Definiten esssplitsarchive.sfl.cnrs.fr/sites/sfl/IMG/pdf/lsalaa2012Loebner.pdf · x is a snowman, x is photographable” (= anaphor context) 1. Observations 2. Concept types 3. Uses

Definiten esssplits

Sebastian Löbner

Heinrich Heine University Düsseldorf, Collaborative Research Centre 991 “The Structure of Representations in Language, Cognition and Science”

www.sfb991.uni-duesseldorf.de/en/sfb991/

Languages with and without articles Paris, 15-16 March, 2012 UMR 7023 CNRS / Paris 8

1. Observations

2. Concept Types

3. Uses

4. The scale

5. Splits

1. Observations 2. Concept types 3. Uses 4. The scale 5. Splits 3

1. Observations about definite descriptions

For languages with definiteness marking:

! " Semantic theory is preoccupied with anaphoric uses of definites.

! " There are certain conceptual types of nouns for which the definite article is (almost) obligatory ! and these types of definites are not anaphoric.

! " In almost all languages there are splits of marking definite NPs ! one split separates the domain of demonstratives from other definite NPs, another one separates morphologically marked definites from unmarked ones.

! " In most cases, definite articles developed from demonstratives.


2. Concept types

[ © stands for: “in need of support by special context” ! indicates: “felicitous without particular contextual support“]

individual concepts

(1) !The/©A pope will visit Switzerland in 2016.

(2) By 2030, the catholic church will have !a/*the different pope.

sortal concepts

(3) !A/©The cat killed !a/©the mouse.

(4) © Our cat caught a mouse yesterday. She killed !the/©a mouse.

functional concepts

(5) !The/©A mother of Jeanne consulted the teacher.

(6) Every person has !a/*the mother.

1. Observations 2. Concept types 3. Uses 4. The scale 5. Splits

5

Conceptual noun types and their respective unmarked determinations

[–U] [+U] conceptually unique

sortal nouns

girl book water

! indefinite

! absolute

logical type: <e,t>

individual nouns

pope; Jeanne; she

! definite

! absolute

logical type: <e>

[–R]

relational nouns

daughter part kin

! indefinite

! possessive

logical type: <e,<e,t>>

functional nouns

mother mouth amount

! definite

! possessive

logical type: <e,e>

[+R]

conceptually

relational


6

Conceptual noun types ! definite and possessive determination


sortal nouns

girl book water

© definite

© possessive

logical type: <e,t>

individual nouns

pope; Jeanne; she

! definite

© possessive

logical type: <e>

[–R]

relational nouns

daughter part kin

© definite

! possessive


functional nouns

mother mouth amount

! definite

! possessive

logical type: <e,e>

[+R]

conceptually

relational


7

Conceptual noun types ! indefinite and absolute determination


sortal nouns

girl book water

! indefinite

! absolute

logical type: <e,t>

individual nouns

pope; Jeanne; she

© indefinite

! absolute

logical type: <e>

[–R]

relational nouns

daughter part kin

! indefinite

© absolute


functional nouns

mother mouth amount

© indefinite

© absolute

logical type: <e,e>

[+R]

conceptually

relational


Concept types and determination

! " The conceptual type of a noun or pronoun is lexically fixed (modulo polysemy): The meaning of a sortal/relational/individual/functional [pro]noun is a sortal/relational/individual/functional concept.

! " However:

When an NP is formed, the noun concept may undergo conceptual shifts,

!" (overtly) by being combined with modifiers or arguments

!" (covertly) by undergoing a general meaning shift (e.g. metonymy)

!" (covertly) by being merged with contextual information on its referent

! " Simple determination ( = definite / indefinite / possessive / absolute without further semantic content) fixes the conceptual type of the NP token. Determination may coerce a type shift of the CNP (common noun phrase = operand of determination).


Definite determination

! " Definite determination means: “Construe the NP token as a conceptually unique description, i.e. as [+U] ! ”.

- The meaning/function of definite determination is the same for singular, plural, and mass CNPs

! Definiteness is neither a matter of extensional/accidental uniqueness nor of familiarity.

Indefinite determination

! " Indefinite determination means: “Construe the NP token as a sortal description, i.e. as [–U] ! ”.

- The meaning/function of indefinite determination is the same for singular, plural, and mass CNPs.

[Possessive and absolute determination will be disregarded in this talk, but see Löbner 2011.]


