Navigation – Plan du site

On definite kinds1

Olga Borik et M. Teresa Espinal
p. 123-146

Résumés

Dans cet article, nous défendons trois hypothèses. Premièrement, toutes les langues qui possèdent des déterminants (nuls ou manifestes) ont des termes d’espèces définis, une possibilité qui n’empêche pas les langues de recourir à d’autres moyens pour faire référence aux espèces. Deuxièmement, les espèces sont dénotées par des DPs définis sans Nombre. Troisièmement, l’interprétation de sous-espèces se fait sur Nombre. Nous soutiendrons empiriquement ces hypothèses en nous appuyant sur l’analyse contrastive de deux langues qui mettent en œuvre des stratégies opposées pour marquer la définitude : l’espagnol, une langue romane avec articles, et le russe, une langue slave sans article. Nous faisons la prédiction que les termes d’espèces définis ne peuvent pas se combiner avec des prédicats qui encode la pluralité, ne peuvent pas déclencher une interprétation générique avec des prédicats s-level, et ont une interprétation différente de celle associée aux définis pluriels.

Haut de page

Texte intégral

1. Introduction

  • 1 This research has been funded by the following two research grants awarded by the Spanish Ministeri (...)

1A central topic in the literature on genericity is how different types of languages refer to kinds. In this respect, a well-established assumption, since Carlson (1977/1980), is that bare plurals (BPs) in English (E) allow for a generic use, a reading that arises because BPs are taken to denote a particular type of entity: names of kinds of things. This view reflects the fact that BPs in E may occur as arguments of generic sentences (with kind-level and individual-level predicates).

2However, a still unresolved and poorly understood phenomenon is the question why E also allows the use of definite generics, as pointed out by Carlson (1977/1980:274-280) and Chierchia (1998:379-383):

(1)

a.

The owl is common/widespread/fast disappearing/often intelligent.

 

 

(Carlson 1977/1980:276, (32a))

 

b.

Owls are common/widespread/fast disappearing/often intelligent.

 

 

(Carlson 1977/1980:276, (32b))

In this paper we will challenge the standard assumption that the generic or kind reading for NPs is basically modeled over pluralities (Chierchia 1998), and we will focus on so-called “singular” definite generics as in (1a).

3We will defend the following hypotheses:

4H1. All languages that have Determiners (null or overt) have definite kinds, a possibility which does not prevent languages from using other means to refer to kinds (e.g. bare plurals in E).

5H2. Kinds are referred to by definite DPs with no Number involved. We will provide empirical support for this hypothesis based on a contrastive analysis of two languages that show opposite strategies for marking definiteness: Spanish (S), a Romance language with articles, which does not allow for generic BPs (Laca 1990; Dobrovie-Sorin and Laca 1996, 2003), and Russian (R), a Slavic language with no article.

6H3. In these languages the subkind interpretation is built on Number. In combination with i(ndividual)-level and k(ind)-level predicates argument DPs type-shift their reading from objects to subkinds, otherwise the interpretation crashes.

7The prediction we make in this paper is that languages that have (null or overt) Determiners do not differ with respect to definite generics: they all have them. However, we think that languages differ a lot in the domain of pluralities (cf. Farkas and de Swart 2007).

2. The kind interpretation

2.1. The definite generic The N according to Carlson

8Carlson distinguishes between two classes of individuals: objects and kinds. Objects can occupy one place at a time. Kinds may function in such a way as to be in many places at a given time. Kinds also serve to tie objects as well as stages together to make them manifestations of the same type of thing. Objects only serve to tie stages together, and never other objects. Unlike kinds, objects cannot have stages that appear at various distinct places at a given instant.

9We interpret this view as corresponding to an extensional view of kinds, which can be represented in a semi-lattice structure (2), where the kind corresponds to the highest node in the lattice.

According to Carlson, semantically, there is a great deal of similarity between the definite generic and the bare plural. Indeed, he claims that the owl and owls behave entirely in parallel.

(3)

a.

The owl has two eyes in the front of its skull.

 

b.

Owls have two eyes in the front of their skull.

 

 

(Carlson 1980:276, exs. (31))

10Let us consider in more detail the definite generic vis-à-vis the bare plural. Carlson takes both to be NPs that denote the property set of a kind. He claims that the definite generic appears to share most (but not all, apparently) of the individual-level properties of the bare plural that corresponds to it. The definite generic is much more limited in the sorts of NP’s it may productively occur in than the bare plural, the latter being virtually unbounded. Referring to E, Carlson states that definite generics are much more restricted in the distribution than bare plurals because they can only refer to well-established kinds. He also regards the definite generic as a proper name: the fact that it takes the definite article makes it seem to be much closer syntactically to proper names, as many proper names have as a part of the name the definite article as well (The Hague, The Mississipi, The Bronx, etc.). According to this view the definite generic appears to be more name-like than the bare plural in several respects, but Carlson attributes to both definite generics and bare plurals a name-like status.

11Our reservations with respect to Carlson’s analysis are the following. First of all, even if we consider the extensional counterpart of owls as denoting the intension of a set (i.e., a function that picks out all existent owls in any possible world at any given time), it seems to us that the owl cannot denote in the same way the intension of a set. The denotation of plural nominal expressions in English seems to be closer to mass nouns, but different from the denotation of definite kinds, which are similar to atomic entities.

12Secondly, Carlson highlights a number of similarities between bare plural kinds and definite kinds, but never explains the relevant differences between them might be, apart from pointing out the restrictions involved in the use of definite generics. We argue that these expressions have different denotations relying on the presence or absence of number.

13In the account that we are going to propose in this paper, we rather follow Jespersen (1927), who characterizes the definite generic as denoting the species itself. Plurals, on the other hand, denote all members of the species.

14We will also rely on the hypothesis that there is no ambiguous Determiner. Romance languages may show overt Determiners preceding proper names. The syntactic form and semantics of this definite Determiner is the same as for definite kinds, where an overt Determiner precedes a common noun. We will come back to this hypothesis in Section 3.

2.2. Singular generic The according to Chierchia

15In Cherchia’s proposal, kinds are “individual concepts of a certain sort: functions from worlds (or situations) into pluralities, the sum of all instances of the kind” (Chierchia 1998:349). Kinds are derived from properties by the down operator , which is defined as follows (Chierchia 1988:351):

(4)

For any property P and world/situation s,

(i.e., the down operator maps a property to the greatest element in the extension of P in a world).

