ppt

advertisement
Telicity features of bare nominals
Henriëtte de Swart
Berlin, Dec 2010
Bare plurals and telicity






Mary ate an/the apple in/*for an hour. [telic]
Mary ate apples for/*in an hour.
[atelic]
Mary ate the apples in/*for an hour.
It took Mary an hour to eat an apple/*apples.
He continued to eat #an apple/#the apple/apples.
English bare plurals lead to atelicity (unbounded
process), most other nominal arguments to
telicity (event with inherent endpoint).
Iterative durativity/bare habituality





John found #a flea/fleas on his dog
for a week.
John repairs #a bicycle/bicycles.
Every day, John repairs a bicycle/bicycles.
Sg indefinite does not allow multiple event
reading, even if one object is involved per event;
no bare habituality.
Sg indef OK under quantifier scope.
Aspectual composition






Semantics of nominal argument determines
aspectual nature of VP (S).
Verkuyl (1972/1993): [±SQA] feature on NPs
Krifka (1989): quantized/non-quantized objects.
Mapping objects
events/path structure.
Quantized object maps onto quantized event/
bounded path (Mary ate an apple)
Cumulative object maps onto cumulative event/
unbounded path (Mary ate apples)
Iterative durativity



With count noun interpretations, cumulative
reference requires plurality (Scha 1984).
Van Geenhoven (2004, 2005): pluractionality
explains combination of accomplishment/
achievement with for-adverbial: bare plural
distributes internal argument over events.
De Swart (2006) on bare habituality: bare plural
behaves like dependent plural on set of events.
Inherent telicity
The dog ate up a/the cake
that I baked for the party.
 The dog ate up the cakes/
*cakes I baked for the party.
He drank up (all) the water/*water in the tap.
Particle verb inherently telic: mapping from
object to event requires object to be
quantized  incompatible with bare
plural/mass noun.



Cross-linguistic support (Italian)



Ha stirato molte camicie in due ore /
*per due ore di seguito.
He ironed many shirts in two
hours/*for two hours.
Ha stirato camicie *in due ore / per
due ore di seguito.
He ironed shirts *in two hours/for
two hours.
Dobrovie-Sorin & Laca (2003).
Broadening our view



Do bare plurals in all languages lead to
atelicity? If so, why? If not, why not?
What about bare singular (count) nominals (in
languages in which they occur)? Predictions
about telicity?
If we want to investigate the telicity features of
bare nominals, where do we start?
Bare nominal semantics



BN: nominal without a determiner ~ no info
about quantity, discourse reference.
Intuition: bare nominals convey (covertly) what
is not expressed (overtly) by determiners
(cf. Chierchia 1998, blocking).
What features of the language come into play in
determining the aspectual nature of
configurations with bare nominals?
A typology of bare nominals




Cross-linguistic variation in the semantics of
bare nominals correlates with variation in
number marking and article use.
Number: sg/pl distinction leads to BS/BPl
distinction ~ investigate number neutrality.
Article use: definite/indefinite article blocks
definite interpretation/discourse reference.
De Swart & Zwarts (2009, 2010): OT typology.
OT typology of number/articles




*FunctN: Avoid functional structure in the
nominal domain (markedness constraint).
FPl: Parse sum reference in the functional
projection of the nominal (faithfulness constr.)
FDef: Parse dynamic uniqueness by means of a
functional layer above NP.
Fdr: Parse a discourse referent by
means of a functional layer above
NP.
No sg/pl, no articles: Mand. Chinese
*FunctN >> {faith constraints number, articles}
 Wò kànjiàn xióng le.
[Mandarin Chinese]
I
see
bear Asp
‘I saw a bear/bears.’
 Gou juezhong le.
Dog extinct Asp.
‘Dogs are extinct.’
 Gou hen jiling.
Dog very smart.
‘The dog/dogs are intelligent.’