Congruency and type shifts

! " If the CNP is [–U], definite determination coerces a type shift [–U] " [+U]

> In particular, definite determination coerces a type shift on sortal CNPs

for anaphoric and deictic DDs

! " If the CNP is [+U], indefinite determination coerces a type shift [+U] " [–U]

> cf. indefinite uses of individual or functional concepts

! " Determination is (in)congruent iffdef the CNP is (not) of the resulting type.

! " A NP is semantically definite iffdef the CNP is [+U]. A DD is pragmatically definite iffdef the CNP is [–U].


Levels of type shifts

Level 0 a. choice of a lexical meaning variant

b. compositional modification: attributes, complements, adjuncts

core semantics

Level 1 general conceptual shifts applying across semantic subclasses of meanings (such as “artefact“, “institution“, “profession“, “attribute“, “property“)

dynamic lexicon

Level 2 merging the concept for the referent of an NP with extralinguistic information

pragmatic enrichment


3. Uses of definites

! " Congruent definite determination: individual and functional CNPs If the CNP is [+U], definite determination is semantically redundant.

!" CNP = lexically [+U] individual or functional noun (cf. the pope and mother examples)

!" CNP = lexically [–U] sortal or relational noun plus a modifier that turns a [–U] concept into a [+U] concept, such as

!" adjectival only

!" superlatives, last, next, favourite (Partee & Borschev), ordinals

!" [+U] appositions, number 2, word ‘kinezumi’, rumour that …

!" autophoric DDs: SC with “establishing relative clause”

!" artefacts-in-exclusive-use-possessives my | the toothbrush

level 0 shifts

level 1 shift


! " Incongruent definite determination: sortal and relational CNPs If the CNP is [–U], definite determination is semantically functional; it inevitably involves a type shift [–U] " [+U] (logically: <e,t> " e).

!" deictic use: The deictic gesture maps the sort described by the [–U] CNP to a particular individual of the sort. Note that “what Speaker points to” is an individual concept (in deictic uses enriched with the sortal information on the value of the function provided by the CNP)

!" anaphoric use: The sentential and wider context of the antecedent plus the sentential context of the anaphoric definite NP yield an individual concept for the referent.

(8) Reinhold met a yeti. He took a picture of the snowman.

individual concept: “the x such that: Reinhold met x; x is a yeti; (= antecedent context) x is a snowman, x is photographable” (= anaphor context)


! " Functional concepts and definiteness (1)

(9) a. Reinhold claims he saw [ [the footsteps]+U of [a yeti ]–U ]–U in the snow. # Reinhold saw [ yeti footsteps ]–U in the snow. $ Reinhold saw [ the yeti footsteps ]+U in the snow.

b. Reinhold claims he saw [ [the footsteps]+U of [the yeti ]+U ]+U in the snow. = Reinhold saw [ the yeti’s footsteps ]+U in the snow.

c. Reinhold claims he saw [ [footsteps]–U of [a yeti ]–U ]–U in the snow. # Reinhold saw [ yeti footsteps ]–U in the snow.

d. Reinhold claims he saw [ [footsteps]–U of [the yeti ]+U ]–U in the snow.

!" Definite or indefinite determination applies only

to the immediate operand, not necessarily to the whole NP !

( > mismatch of constituent structure and semantic composition)



!" The [U] value of a NP with possessor complement is the minimum of the [U] values of the possessum concept and the possessor concept.

!" If the possessum is a functional concept (FC), it inherits its [U] value from the possessor concept.

!" Referential transparency of FCs If the possessum is an FC, it inherits the total determination from the possessor concept, i.e. being (in)definite, possessive, deictic, anaphoric, quantifying, generic etc.



!" A functional CNP in absolute use (i.e. with no explicit possessor specification) with definite determination has an implicit [+U] possessor.

(10) special case: definite associative anaphor (DAA): definite [+U][+R] CNP with implicit anaphoric possessor argument

a. “How much is this?” – “[The price]+U [= of this+U] is attached on the back.”

b. I’ve bought a car, but something’s wrong with [the clutch]+U [of the car+U].

!" A functional CNP in absolute use with indefinite determination has a [–U] possessor.

(11) a. [A father]–U [of a student–U] came to my office hours the other day.

b. [A father]–U [of the student+U] came to my office hours the other day.

!" With functional CNPs in absolute use, explicit definite determination is pragmatically not redundant, as it entails that the possessor argument is [+U].