16Under this view, the representation of the BP kind-denoting NP owls is as in (5):

(5)

OWL = o

By contrast, singular generic NPs with the definite article the (like the owl) pick out a group individual which do not allow any reference to its members. This accounts for the fact that these nominal expressions cannot combine with predicates that require a count interpretation (cf. (6a) and (7a)). In this respect, definite kinds show a contrast with “regular” bare plural kind NPs, as shown in (6):

(6)

a.

*The owl is numerous.

 

b.

Owls are numerous.

(7)

a.

*The Jones family is numerous.

 

b.

The members of the Jones family are numerous

According to Chierchia, the semantic derivation of singular generic NPs is complex and proceeds with the following two steps, as illustrated in (8):

(8)

a.

Step1: a MASS operator applies to the denotation of a singular count noun like owl to produce MASS(owl). It has a mass reference, i.e. it applies both to atomic owls and pluralities of owls.

 

b.

Step2: the iota-operator applies, but its semantics is different for count and mass noun denotations (Chierchia 1998:380):

The resulting representation of the definite kind NP the owl is given in (9):

(9)

the owl THE(MASS(owl) = g(ι MASS(owl))

It should be noted that, according to the analysis of Chierchia, both “the sums of all instances of the kind” (bare plural kinds) and groups (definite singular kinds) are atomic, although they have different internal structures. But we think that this, in fact, poses a conceptual problem, since kinds denoted by BPs seem to have a dual nature: on the one hand, they are assumed to represent “the sums of all instances” or “the largest plurality” and, on the other hand, kinds constitute a subset of atomic entities in the semi-lattice structure.

  • 2 Krifka actually notices that the group semantics might explain the use of definite generics in sent (...)
  • 3 The prediction of Chierchia’s analysis with respect to bare plurals are not discussed in this paper (...)

17Apart from this, we have an empirical objection to Chierchia’s analysis. It should be observed that we do not normally find generic the-NPs in episodic sentences (as noted by Chierchia himself, but also Krifka 2004).2 Given that the definite article in his analysis is basically assumed to be ambiguous (cf. (8b)), it is unclear why (10) does not convey this ambiguity. In particular, nothing seems to prevent that in (10), which is an episodic sentence with a stage-level predicate, the derivation in (9) applies, giving the generic interpretation to the subject.3

(10)

The owl hunted/was hunting a rat this morning

In the account that we are going to develop in this paper, we assume that BNs do not have to be shifted into a mass reading to derive a kind interpretation and we argue that definite kinds do not rely on the notion of plurality at all. Furthermore, we treat the definite D uniformly as the iota operator without encoding any sort of ambiguity. With regard to the empirical problem posed in (10), we rely on the hypothesis that definite kinds do not combine with stage level predicates, as will be explained in the next section.

3. The denotation of nominal expressions and definite kinds

18Let us start with the assumption that the canonical syntactic structure (Chierchia 1998, Longobardi 2001, Zamparelli 1995) for nominal expressions in languages with number morphology and determiners is that given in (11):

(11)

[DP D [NumP Num [NP N ]]]

This tripartite syntactic) organization schema is not, however, the only possible one, since nominal structures may lack a Number specification (12b), a Determiner (12c), or both (12a).

(12)

a.

[NP N ]

 

b.

[DP D [NP N ]]

 

c.

[NumP Num [NP N ]]

Therefore, we have to consider three additional syntactic structures in 12(a-c) that emerge from this schema and which have been postulated respectively for (i) bare count nominals (BNs) in object position in several Creole and Romance languages (Déprez 2005, Dobrovie-Sorin et al. 2006); (ii) singular expressions occurring in subject position in languages such as Brazilian Portuguese (Munn and Schmitt 2005); and (iii) existential BPs in object position in languages such as Catalan, Romanian and S (Dobrovie-Sorin and Laca 1996, 2003; Dobrovie-Sorin et al. 2006). The fact that these different syntactic possibilities exist does not imply that languages necessarily have predicates that select for these different types of nominal expressions (see below in this section). We think that semantic arguments are basically of two types: individual objects or kinds, but the four structures in (11) and (12) offer more possibilities and, therefore, additional mechanisms are needed in natural languages to match the requirements of the predicate with the meaning of the nominal expression it combines with.

19Let us focus on the semantics of (12a) in contrast to (12b). We start from the assumption that of the four nominal structures in (11) and (12), the “least marked” one (see Swart and Zwarts 2009) is the representation in (12a), which corresponds to BNs when occurring as objects of verbs and as objects of prepositions, the situation illustrated for S in (13).

(13)

a.

Este profesor tiene libro. (P. Goenaga, p.c.

 

 

this professor has book

 

 

‘This professor has a book/ has published.’ (It could be one book, or more than one

 

b.

Estuvieron en prisión.

 

 

were in prison

 

 

‘They spent time in prison.’ (It may have been more than one prison

Note that these BNs, in object position of Vs and Ps, have a number neutral interpretation, as suggested by the glosses (see Espinal 2010, Espinal and McNally 2011), and hence do not have inherent number. Therefore, we follow the hypothesis that they are morphologically unmarked for Number, and do not require a Num projection in syntax (see Munn and Schmitt 2005 for a similar claim with respect to Brazilian Portuguese).

20Besides Num, BNs in S also lack a Determiner. Several arguments support this claim: (i) BNs can occur neither in subject position (which is possible in BrP; Munn and Schmitt 2005, Müller 2002) nor as indirect objects, and (ii) cannot be assumed to move to a strong D0 position (as postulated for canonical arguments in Italian by Longobardi 1994, 2001) because BNs are not associated with a marker of determined atomic reference.

21We assume that the denotation of bare common count nouns, corresponding to structure (12a), is a property. More specifically, we assume that the BNs in (13) denote properties of kinds, where kinds are understood as intensional entities, without any reference to the particular individuals that might instantiate a kind (Espinal and McNally 2007, 2011; Dobrovie-Sorin and Pires de Oliveira 2007).

22Definite nominals may have either the structure in (11) or the one in (12b). We postulate that the latter is the one that corresponds to definite kinds, while the former corresponds to definite singular or plural individual objects, i.e. non-kind entities.