Induced telicity in Mandarin



Wo he-guan le tang.
I drink-up asp soup
‘I drank up the soup/*soup.’
Wo mai-zhao le shu.
I
buy-get asp book
I
managed to buy the books/*books.’
Sybesma (1999): RV construction requires
definite/specific interpretation of bare nominal.
Telicity features of Mandarin BN


BNn:  quantized (‘indef’, ‘specific’, ‘definite’),
 cumulative (‘unbounded plurality’)
No blocking of form/meaning combination:
telic/atelic interpretation for number neutral BN.
Sg/pl distinction, no article: Slavic



FPl >> *FunctN >> {faithfulness constraints
definiteness/discourse reference}
On ot-krylperf
okno.
[Russian]
he open.past.perf window.acc
‘He opened (the/a) window.’
Petja čitalimp
stat’i/literaturu
Peter read-imp-past.sg. articles/literature-acc
‘Peter was reading articles/the articles/
literature/the literature/read articles/literature.’
BS in Slavic semantically singular
BSs in Slavic languages
have atomic reference:
complement of BPl
BS
under bidirectional
optimization (Farkas &
de Swart 2010).
at
sum


BPl 



Bare habituality with BPl


Cumulativity of count noun depends on plurality
(Scha 1984) ~ no cumulative interpretation for BSs.
Petja čitaet
lekcii v universitete [Russian]
Peter read-IMP-pres lectures in university

‘Peter gives lectures (is a lecturer) at the university
Petja zavtra čitaet
lekciju v universitete
Peter tomorrow read-IMP-pres.3sg lecture in university
‘Tomorrow, Peter is giving (will give) a lecture at
the university’
Borik (2002: 140).
BPl definite/indefinite in Slavic


Petja pro-čital
stat’i/literaturu
Peter perf-read-past.sg articles/literature-acc
‘Peter read the articles/the literature’
No definite article, no competition: BPl
underspecified ~ adapts under contextual
pressure to define inherent endpoint by
taking up definite/specific interpretation: Filip
(1999), Piñón (2001), Gehrke (2008),..
Perfectivity induces telicity




Piñón (2001): Perfective prefix requires
quantized (not cumulative) object.
Czytaći: Imp(Read) = yxe [Read(e,x,y)]
Prze-czytaćp: Perf(Imp(Read)) =
PQe[Q(e,xe’[P(e’, xe” [Read(e”,s,y)])])
 x[CUM(Q(xe’[Read(e’,x,y)]))]
 y[CUM(P(xe’[Read(e’,x,y)]))]]
PQ[CUM(Perf-Imp-Read(P )(Q))]
Slavic BS/BPl and telicity


BSs:  quantized (‘indef’, ‘specific’, ‘definite’)
 cumulative
BPl:  quantized (‘specific’, ‘definite’)
 cumulative (‘unbounded plural’)
Definite article (Hebrew)




{FPl, Fdef} >> *FunctN >> Fdr
ra’iti kelev. hu navax/ #hem navxu
I-saw dog. It barked/ #they barked
‘I saw a dog. It barked/ #they barked.’
novxim klavin.
Bark dogs ‘Dogs are barking.’
Doron (2003). Strong contrast sg/pl ~ BS has
atomic reference: BSs. Fully discourse
referential. Restricted to indefinite
interpretation under bidirectional optimization.
BS in Hebrew semantically indefinite
Blocking by DefSg
restricts BSs in
Hebrew to indefinite
interpretation.
Idem for BPl
(non-definite only)
BS





DefSg 


Telicity features of Hebrew BS/BPl




hu kara sefer be-ša’a/ be-mešex ša’a
he read book in-hour/ for
hour
‘He read a book in an hour/for an hour.’
(weak telicity features, no cumulative reading)
hu nipeax balonim bemešex šaa
he blew balloons for an hour
hu nipeax et ha-balonim tox 5 dakot.
he blew acc the balloons in 5 minutes
Cabredo Hoffher (2009), Yitzhaki (2003)
No iterative durativity for Hebrew BS




Lack of plurality blocks iterative durativity/bare
habituality of Hebrew BSs
John me’ašen sigariya
John smokes cigarette
 John is smoking a cigarette (episodic)
 John smokes cigarettes (habitual)
John me’ašen sigariyot
John smokes cigarettes
 John smokes cigarettes (habitual)
Cabredo Hoffher (2009), Yoad Winter (p.c.)
Telicity features of Hebrew BS/BPl


BSs:  quantized (‘indefinite’)
 cumulative
BPl:  quantized (‘specific’, ‘definite’)
 cumulative (‘unbounded plural’)
Def/indef article (Romance, Hungarian)



{Fpl, Fdef, Fdr} >> *FunctN
Morphological sg/pl contrast, def/indef sg, and
bare/indef plural (depending on discourse role
plural morphology, cf. Farkas & de Swart 2003).
Strong contrast BS  everything else: BS does
not satisfy Fdr ~ restricted to constructions with
‘weak’ discourse referentiality features: object
position of ‘have’ verbs, bare predication, bare
coordination, bare PPS..
Number neutrality of BS