Evidence for the proposed theory of definiteness

!" Statistical evidence:

Congruent uses of definite and indefinite determination are more frequent

than incongruent uses.

from: Horn, Kimm & Gerland (to appear)

zero

zero

indef

zero

def

sortal

individual (lex.)

indiv. (p.n., p.p.)

relational

functional

[–U]

[+U]

[+U]

[–U]

[+U]

50%


Evidence (ctd.)

! " Psycholinguistic evidence:

Incongruent determination requires more processing time.

(work in progress, Peter Indefrey)

! " Typological evidence:

Incongruent determination receives more salient marking. Some options:

!" Incongruent uses are marked, while congruent uses are not

!" Congruent uses receive reduced marking as compared to incongruent uses.

!" Definiteness splits: > Existence of definiteness marking entails marking of pragmatic definiteness. > Certain types of semantically definite NPs are left unmarked


4. The scale of uniqueness / definiteness

deictic definites

< anaphoric definites

pragmatic definites < semantic definites

autophoric NPs

< definite associative anaphors (DAA)

< lexical IC, complex IC (SC with superlative, etc.)

< proper names

< personal pronouns


4. The scale of uniqueness / definiteness

general nouns proper names pers.pron.

with adnominal demonstratives

Grammatical distinctions

deictic anaph. autoph. DAA IC proper names pers.pron.

pragmatic def. semantic definiteness

Types of definite NPs


5. Definiteness splits

Adnominal demonstratives

! " The standard uses of adnominal demonstratives – deictic and anaphoric – require a [–U] CNP for enabling the deictic choice.

! " Demonstrative determination turns a [–U] CNP into a [+U] NP.

Demonstrative determination inevitably requires a level-2 type shift, i.e. reference draws on extralinguistic information.

! " Historically, anaphoric uses of demonstratives emerge from deictic uses.

! " Some languages have separate anaphoric determiners (e.g. Lakhota, Hausa. Lyons 1999: 53ff).

! " Adnominal demonstratives have the same range of application as pragmatic definiteness, i.e. [–U] CNPs.


Split type A : no definiteness marking

(13) Japanese (similarly: Chinese, Russian, Latin) a. sono hon wa nani? deictic, anaphoric

DEMADN book TOP what ‘what’s about this book?’

b. kin! katta (*sono) hon wa tsumaranai autophoric

yesterday bought DEMADN book TOP boring ‘the book I bought yesterday is boring’

c. kin! hon o katta. (*sono) taitoru wa oboe- nai DAA DEMADN title TOP remember-NEG ‘I bought a book yesterday. I don’t remember the title’

deictic anaph. autoph. DAA IC proper n. pers. pron.

demonstratives zero definites


Split type B : demonstratives extended to semantic definites

Polish Upper Silesian [Czardybon 2010], West Slavic: Upper Sorbian [Breu 2004]

(14) a. Dej mi ta flaszk-a. deictic

give me that/the bottle

b. Dziy" pryndzy paczka przisz-#-a i jo ta paczka #odebra-#-a. anaphoric

day before parcel arrived and I the parcel collected

c. tyn doktor co mie dzisiej autophoric

the doctor who me examined

d. …cha!pa. … (tyn) dach DAA … house. … (the) roof


demonstratives zero definites


Split type C : definite article different from demonstratives

English

(15) a. Is the dog yours? deictic

b. the dog that attacked me the other day autophoric

c. … a car … the clutch … DAA

d. The station is 10 minutes from here. IC (with article)

e. I’ll see her at school / at lunch. IC (without article)

f. (*The) John is visiting us tomorrow. proper name


demonstratives

definite article zero definites


Split type D1 : demonstratives, strong def, weak def, no proper names, zeros

Standard German (cf. Schwarz 2009)

(16) a. ?Am / An dem Tag, als ich geboren wurde, … autophoric

on the day when I was born

b. Er ist am / (*an dem) Kopf verletzt . DAA

he is at the head injured

c. Ich gehe jetzt zur / (*zu der) Post. IC

I go now to the post office

d. Das ist das Buch *vom / von Peter. proper name

that is the book of (the) Peter


demonstratives

full definites zero definites weak definites


Split type D2 : demonstratives, strong def, weak def, proper names, zeros

German dialects: weak def = weak article, or contraction [Studler 2004] North Frisian: weak “a-article”, strong “d-article” (Fering) [Ebert 1971]

Alemannic (Switzerland, Studler [2004])

(17) a. De Paul het es Ross gchouft. ... Das Ross laamt. proper name, theweak P. has a horse bought. … thestrong horse is lame anaphor

b. de sterchscht Maa vo de Wäut IC, IC

theweak strongest man of theweak world


demonstratives

stong definites zero definites weak definites


Split type D3 : demonstratives, strong def, weak def, no proper names, zeros

Standard Dutch (cf. Ortmann, to appear) also: Swedish (Stroh-Wollin 2003)

strong die–dat vs. general de–het def. determiner vs. suffix (plus det.)