23Let us start with by considering the minimal pair in (14a-b), as discussed in Espinal (2010).

(14)

a.

Tener gripe (porcina)

 

 

have flu swine

 

 

‘To have swine flu.’

 

b.

Tener la gripe (porcina)

 

 

have the flu swine

 

 

‘To have the swine flu.’

 

c.

*Tener las gripes porcinas.

 

 

have the.pl flu.pl swine.pl

  • 4 There is a nice parallel in Russian for this type of examples. Consider the contrast in (i):
    (i) a.
    (...)

In (14a) “have–flu” or “have–swine flu” (“swine flu” being a subtype of “flu”) is a characterizing predicate of the external subject, in the sense that it adscribes a complex property of “having (swine) flu” to it. In (14b) the definite D conveys a generalization and denotes the kind/class of viruses known as la gripe porcina the swine flu”; this definite D does not refer to the set of entities which belong to the kind, but rather names the kind itself, as proved by the fact that this nominal expression cannot occur in a definite plural form (14c). This example shows that definite kinds have no Number associated with them.4

  • 5 As will soon become clear, the prediction we make here is that the plural counterpart of (15a) can (...)

24Furthermore, only the definite expression la gripe (porcina) (but not the BN) can occur in subject position of a kind-selecting predicate like estar por todas partes to be widespread”, as shown in the minimal pair exemplified in (15).5

(15)

a.

La gripe porcina está por todas partes.

 

 

the flu swine is around every parts

 

 

‘The swine flu is widespread.’

 

 

*Gripe porcina está por todas partes.

 

 

flu swine is around every parts

The meaning of bare count nouns, as denoting properties of kinds of individuals of type ⟨ek,t⟩ (Espinal and McNally 2007, Espinal 2010) is represented in (16b), where the superscript k indicates a kind-level individual x.

(16)

a.

[NP N ]

 

 

b.

⟦N⟧ = λPλxk [P(xk]

b′. ⟨ek,t⟩ properties of kinds

25The definite D is formally translated as the iota-operator (Partee 1987), which combines with a property to give an entity-denoting expression, the atomic kind (xk). It is crucial to note that the iota-operatorin this analysis binds variables of kinds, and this derivation syntactically corresponds to definite nominal expressions not marked for Number. Therefore, structure (12b), repeated here as (17a), corresponds to what we name definite kinds, not singular definite kinds or singular generics.

(17)

a.

[DP D [NP N ]]

 

b.

⟦D N⟧ = λPιxk[P(xk)] b′. ⟨ek⟩ kind denotation

  • 6 As an anonymous reviewer pointed out, it is not obvious that the definite is not ambiguous if it ca (...)

Recall that we rely on the hypothesis that there is no ambiguous Determiner. Romance languages may show overt Determiners preceding proper names. The syntactic form and semantics of this definite Determiner is the same as for definite kinds, where an overt Determiner precedes a common noun.6

26An argument in support of the syntactic-semantic correlation just made relies on speakers judgements about possible discourse relationships. Properties of kinds cannot introduce discourse referents, while kind denoting expressions can. The examples in (18) illustrate the fact that kinds are referred to in discourse by means of token discourse anaphors (like the Catalan third person accusative clitic el), whereas BNs introduce properties, and reference to them in discourse is only possible by means of a type anaphor (like the Catalan property-type anaphoric clitic en).

(18)

 

CATALAN

 

a.

Alguna cosa va exterminar el dodo, però què el / #en va

 

 

some thing past exterminate the dodo but what it.acc / prop past

 

 

exterminar?

 

 

exterminate

 

 

‘The dodo was exterminated by something, but what was it?’

 

 

(Espinal 2010: note 19)

 

 

 

b.

Avui porta faldilla. Li ’n vam regalar

 

 

today wear.3sg skirt to.her prop past.1pl give.present

 

 

una l’any passat.

 

 

one the.year last

 

 

‘Today she is wearing a skirt. We gave her one as a present last year.’

 

 

(Espinal and McNally 2011: ex.(17))

With these assumptions in mind, our main theoretical claim is that a definite determiner, standardly interpreted as the iota operator ι, is responsible for instantiating a kind-denoting expression if, and only if, it applies directly to a Noun. Crucially, no Number is involved in the composition of a definite kind interpretation.

27Let us now consider what happens when a syntactic projection that encodes Number is present. If Number is available, we assume that it relates properties of kinds to properties of objects, singular or plural, of that kind. It correlates with Carlson’s (1977) realization relation R (cf. Déprez 2005), although for him R is a two-place relationhip between stages and individuals.

(19)

a.

⟦Num[–PL] N⟧ = λPλxkλyo [P(xk) ∧ R(P(xo), P(xk))] ⟨eo,t⟩ properties of objects

 

 

⟦Num[+PL] N⟧ = λPλxkλyo [P(xk) ∧ R(*P(xo), P(xk))]

In addition, the iota operator has a uniform semantics and applies either to properties of kinds (17b) or to properties of objects of this kind (20b). In the former case the output is a kind entity, in the latter case the output is an object entity.

(20)

a.

[DP D [NumP Num [NP N ]]

 

b.

⟦D Num N⟧= λPλxkιxo [P(xk) ∧ R(P(xo), P(xk))] b′. ⟨eo⟩ object denotation

We therefore claim and postulate that the difference between two types of entities, kinds and objects, initially proposed by Carlson, is morphosyntactically encoded by the presence or absence of Number.

28As for predicates, we assume the following typology: (i) kind-level predicates select for ⟨ek⟩ arguments, (ii) individual-level predicates may select for ⟨ek⟩ or ⟨eo⟩, and (iii) stage-level predicates only select for ⟨eo⟩. A model based on this predicate typology predicts the possibility to use a kind-referring definite NP with k-level and i-level predicates, as illustrated in (21a-b). This predicts that definite kinds must also be allowed as subjects of conjoined k-level and i-level predicates, as (21c) illustrates.

(21)

a.

The blue whale is the largest mammal in the world.

 

b.

The blue whale lives in all oceans, mainly in the Artic and the Antartic.

 

c.

The blue whale lives in all oceans and is the largest mammal in the world.

However, a s-level predicate makes it impossible to interpret the subject DP as a kind. In order to make a generic statement with a s-level predicate the sentential operator GEN is required (Krifka et al. 1995). This type of abstract operator is what guarantees the appropriate interpretation for (22).