Busco pis. Un a Barcelona i un a Girona. [Catalan]
look.for-1sg appartment. One in B. and one in G.
‘I’m looking for an apartment. One in Barcelona
and one in Girona.’
Espinal & Mcnally (2010)
Mari belyeget gujt.
[Hungarian]
Mari stamp-acc collect ‘Mari collects stamps.’
BS in Romance/Hungarian number neutral: BSn.
Farkas & de Swart (2003): number defined for
discourse referents, not for thematic arguments
(DRT). Weak referentiality ~ number neutrality.
Bare singulars with ‘have’ verbs




Spanish, Catalan, Romanian: fairly liberal use of
bare singulars in object position of ‘have’ verbs,
cf. Dobrovie-Sorin, Bleam & Espinal (2006),
Espinal & McNally (2010).
Lleva sombrero. [Sp] / Porta barret. [Catalan]
wears hat
wears hat
‘(S)he wears a hat.’
Ion are casă
[Romanian]
Ion has house. ‘Ion has a house.’
But: mostly stative verbs  no telicity effects.
Accomplishment verbs: telicity



Encontraron aparcamento (en diez minutos) [Sp]
Found
parking
(in ten minutes)
‘They found a parking place in ten minutes
Espinal (2009): there could be more than one
parking place if more than one driver (NN).
Telic interpretation of bare nominal possible, at
least with certain verbs. Espinal (p.c.): BSn must
be aspectually inert (property interpretation).
Collectivity vs. iteration in H.



Ma delutan szaraz levelet szedtem ossze a haz korul.
This afternoon dry leaf gathered together the house around
‘This afternoon, I gathered dry leaves around the house.’
Ma delutan szaraz leveleket szedtem ossze egy-es-è-vel
This afternoon dry leaves gathered together one-by-one
a haz korul
[Hungarian]
the house around
‘This afternoon, I gathered dry leaves one by one around
the house.’
Number neutrality in object position ‘collect’ verbs, but no
iterative durativity. Dayal (2009).
No iterative durativity in H


János (*egy hétig) bolhát talált a utyáján.
John(*one week-till) flea.acc found the dog-3sg-on.
John found some fleas on his dog (on one
occasion).
[Hungarian]
Not: John found fleas on his dog for a week
(iterative durative reading), Bende-Farkas (2001).
Number neutrality in Romance/Hungarian does not
lead to atelicity via plurality (no cumulativity).
Telicity features of BS/BPl in
Romance/Hungarian




BSn:  quantized (‘indefinite’, ‘definite’)
 cumulative (‘unbounded plurality’)
BPl:  quantized (‘specific’, ‘definite’)
 cumulative (‘unbounded plural’)
Def/indef and sg/pl contrast do not apply to
non-referential arguments (require dr).
Cumulative BSn requires (dr) plurality for event
distributivity: not available for BSn in Romance/
Hungarian.
Recap: role of number in telicity


*FunctN >> FPl or FPl 0 *FunctN leads to
number neutrality ~ BSn  cumulative 
atelic, iterative durativity/bare habituality
(Mandarin Chinese, Hindi, Braz. Portuguese)
FPl >> *FunctN leads to atomic reference for
BSs ~  cumulative  telic, no iterative
durativity/bare habituality (Slavic, Hebrew).
Recap: role of definite article


*FunctN >> Fdef makes definite/specific
interpretations available for both BS and BPl ~
 quantized  telic interpretations available
with BS and BPl (Mandarin Chinese, Hindi,
Slavic).
Fdef >> *FunctN restricts BS/BPl to indefinite
interpretation ~ BPl  quantized  atelic
interpretation only for BPl (Hebrew, Brazilian
Portuguese).
Recap: role of indef. article


In Brazilian Portuguese, Papiamentu indefinite
sg competes with BSn ~ BSn  quantized 
atelic interpretation only, iterative durativity/ bare
habituality OK. Why?
Fdr >> *FunctN: BS restricted to non-referential
position, number and definiteness irrelevant, but
no asserted plurality. BSn  cumulative  quasi
telic interpretation verb driven, no iterative
durativity/bare habituality (Romance,
Hungarian).
Project Info


Weak referentiality: bare nominals at the
interface of lexicon, syntax and semantics
(2008-2012).
http://www.hum.uu.nl/medewerkers/b.s.w.lebr
uyn/weakreferentiality/index.htm
Download