(18) a. In dat gedicht wilde ik de liefde vergelijken ... anaphoric vs. IC

in thestr poem would I theweak love compare …

b. Jij was die vrouw die ik zag (Google) autophoric

you were thestrong woman who I saw

c. Dit is de vrouw die Osama verdedigte (Google) autophoric this is theweak woman who Osama defended


demonstratives

weak definites zero definites strong definites


Split type E : demonstratives, strong def, weak def, preproprial article, zeros

Swedish

(19) a. det witte hus-et ‘the white house’ SC DEF white house-DEF

b. witte hus-et ‘the White House’ IC

Swedish and Norwegian dialects

reduced 3rd person pronouns a–n with proper names as “preproprial” articles are “obligatory with persons’ given names” (Dahl 2007: 91)

c. a Brita n Erik proper name


demonstratives

weak definites zero definites

strong definites preproprial


Split type F : demonstratives, def. article, personal article

Maori [Bauer 2004: 114, 109]

(19) a. te whare kei te huringa autophoric, IC DEFART house at the corner ‘the house at the corner’

b. ka kite au i a ia personal pronouns T/A see 1sg DO PERSART 3sg ‘I saw him’


demonstratives

general definite article zero definites

preproprial


1. Observations about definite descriptions revisited

! " Semantic theory is preoccupied with anaphoric uses of definites

(Russell, GQT, DRT, File Change Semantics, Heim & Kratzer).

!" These are theories restricted to definites with sortal nouns/CNPs: [–U][–R]

Probable reasons:

!" By far the most general nouns are sortal.

!" Only with [–U] nouns, definiteness marking is semantically fully functional.

!" Linguistics preferably uses written data ! where anaphoric uses by far prevail.

!" While relational nouns enjoy increasing attention in linguistic theory, individual and functional nouns are hardly recognized as noun classes of their own.


1. Observations about definite descriptions revisited

For languages with definiteness marking:

! " There are certain conceptual types of nouns for which the definite article is (almost) obligatory: individual and functional nouns.

Exceptions are due to incongruent, i.e. indefinite, uses of [+U] nouns

(cf. (2) and (6) above). These decrease in frequency in the following order:

FN < IN < proper names < 3rd person < 2nd, 1st person

To the extent that there are incongruent uses, marking definiteness with these subtypes of nouns is functional even if semantically redundant.

> The more frequent the incongruent uses are, the more likely the congruent uses will be marked as definite.


! " For semantic definites, definiteness marking is semantically redundant.

Further reasons for the existence of definiteness marking of semantic

definites:

!" Indicating indexicality is functional even for semantically unique CNPs (> situational argument for IC, FC)

!" Analogy pressure towards a uniform syntactic structure of NP/DP motivates explicit determination for all general nouns.

!" Often articles carry nonsemantic functions: marking of case, number, gender, or noun class


Splits

!" A universal split separates pragmatic from semantic definites.

additional splits > within pragmatic definites: deictic vs. anaphoric

> within semantic definites autophoric vs. other semantic definites

general nouns vs. proper names and personal pronouns

proper names vs. personal pronouns


! " In most cases, definite articles developed from demonstratives.

!" Most general nouns are sortal (type <e,t>).

!" Demonstratives take sortal nouns and yield definite NPs ! they map a [–U] CNP on a [+U] NP.

!" In the course of grammaticalization of definiteness markers from demonstratives, the following steps are involved:

-" demonstratives are getting used without a deictic gesture (cf. anaphoric uses, also “anamnestic” uses [Himmelmann 1997])

-" deictic distinctions are neutralized (cf. anaphoric uses)

-" the requirement of [–U] input is dropped (first: autophoric uses, later “abstract situational uses” [Hawkins 1978] with IN and FN)

!" Finally, demonstratives-turned-definiteness-markers admit [±U] CNPs and just mark the result as a [+U] NP.