(22)

A blue whale eats an average of three tones of food a day.

This approach predicts that a distinction is to be made in natural languages between definite kinds (no Number projection in syntax, selected by k-level predicates and possibly also by i-level predicates), and (in)definite generics (Number projection in syntax, GEN operator, selected by s-level predicates). And, more importantly, this difference in meaning is encoded, on the one hand, by the morphosyntactic specification of Number on the DP, and, on the other hand, by a morhphosyntactic realization of the GEN operator.

4. Definite kinds in Spanish

  • 7 Note that (23b) is ambiguous between the definite kind reading, associated with a DP structure that (...)

29The model we have just presented straightforwardly accounts for the S data in (23). Note that a kind-denoting subject is allowed with both k- and i-level predicates. If a singular definite DP subject is combined with either an i-level or a s-level predicate, then the existence of a unique individual object is to be inferred.7

(23)

a.

El dodó se extinguió en el siglo XVII. k-level; [DPD [NPN ]]

 

 

the dodo CL extinguished in the century XVII

 

 

‘The dodo was extinct in the XVII century.’

 

b.

El dodó vivió en la isla Mauricio. i-level; [DP D [NP N ]]

 

 

the dodo lived in the isle Mauritius or [DP D [NumP Num [NP N ]]]

 

 

‘The dodo lived in Mauritius Island.’

 

c.

El dodó fue disecado en el Museo Ashmolean. s-level;

 

 

the dodo was dissected in the museum Ashmolean [DP D [NumP Num [NP N ]]]

 

 

‘The dodo was dissected in the Ashmolean Museum.’

Our claim that there is no Number involved with the kind-referring DP subject is supported by the following piece of data:

(24)

a.

La nevera se inventó en el siglo XVIII.

 

 

the fridge CL invented.3sg in the century XVIII

 

 

‘The fridge was invented in the XVIII century.’

 

b.

*Las (dos) neveras se inventaron…

 

 

the two fridges CL invented.3pl

The ungrammaticality of (24b) supports the correlation we make between definite kinds and absence of morphosyntactic Number. In other words, as already mentioned in the previous section, in the absence of any explicit evidence for semantic or morphosyntactic Number in subject DP position, we assume the minimal hypothesis that there is no Number projection. Definite kinds are numberless. The reason for this generalization is that in S, as in many other languages (cf. Beyssade 2005 for French and Doron 2003 for Hebrew), there is a clear contrast between definite kinds, conceived as intensional atomic entities, and maximal sums of individuals, conceived as extensional non-atomic (plural) entities. We come back to this contrast at the end of this section.

30Now, if Number is assumed to relate properties of kinds to properties of objects of that kind (as presented in Section 3), the next question that we have to answer is how we can account for Number (singular and plural) in DP structures that refer to subkinds, like in (25). Note that the DPs in (25a,b) contain either a demonstrative or a numeral.

(25)

a.

Esta ballena está en peligro de extinción. subkind

 

 

this whale is on verge of extinction [DP D [NumP Num[–PL] [NP N ]]]

 

 

‘This whale is on the verge of extinction.’

 

b.

Dos ballenas están en peligro de extinción. subkind

 

 

two whales are on verge of extinction [DP D [NumP Num[+PL] [NP N ]]]

 

 

‘Two whales are on the verge of extinction.’

 

c.

La ballena está en peligro de extinción. kind

 

 

the whale is on verge of extinction [DP D [NP N ]]

 

 

‘This whale is on the verge of extinction.’

  • 8 See also Carslon (1977/1980:26) for a specific reference to English names of kinds of things.

One important difference between (25a-b) and (25c) is that in the former a lexical item such as clase, tipokind”, a kind-noun (Zamparelli 1995), can always be inserted.8 With the definite kind DP this possibility is discarded.

  • 9 If the subject of this sentence is modified by a relative clause, the example is grammatical with a (...)

(26)

a.

Esta clase de ballena está en peligro de extinción

 

 

this kind of whale is on verge of extinction

 

 

‘This kind of whale is on the verge of extinction.’

 

b.

Dos tipos de ballenas están en peligro de extinción.

 

 

two kinds of whales are on verge of extinction

 

 

‘This kind of whale is on the verge of extinction.’

 

c.

*La clase de ballena está en peligro de extinción.9

 

 

the kind of whale is on verge of extinction

Going back to example (24b), it is interesting to note that it can only be turned into a grammatical sentence if a noun that names kinds of things is overtly specified. Example (27) necessarily conveys a taxonomic interpretation.

(27)

Las dos clases de neveras se inventaron …

 

the two kinds of fridges CL invented.3pl

 

‘The two kinds of fridges were invented…’

The morphosyntactic DP structure corresponding to the examples in (26) and (27) is the maximal one, containing both Number and Determiner, as represented in (20b) and with the semantic type in (20b′).

  • 10 See also Aguilar-Guevara and Zwarts (2010) for a kind lifting rule that shifts transitive verbs (or (...)

31Let us now consider the problem posed by the data in (25a-b). It is obvious that an ⟨eo-type object denotation cannot combine in a straightforward way with a k-level predicate like estar en peligro de extinción to be on the verge of extinction”, unless a semantic operation of type shifting guarantees that the expression denoting individual entities (i.e. esta ballena, dos ballenas) is turned into a class-denoting expression, namely a subkind. Provisionally, we propose the operation in (28)10.

(28)

Type-shifting to subkinds

 

eoesk

 

Condition: when an individual object is the semantic argument corresponding to a selected k-argument of a k-level or an i-level predicate.

  • 11 It seems that there is an additional puzzle that appears to be related to the presence of a demonst (...)

This type-shifting operation, conceived as a last resort process, guarantees an appropriate semantic composition for the sentences in (25a,b). It also predicts that an example such as (29) only allows a subkind interpretation, even in the case the speaker holds a specific iPod computer in his hand.11

(29)

Steve Jobs inventó este iPod.

 

Steve Jobs invented this iPod

 

‘Steve jobs invented this iPod.’

The operation stated in (28) predicts the possibility of obtaining a subkind reading when a Number projection, either singular or plural, is overt and a Determiner projection, either an (in)definite or a demonstrative, is also overt, but the semantics of the DP cannot combine with the selecting requirements of the VP predicate.

