 Distribution bare singulars

Ranking *FunctN >> Fdr: bare singulars OK in regular argument position (Mandarin, Hindi, Russian, Hebrew..)

Ranking Fdr >> *FunctN: bare singulars blocked from regular argument position (English, French, St’átimcets,…).

Page 43: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

3 Semantic constraint: Arg

Why do argument positions need marking? Step 1: Argument positions require referentiality

(Arg). Step 2: Referentiality requires marking. (Fdr)

Semantic faithfulness constraint:Arg: parse an XP in argument position as a discourse referent (where X= N, Num or D).

Since Fdr requires discourse referents to be expressed by a strong functional layer, arguments will have marking.

Page 44: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

3 Bare constructions

See the separate handout for bare constructions in English.

What about bare constructions in other languages?

Page 45: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

3 Bare sg escaping Arg

John is in prison. #It is a brick building. Ik weet dat Peter viool speelt. #Kan hii

‘m meenemen? [Dutch] I know that Peter plays violin. #Can he bring it?

Lack of discourse anaphoric binding lack of marking

Not all objects of V and P are arguments.

Page 46: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

3 Extension

Is this true for other environments in which bare nominals occur?

Is it true for other languages? For production experiments on

discourse anaphora see part III

Page 47: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

4 Semantics of bare sg

What do bare singulars mean in ‘weakly referring’ environments?

Lack of discourse referentiality in languages that have a high ranking of Fdr.

Also: pragmatic ‘enrichment’ of the bare nominal.

Page 48: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

4 Bare vs. marked I

John is in jail. John is in the jail. Full PP: location. Bare PP: location + activity sense

(John is a prisoner). Full PP: location – activity sense

(John is in the building, but not as a prisoner)

Page 49: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

4 Bare vs. marked II

Henriëtte is manager. [Dutch] Henriëtte is een manager. Henriëtte is (a) manager. Bare predication: professional

interpretation (‘capacity’ reading). Non-bare predication: general (minus professional reading).

Page 50: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

4 Horn’s division of pragm. labor

Unmarked forms pair up with unmarked meanings, marked forms pair up with marked meanings.

Minimal form preferred: bare nominal is unmarked form.

Stereotypical interpretation preferred: unmarked meaning.

Page 51: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

4 Bare location (weak biOT)

‘prisoner’

‘visitor’

bare PP

def PP

Page 52: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

4 Bare predication (weak biOT)

‘capacity’

‘metaphor’

bare pred

indef pred.

Page 53: Semantic Structures 2011 Henriëtte de Swart revisions by Joost Zwarts

4 Bare vs. marked III

Is this contrast between bare and marked also seen in other constructions?

What about other distinctions between marked and bare (modification, number neutrality, idiomaticity)?