Properties, entity correlates of properties, and existentials

1
Properties, entity correlates of properties, and
existentials

• Louise McNally
• Universitat Pompeu Fabra
• Workshop on Indefinites and Weak Quantifiers
• 1/7/2005

2
Goals of the talk

• Review two underappreciated aspects of McNallys
  1992 semantics for English existential sentences
• Lexical semantics of There be
• Formalization of the notion of property
• Reflect on some of their implications for the
  study of incorporation(-like) phenomena

3
McNally 1992 lexical semantics

• There be denotes a property of properties
  paraphraseable as to be instantiated.
• Existential quantification is only entailed it
  plays no role in semantic well-formedness (cf.
  Strawson 1959).
• Similarities between There be sentences and
  incorporation constructions or presentational-ther
  e sentences are accidental or superficial.
• Translation ?P.There-be(P)

4
Alternative lexical semantics

• There be denotes a property of entities
  paraphraseable as to exist but (for whatever
  reason) it can only combine with its argument via
  something like semantic incorporation (see e.g.
  Van Geenhoven 1998).
• The role of existential quantification depends on
  how semantic composition works.
• A deeper connection to presentational sentences
  is in principle possible (see e.g. Landman 2004).
• Translation ?P?x.There-be(x) ? P(x)

5
The formalization of property

• McNally 1992 The postverbal NP in
  existential-there sentences denotes the entity
  correlate of a property (a sort within type e,
  using a property-theoretic semantics inspired in
  Chierchia 1984)
• McNally 1998 and various others The NP denotes a
  functional type in a set-theoretic semantics
  (e.g., type lte,tgt ex-/intensionality will be set
  aside).

6
Basic case

• (1) There were problems.
• McNally 1992 ?y.There-be(y)(n?x.problems(x))
  There-be(n?x.problems(x))
• McNally 1998 ?P.There-be(P)(?x.problems(x))
  There-be(?x.problems(x))

7
The problem case

• Quantification into postverbal position (Lumsden
  1988)
• (2) Among our dresses there were most kinds of
  shabby and greasy wear,.
• C. Dickens, The Uncommercial Traveler, Chap. 4

8
The problem case McNally 1992

• most kinds of shabby and greasy wear can quantify
  unproblematically over the existentials argument
  if we model kinds as entity correlates of
  properties (Cocchiarella 1976)
• kind(s) of shabby and greasy wear denotes a
  property of subkinds of the kind shabby greasy
  wear
• ?x.kind(x,n?y.shabby-greasy-wear(y))
• Most xkind(x, n?y.shabby-greasy-wear(y))
• There-be(x)

9
The problem case McNally 1998 (1)

• How to make most kinds of shabby and greasy wear
  quantify over a type lte,tgt argument?
• Option 1 kind(s) of shabby and greasy wear
  denotes a property of properties (see also de
  Swart 2001)
• ?P.kind(P, ?y.shabby-greasy-wear(y))
• Most Pkind(P, ?y.shabby-greasy-wear(y))
• There-be(P)
• Problem Motivating a higher type for nouns like
  kind and resulting proliferation of types for
  other categories

10
The problem case McNally 1998 (2)

• Option 2 kind(s) of shabby and greasy wear
  denotes a property of subkinds of the kind
  shabby greasy wear, as on McNallys 1992
  analysis use type coercion to permit
  quantificationbut how?
• If the argument NP is coerced, its unclear how
  quantification will work
• Most xkind(x,n?y.sh-gr-wr(y))?P.There-be(P)(?x
  )

11
The problem case McNally 1998 (3)

• Option 2, cont. Coercing the type of the
  existential predicate (ltlte,tgt,tgt ? lte,tgt) amounts
  to adopting the McNally 1992 analysis for the
  problem case
• Most xkind(x,n?y.sh-gr-wr(y))?z.There-be(?z)(x
  )
• Conclusion On the basis of simplicity, the
  analysis on which the existential predicate
  always selects for an entity correlate of a
  property appears superior.

12
Another argument for entity correlates of
properties

• The failure of non-DP, property-type expressions
  to appear in existentials is not
  straightforwardly predicted by a property
  analysis
• (3) There was happiness/happy.
• (4) There was president. (cp. She is president.)
• If There be selects for an entity correlate of a
  property, we need only exclude APs, PPs, etc.
  from denoting such entities. A theory of the
  semantics of nominalization should handle this
  (e.g. Chierchia 1984).