32The conclusion to be drawn is that both subkinds and kinds require a D projection, a definite D in the case of kinds and an (in)definite D in the case of subkinds. But, in addition, subkinds are built on Number, either singular or plural.

33The final contrast we would like to point out for a Romance language like S is that definite kinds must be distinguished from maximal sums or maximal sets, as has already been claimed for French (Beyssade 2005). Let us consider (30).

(30)

a. La ballena está en peligro de extinción. kind

 

the whale is on verge of extinction

 

‘All whales are on the verge of extinction.’

 

las ballenas están en peligro de extinción. maximal sum

 

all the whales are on verge of extinction

 

‘All whales are on the verge of extinction.’

The definite DP in subject position in (30a) corresponds to a definite kind. It denotes neither an individual atomic reference nor a sum of atoms; it denotes the name of a kind, the name of the property denoted by the Noun ballena whale”. By contrast, the definite plural DP in subject position in (30b) is ambiguous: it may either refer to the maximal sum of individuals which instantiate the kind, or the maximal sum of subkinds of the kind ballena whale”. This last claim suggests a possible alternative to our account in (28) for subkind interpretation. In particular, it could be that Number is responsible for two possible outputs: either properties of objects, as represented in (19), or properties of subkinds, as represented in (31) (cf. Zamparelli 1995), thus suggesting that Number is ambiguous.

(31)

a.

Num[–PL] N = λPλxkλysk [P(xk) R(P(xsk), P(xk))]

esk,t properties of subkinds

 

b.

Num[+PL] N = λPλxkλysk [P(xk) R(*P(xsk), P(xk))]

 

Disambiguation between (19) and (31) will come up at the time of meaning composition, when the requirements of the predicate will be attended.

5. Definite kinds in Russian

34In this section, we show how our analysis extends to R. As we argue below, in the absence of an overt determiner, R has the same type of defnite kind-referring expressions as S, both from a syntactic and a semantic viewpoint.

35Consider first the set of representative examples in (32). Just like in S, if what looks like a bare subject drontdodo” is selected by either a k-level (32a) or an i-level (32b) predicate, its interpretation corresponds to a kind denoting entity. But reference to an individual object is allowed with an i-level predicate, as shown in (32b), and required with an s-level predicate, as in (32c).

(32)

a.

Dront ischez s lica zemli v XVII veke. k-level;

 

 

dodo disappeared from surface of.earth in XVII century [DP D [NP N]]

 

 

‘The dodo was extinct in the XVII century.’

 

b.

Dront ne umel letat.’ i-level;

 

 

dodo not know.pst fly [DP D [NP N ]] or

 

 

‘The dodo could not fly.’ [DP D [NumP Num [NP N ]]

 

c.

Dront byl raschlenen v muzee Ashmola. s-level;

 

 

dodo was dissected in museum of.Ashmol [DP D [NumP Num [NP N ]]

 

 

‘The dodo was dissected in the Ashmolean museum.’

These two interpretations of the subject (i.e. ek and eo) follow from our assumptions laid out in Section 3, especially from the syntax-semantics mappings of the nominal expressions in combination with the typology of predicates.

36Next, we argue that kind-referring subjects in R involve a null Determiner, which encodes the ι operator in the examples in (32). In making this claim we rely on previous work by Pereltsvaig (2006, 2007) who provides convincing arguments in favour of the Universal DP hypothesis by showing that both DP structures and QP/NP structures are available in R and the difference between them is syntactically and semantically motivated. In this paper, we follow Pereltsvaig in assuming that the DP structure is available in R, even though no overt article is used to realize the D projection.

37Among the diagnostics proposed by Pereltsvaig (2006, 2007) to distinguish between NP and DP structures in R are the following: (i) DPs, but not NPs, allow for the expression of specificity; (ii) DPs, but not NPs, allow for the expression of quantity; (iii) DPs, but not NPs, control PRO, anaphora and trigger agreement.

38We use one of these diagnostics, namely, pronoun and anaphora control, to show that kind-referring subjects have a DP structure. As shown in (33) below, kind-referring subjects can be antecedents for personal and reflexive pronouns:

(33)

a.

Panda nahodit’sja na grani ischeznovenija.

 

 

panda is.found on verge extinction.gen

 

 

Ona javljaetsja oficial’nym simvolom vsemirnogo fonda dikoj prirody.

 

 

she appears official symbol world fund wild nature

 

 

‘The panda is on the verge of extinction. It is the official symbol of WWF.’

 

 

Slon skoro budet zanesen v Krasnuyu Knigu esli

 

 

elephant soon will listed in red book if

 

 

na nego ne perestanut ohotit’sja.

 

 

on him not stop hunt

 

 

‘The elephant will soon be listed in the IUCN Red List if people don’t stop hunting it.’

 

c.

Dront ischez s lica zemli potomu chto ne mog

 

 

dodo disappeared from surface of.earth because that not could

 

 

zaschitit’ sebja ot napadenija.

 

 

protect self from attacks

 

 

‘The dodo was extinct because it could not protect itself from being attacked.’

In contrast, a property-type BN, even when morphologically specified for plural number, does not allow for pronominal reference:

(34)

Ja budu ballotirovat’sja v presidenty. *Ih/*Ego vybirajut raz v shest’ let.

 

I will run for presidents. *They/Him elect once in six years.

 

‘I will run for president. The president is elected once in six years’

We take examples in (33) to support the DP status of kind-referring subjects. We therefore claim that the semantic representation of the kind-referring subjects in (32) and (33) involves the ι operator (i.e. a covert definite article) and its syntactic representation involves a DP projection, just like in the case of definite kinds in S (see (17) above, repeated in (35)):

(35)

a.

[DP D [NP N ]]

 

b.

D N = λPιxk[P(xk)] b′. ek kind denotation

Having argued for the presence of D, we now turn to illustrate that there is no Number involved in kind-referring expressions. Supporting evidence for the absence of Number in the denotation of kinds is given in the following example, which illustrates that numerical expressions are incompatible with definite kinds:

(36)

*Dva dronta ischezli v XVII veke.

 

two dodos disappeared in XVII century

 

‘??Two dodos disappeared in the XVII century’

This example illustrates that if Number is imposed on nominal expressions that refer to species, the output is ill-formed, even when a taxonomic reading is intended. Hence, we conclude, that a kind-referring expression is incompatible with number, both syntactically and semantically.