Selected references 35

Bauer, Winnifred (1993), Maori. Routledge London.

Breu, Walter (2004), Der definite Artikel in der obersorbischen Umgangssprache. In: Marion Krause & Christian Sappok (eds.), Slavistische Linguistik 2002. Referate des XXVIII. Kon-

stanzer Slavistischen Arbeitstreffens. Bochum, 10.-12.9.2002, 9–57. München: Sagner.

Czardybon, Adrian (2010), Die Verwendung des definiten Artikels im Oberschlesischen im

Sprachvergleich. Master Thesis, University of Düsseldorf. Dahl, Östen (2007), Grammaticalization in the North: Noun Phrase Morphosyntax in Scandinavian

Vernaculars. Stockholm University. www2.ling.su.se/staff/oesten/downloads/Gram_north.pdf

Ebert, Karen H. (1971), Referenz, Sprechsituation und die bestimmten Artikel in einem

nordfriesischen Dialekt (Fering). Bredstedt: Nordfriisk Instituut.

Hawkins, John A. (1978), Definiteness and Indefiniteness. Croom Helm. London.

Himmelmann, Nikolaus P. (1997), Deiktikon, Artikel, Nominalphrase. Zur Emergenz syntaktischer

Struktur. Niemeyer. Tübingen.

Horn, Christian, Nicolas Kimm, Doris Gerland (to appear), Empirical Evidence for Concept Types in German Texts. In Th. Gamerschlag, D. Gerland, R. Osswald, W. Petersen (eds.), Concept

types and frames - Applications in Language, Cognition and Philosophy.

Löbner, Sebastian (1985), Definites. Journal of Semantics 4: 279-326.

Löbner, Sebastian (2011), Concept types and determination. Journal of Semantics 28: 279-333.

Lyons, Christopher. (1999), Definiteness. CUP. Cambridge.

Selected references 36

Ortmann, Albert (to appear), Definite article asymmetries and concept types: semantic and pragmatic uniqueness . In Th. Gamerschlag, D. Gerland, R. Osswald, W. Petersen (eds.), Concept types and frames - Applications in Language, Cognition and Philosophy.

Partee, Barbara, & Vladimir Borschev (2002), Integrating lexical and formal semantics: Genitives, relational nouns, and type-shifting. In Robin Cooper, Thomas Gamkrelidze (eds.), Proceedings of the 2nd Tbilisi Symposium on Language, Logic and Computation. Tbilisi State University. Tbilisi. 229–241.

Schwarz, Florian (2009), Two Types of Definites in Natural Language. PhD dissertation, University of Massachusetts Amherst. Studler, Rebekka (2004). Voller und reduzierter Artikel in der schweizerdeutschen DP. In: Bra%i%, S., &uden, D., Podgor'ek, S. & Poga%nik, V. (eds.), Linguistische Studien im Europäischen Jahr der Sprachen. Akten des 36. Linguistischen

Kolloquiums in Ljubljana. Frankfurt a.M.: Lang. 625–635.

Stroh-Wollin, Ulla (2003), (Double) definiteness in Swedish. In: Hans-Olof Delsing et al. (eds.), Grammatik i fokus Vol. 2: Festschrift for Christer Platzack, 335-342. Lunds universitet: Institutionen för nordiska språk.

Studler, Rebekka (2004), Voller und reduzierter Artikel in der schweizerdeutschen DP. In: Bra%i%, S., &uden, D., Podgor'ek, S. & Poga%nik, V. (eds.), Linguistische Studien im Europäischen Jahr der Sprachen. Akten des 36. Linguistischen Kolloquiums in Ljubljana. Frankfurt a.M.: Lang. 625–635.

Acknowledgements 37

The research underlying this talk was largely funded by the Deutsche Forschung-gemeinschaft (DFG) grants Research Unit RU 600 “Functional Concepts and Frames” and Collaborative Research Centre CRC 991 “The Structure of Representations in Language, Cognition, and Science”.

I discussed matters of this talk with many colleagues including Adrian Czardybon, Christian Horn, Nicolas Kimm, and Doris Gerland; I would like to thank Albert Ortmann in particular for numerous discussions and for providing many of the data I quote.

Documents

Definiten esssplitsarchive.sfl.cnrs.fr/sites/sfl/IMG/pdf/lsalaa2012Loebner.pdf · x is a snowman, x is photographable” (= anaphor context) 1. Observations 2. Concept types 3. Uses