13
Reflection on existentials

• We shouldnt be surprised if the semantics of
  existential constructions is cross-linguistically
  variable.
• Still, finding empirical arguments to distinguish
  obligatory incorporation over an ordinary entity
  from lexical selection for a(n entity correlate
  of a) property might be difficult.
• However, in the spirit of Chierchia 1984, we
  might expect a correlation between lexically
  selecting for a property and that property being
  best formalized in type e.
• If so, it might be useful to test for whether an
  existential construction selects for a
  property-qua-function or its entity correlate.

14
Example W. Greenlandic?

• Unlike There be, incorporating verbs in W.
  Greenlandic can incorporate non-nominal
  predicates (see refs. in Van Geenhoven 1998)
• If W. Greenlandic existential qar- combines with
  non-nominal predicates, we can explain the
  similarities and contrasts with There be by
  positing that qar- denotes a property of
  ordinary entities but is obligatorily
  incorporating.

15
Reflection on incorporation

• Entity correlates of properties are not the right
  sorts of input to operations such as Farkas de
  Swarts Unification or Chung Ladusaws 2004
  Restrict.
• We might expect clear cases of Unification or
  Restrict to contrast with the English existential
  construction in terms of quantification over
  properties/kinds.

16
Example Chamorro?

• Unlike There be, the incorporating Chamorro
  existential predicates guäha exist, taya not
  exist, and täi- not have do not permit
  quantification over properties or kinds (Chung
  Ladusaw 2004, chapter 3, fn. 5).
• This contrast with English follows if the
  composition mechanisms involved in the two cases
  are distinct ordinary function application for
  There be, and Unification or Restrict for the
  Chamorro existential predicates.

17
Reflection on properties and kinds

• If the English existential construction selects
  for the entity correlate of a property and if
  kinds are modeled as entity correlates of
  properties, the distinction between properties
  and kinds must be more subtle than we thought
  Expressions like dogs and three dogs will denote
  in the same general sort in English.
• Perhaps opens the door for reconciling certain
  aspects of Chierchias (1998) work on Romance
  with that represented by e.g. Dobrovie-Sorin and
  Laca 2003.

18
Final comments

• The possibility that some expressions might
  really lexically select for properties should not
  be forgotten in the move to abandon the original
  lexical formulation of semantic incorporation in
  favor of alternative composition mechanisms.
• Entity correlates of properties have been
  prematurely marginalized even by their few
  promotors their utility beyond the modeling of
  kinds should be reexplored.

19
References
Chierchia, G. 1984. Topics in the syntax and
semantics of infinitives and gerunds. Ph.D.
diss., U. MA Amherst. Pub. 1989, Garland, New
York. Chierchia, G. 1998. Reference to kinds
across languages. NLS 6, 339-405. Chung, S. W.
A. Ladusaw. 2004. Restriction and saturation.
MIT Press, Cambridge, MA. Cocchiarella, N. 1976.
On the logic of natural kinds. Philosophy of
Science 43, 202-222. de Swart, H. 2001. Weak
readings of indefinites type-shifting and
closure. The Linguistic Review 18,
69-96. Dobrovie-Sorin, C. B. Laca. 2003. Les
noms sans déterminant dans les langues romanes.
In D. Godard (ed.), Les langues
romanes. Problèmes de la phrase simple.
Editions du CNRS, Paris, 235-281. Farkas, D. H.
de Swart. 2003. The semantics of incorporation.
CSLI Pubs., Stanford, CA.
20
References, cont.
Landman, F. 2004. Indefinites and types of sets.
Blackwell, Oxford. Lumsden, M. 1988. Existential
sentences Their structure and meaning.
Routledge, London. McNally, L. 1992. An
interpretation for the English existential
construction. Pub. 1997 as A semantics for the
English existential construction, Garland, New
York. McNally, L. 1998. Existential sentences
without existential quantification. LP 21,
353-392. Strawson, P. F. 1959. Individuals An
essay in descriptive metaphysics. Routledge,
London. Van Geenhoven, V. 1998. Semantic
incorporation and indefinite descriptions. CSLI
Publications, Stanford, CA. More eventually at
http//mutis.upf.es/mcnally louise.mcnally_at_upf.e
du