39Note, however, that in R it is possible to use an overt demonstrative pronoun to modify a kind-referring subject, as in (37a). Such an example only allows a taxonomic (subkind) interpretation, similar to (37b), which involves the lexical noun vidtype”:

(37)

a.

Etot dront ischez v XVII veke.

 

 

this.sg dodo disappeared in XVII century

 

 

‘This dodo disappeared in the XVII century.’

 

b.

Etot vid dronta ischez v XVII veke.

 

 

this.sg type dodo.gen.sg disappeared in XVII century

 

 

‘This type of dodo disappeared in the XVII century’

Unlike in S, this taxonomic interpretation is not available with numerals in R (cf. the ungrammatical example in (36)). Numerals in this language must pattern with lexical items such as vid, tip, class type, kind, class”, which cannot directly combine with a kind-referring expression, as shown in (38c). For S, this last phenomenon was illustrated in example (26). However, the lexical items referring to classes in R occur freely with demonstrative pronouns and numerals, as shown in (38a) and (38b).

(38)

a.

Etot vid kita/kitov ischez s lica zemli

 

 

this.sg type. sg whale.gen.sg/gen.pl disappeared from surface of.earth

 

 

sovsem nedavno.

 

 

just recently

 

 

‘This type of whale has become extinct only recently’

 

b.

Dva vida kitov ischezli s lica zemli

 

 

two type. gen.pl whale. gen.pl disappear from surface of.earth

 

 

sovsem nedavno.

 

 

just recently

 

 

‘Two types of whales has become extinct only recently’

 

c.

*Vid kita ischez s lica zemli sovsem nedavno.

 

 

type whale disappeared from surface of.earth just recently

In our analysis, a subkind interpretation is built on Number, either singular or plural, as the examples in (37) and (38a,b) properly illustrate for R. In these examples, we see the overt number specification on the noun in Genitive case.

40It is interesting to observe that in more traditional analyses of R nominals, demonstrative pronouns, as well as quantifiers and numerals, are taken to be “actualizers”, i.e. the elements that, in the absence of articles, actualize a common noun or, in other words, the elements that indicate a referential status of a common noun (Padučeva 1985).

41We interpret these elements, actualizers, as functions that are “parasitic” on number. According to our hypothesis, number corresponds to a semantic function that maps properties of kinds to properties of individuals. If actualizers are referential indicators, they can appear only in those configurations where number is involved, because they indicate the referential status of individual objects.

42Crucial to our argumentation is the fact that definite kinds are incompatible with the presence of overt morpho-syntactic actualizers of any type. We interpret this fact as supporting our conclusion that all actualizers rely on the presence of Number in morphology, syntax and semantics.

43Finally, we would like to сonsider the contrast in (39):

(39)

a.

Kit nahoditsja pod ugrozoj ischeznovenija kind (ι)

 

 

whale is.found under danger of.extinction

 

 

‘The whale is in danger of extinction.’

 

b.

Vse kity nahodjatsja pod ugrozoj ischeznovenija maximal sum

 

 

all whale.pl are.found under danger of.extinction

 

 

‘All the whales are in danger of extinction.’

 

c.

kity nahodjatsja pod ugrozoj ischeznovenija kind ()

 

 

whale.pl are.found under danger of.extinction

 

 

‘Whales are in danger of extinction.’

The subject in (39a) corresponds to what we have called a definite kind and is parallel to the S example in (30a). In the case of (39b), the plural subject shows the same type of ambiguity as in the corresponding S example in (30b). In particular, vse kity all whales” can either refer to a maximal sum of individuals instantiating the whale kind or the maximal sum of subkinds of whales (the blue whale, the white whale, etc.). Finally, (39c) contains a bare plural in subject position, which does not have any correspondence in S, but correlates with a mass-like interpretation of bare plural kind-referring expressions, like in E.

44We take (39a) and (39c) to indicate that languages can have two ways to refer to kinds: either by means of definite kinds, which have been the object of our analysis in this paper, or by means of bare plurals, a discussion of which is outside the scope of this paper. It seems that R, like E, can employ both means, whereas in S, only definite kinds are possible. Note that, according to the nominal mapping parameter of Chierchia (1998), both E and R belong to the type of [+arg, +pred] languages, whereas S, as well all the other Romance languages, is of the type [–arg, +pred]. We think that Chierchia’s typology, at least at first sight, correlates nicely with the possible ways of referring to kinds that a language allows for. Assuming that bare plurals are derived by Chierchia’s operator, it follows that [–arg, +pred] languages cannot have bare plurals in argument positions unless a D is applied. Both R and E, however, have an option of referring to kinds by means of bare plurals. We find this observation very interesting but we will have to elaborate on this idea in future research.

6. Conclusions and predictions

45Our proposal, based on the general assumption that the difference between object entities ⟨eo and kind entities ⟨ek relies on the presence or absence of Number in the structure of nominal expressions, has several consequences.

46Our first prediction is that definite kinds cannot combine with predicates that encode plurality, like, for instance, gather, be numerous, etc. Chierchia (1998:381) make a similar observation, but his explanation of this phenomena is different from ours and relies on the presence of a MASS operator (see (8) above).

(40)

a.

*The tiger gathered in the jungle

 

b.

*El tigre se reunió en la selva

 

 

the tiger cl gathered in the jungle

 

c.

*Tigr sobralsja v dzungljah

 

 

tiger gathered.refl in jungle.pl

Our explanation of this fact relies on the absence of Number in the syntactic and semantic representation of definite kinds. In the absence of a projection that encodes Number, the subject cannot agree with a predicate that requires plurality. In contrast to numberless definite kinds, plural definite DPs, which, according to our proposal, denote maximal sums of individuals, are fine in this position:

(41)

a.

The tigers gather in the jungle.

 

b.

Los tigres se reúnen en la selva.

 

 

the.pl tiger.pl CL gather in the jungle

 

c.

Vse tigry sobirajutsja v dzungljah.

 

 

All tiger.pl gather.refl in jungle.pl

The second consequence of our analysis is that definite kinds cannot trigger a generic interpretation with s-level predicates, because they are not selected by this type of predicates:

(42)

a.

The tiger ate in the jungle

 

b.

El tigre comió en la selva.

 

 

the tiger ate in the jungle

 

c.

Tigr poel v dzungljah.

 

 

tiger ate.perf in jungle.pl

The subject of the three examples in (42) can only have an individual reference. The contrast between (42) and (43) shows that a generic (or habitual) interpretation with a s-level predicate is independent of the definite DP in subject position, but is related to the presence of a generic operator instantiated by means of the present tense:

(43)

a.

The

tiger

eats

in

the

jungle

 

b.

El

tigre

come

en

la

selva

 

 

the

tiger

eats

in

the

jungle

 

c.

Tigr

est

v

dzungljah

 

 

tiger

eat.imperf

in

jungle.pl

Our third prediction is that definite kinds and definite plurals have different interpretations in the languages that we have considered. Definite plurals are not to be interpreted as denoting kinds, but the maximal sum of individuals that satisfy the property denoted by the noun (Beyssade 2005).

47The fourth consequence of our analysis is that we expect all languages that have Determiners (null or overt) to have definite kinds. However, as we noted at the end of the previous section, nothing prevents languages from using other means to refer to kinds, as, for instance, bare plurals in E (Carlson 1980, Chierchia 1998) and in R. Our hypothesis is that the availability of bare plurals depends on the semantic status of the nominal expressions in a language, in accordance with the Nominal Mapping Parameter (Chierchia 1998).

48In particular, languages whose nominal expressions are [–arg, +pred], like S, always require a Determiner (null or overt) to realize a semantic argument, and definite kinds are one sort of semantic argument. On the other hand, languages whose nominal expressions are [+arg, +pred], like E and R, in addition to definite kinds, also allow for bare plurals as kind arguments, because they can derive an argument interpretation by means of the operator. This means that Germanic and Slavic languages show two means to express kinds in the nominal domain: the ι operator, that corresponds to D, and the  operator, that corresponds to Num[+PL], whereas Romance languages can only rely on the ι operator.

49Finally, it should be noted that our proposal comes closer to the analysis of Dayal (2004) in treating the definite article uniformly as a ι operator and in assuming that there are two ways to refer to kinds, by means of ι or by means of . However, in Dayal's approach, definite kinds are singular, whereas we argue that they are numberless.  Dayal takes definite kinds to be syntactically singular, although conceptually based on pluralities. In our approach, definite kinds are the basic entities derived by a direct application of the ι operator to the property of kind denoted by a common noun. There is no plurality involved in the denotation of definite kinds, neither conceptually nor grammatically.

50Another crucial difference is that in order to derive a subkind (taxonomic) interpretation of nominal expressions, Dayal relies on the hypothesis that “common nouns […] systematically denote properties of ordinary individuals or properties of sub-kinds” (p.424). We do not derive the taxonomic reading by building an ambiguity into the denotation of the common noun or into the determiner: the presence or absence of Num does this work.

51To match the syntax and semantics of nominal expressions, we commit to stricter mapping rules. In particular, if the interpretation requires the ι operator, we assume that it is reflected in syntax by the presence of D projection. Thus, unlike Dayal, we argue that definite kinds in R do involve a null determiner and that Number is independent of D.

52As for crosslinguistic differences, Dayal attributes the difference between E and Romance languages to different lexicalization patterns. In particular, in E the operator is not lexicalized, whereas in Romance languages the definite article is said to lexicalize both the ι and the operators. In contrast, we attribute the difference between E and Romance languages to Chierchia’s typology of nominal reference, which blocks the application of the operator in Romance altogether. According to our analysis, in S kind nominal expressions corresponding to E bare plurals are absent, and definite plurals do not refer to kinds but to the maximal sum of individuals satisfying the property denoted by the noun. By contrast, in R neither nor ι are lexicalized, but there is syntactic and semantic evidence of the existence of these operators.

53As a final note, it seems reasonable to suggest that if a language has more than one way to refer to kinds, like E or R, only one of the possible constructions would be the unmarked option. In E and R, this is what bare plurals do, whereas definite kinds are pragmatically restricted. However, in S and in other Romance languages definite kinds are the only option to refer to kinds, therefore they are not pragmatically restricted.

Haut de page

Bibliographie

Des DOI (Digital Object Identifier) sont automatiquement ajoutés aux références par Bilbo, l'outil d'annotation bibliographique d'OpenEdition.
Les utilisateurs des institutions abonnées à l'un des programmes freemium d'OpenEdition peuvent télécharger les références bibliographiques pour lesquelles Bilbo a trouvé un DOI.
Format
APA
MLA
Chicago
Le service d'export bibliographique est disponible pour les institutions qui ont souscrit à un des programmes freemium d'OpenEdition.
Si vous souhaitez que votre institution souscrive à l'un des programmes freemium d'OpenEdition et bénéficie de ses services, écrivez à : access@openedition.org.

Aguilar-Guevara, Anna; Zwarts, Joost (2010). Weak definites and reference to kinds. Proceedings of SALT 20: 179-196.

Beyssade, Claire (2005). Les définis génériques en français: noms d’espèces ou sommes maximales. In Dobrovie-Sorin (2005): 33-63.

Carlson, Greg (1977). Reference to kinds in English. PhD dissertation. University of Massachusetts at Amherst. Published in New York: Garland. 1980.

Carlson, Greg; Pelletier, Francis J. (eds.) (1995). The Generic Book. Chicago: The University of Chicago Press.

Chierchia, Gennaro (1998). Reference to kinds across languages. Natural Language Semantics 6: 339-405.

Dayal, Veneeta (2004). Number marking and (in)definiteness in kind terms. Linguistics and Philosophy 27-4: 393-450.

Format
APA
MLA
Chicago
Le service d'export bibliographique est disponible pour les institutions qui ont souscrit à un des programmes freemium d'OpenEdition.
Si vous souhaitez que votre institution souscrive à l'un des programmes freemium d'OpenEdition et bénéficie de ses services, écrivez à : access@openedition.org.

Déprez, Viviane (2005). Morphological number, semantic number and bare nouns. Lingua 115: 957-883.
DOI : 10.1016/j.lingua.2004.01.006

Dobrovie-Sorin, Carmen (ed.) (1995). Noms nus et généricité. Paris: Presses Universitaires de Vincennes.

Dobrovie-Sorin, Carmen; Bleam, Tonia; Espinal, M. Teresa (2006). Bare nouns, number and types of incorporation. In Vogeleer & Tasmowski (2006): 51-79.

Dobrovie-Sorin, Carmen; Pires de Oliveira, Roberta (2007). Reference to kinds in Brazilian Portuguese: bare singulars vs definite singulars. Paper presented at Sinn und Bedeutung, Oslo.

Doron, Edit (2003). Bare singular reference to kinds. Proceedings of SALT 13: 73-90.

Format
APA
MLA
Chicago
Le service d'export bibliographique est disponible pour les institutions qui ont souscrit à un des programmes freemium d'OpenEdition.
Si vous souhaitez que votre institution souscrive à l'un des programmes freemium d'OpenEdition et bénéficie de ses services, écrivez à : access@openedition.org.

Espinal, M.Teresa (2010). Bare nominals in Catalan and Spanish. Their structure and meaning. Lingua 120: 984-1009.
DOI : 10.1016/j.lingua.2009.06.002

Espinal, M.Teresa; McNally, Louise (2007). Bare singulars: Variation at the syntax-semantics interface. Paper presented at the Workshop on Bare nouns and nominalizations. Stuttgart: University of Stuttgart.

Format
APA
MLA
Chicago
Le service d'export bibliographique est disponible pour les institutions qui ont souscrit à un des programmes freemium d'OpenEdition.
Si vous souhaitez que votre institution souscrive à l'un des programmes freemium d'OpenEdition et bénéficie de ses services, écrivez à : access@openedition.org.

Espinal, M.Teresa; McNally, Louise (2011). Bare nominals and incorporating verbs in Catalan and Spanish. Journal of Linguistics 47: 87-128.
DOI : 10.1017/S0022226710000228

Format
APA
MLA
Chicago
Le service d'export bibliographique est disponible pour les institutions qui ont souscrit à un des programmes freemium d'OpenEdition.
Si vous souhaitez que votre institution souscrive à l'un des programmes freemium d'OpenEdition et bénéficie de ses services, écrivez à : access@openedition.org.

Farkas, Donka; de Swart, Henriëtte (2007). Article choice in plural generics. Lingua 117: 1657-1676.
DOI : 10.1016/j.lingua.2006.06.011

Format
APA
MLA
Chicago
Le service d'export bibliographique est disponible pour les institutions qui ont souscrit à un des programmes freemium d'OpenEdition.
Si vous souhaitez que votre institution souscrive à l'un des programmes freemium d'OpenEdition et bénéficie de ses services, écrivez à : access@openedition.org.

Jespersen, Otto (1927). A modern English grammar. Heidelberg: Carl Winter’s Universitätbuchhanlung.
DOI : 10.1086/435626

Krifka, Manfred (2004). Kinds of Kind Reference: Bare Plurals—Ambiguous or Not? Proceedings of SALT 13: 180-203.

Krifka, Manfred; Pelletier, Francis Jeffry; Carlson, Greg; ter Meulen, Alice; Chierchia, Gennaro; Link, Godehard (1995). Introduction. In Carlson & Pelletier (1995): 1-124.

Landman, Fred (1989). Groups, I. Linguistics and Philosophy 12/5: 559-605.

Longobardi, Giuseppe (1994). Reference and proper names. Linguistic Inquiry 25: 609-665.

Longobardi, Giuseppe (2001). How Comparative is Semantics? A unified parametric theory of bare nouns and proper names. Natural Language Semantics 9: 335-369.

Format
APA
MLA
Chicago
Le service d'export bibliographique est disponible pour les institutions qui ont souscrit à un des programmes freemium d'OpenEdition.
Si vous souhaitez que votre institution souscrive à l'un des programmes freemium d'OpenEdition et bénéficie de ses services, écrivez à : access@openedition.org.

Munn, Alan; Schmitt, Cristina (2005). Number and indefinites. Lingua 115: 821–855.
DOI : 10.1016/j.lingua.2004.01.007

de Swart, Henriëtte; Zwarts, Joost (2009). Less form—more meaning: why bare singular nouns are special. Lingua 119/2: 280-295.

Padučeva, Eleva Viktorovna (1985). Vyskazyvanije i ego sootnesennost’s dejstvitel’nostju. Moskva: Nauka.

Format
APA
MLA
Chicago
Le service d'export bibliographique est disponible pour les institutions qui ont souscrit à un des programmes freemium d'OpenEdition.
Si vous souhaitez que votre institution souscrive à l'un des programmes freemium d'OpenEdition et bénéficie de ses services, écrivez à : access@openedition.org.

Pereltsvaig, Asya (2006). Small nominals. Natural language and Linguistic Theory 24/2: 433-500.
DOI : 10.1007/s11049-005-3820-z

Format
APA
MLA
Chicago
Le service d'export bibliographique est disponible pour les institutions qui ont souscrit à un des programmes freemium d'OpenEdition.
Si vous souhaitez que votre institution souscrive à l'un des programmes freemium d'OpenEdition et bénéficie de ses services, écrivez à : access@openedition.org.

Pereltsvaig, Asya (2007). On the universality of DP: A view from Russian. Studia Linguistica 61/1: 5-38.
DOI : 10.1111/j.1467-9582.2007.00129.x

Vogeleer, Svetlana; Tasmowski, Liliane (eds.) (2006). Non-definiteness and plurality. Amsterdam: John Benjamins, Linguistics Today.

Zamparelli, Roberto (1995). Layers in the determiner phrase. PhD Dissertation. University of Rochester. Published in New York: Garland. 2000.

Haut de page

Notes

1 This research has been funded by the following two research grants awarded by the Spanish Ministerio de Ciencia e Innovación (HUM 2006-13295-C02-02, FFI2011-23356), and by a grant awarded by the Generalitat de Catalunya to the Centre de Linguística Teòrica (2009SGR-1073). The first author also acknowledges a Ramón y Cajal award (RyC-2008-02856), and the second author an ICREA Acadèmia award.

2 Krifka actually notices that the group semantics might explain the use of definite generics in sentences like The American customer bought 74.000 BMWs last year or The rat reached Australia in 1770. In The Generic book these examples are analyzed as collective predications which attribute important properties of members of the group to the group, although we don’t see how this interpretation would work for the first example (i.e., members of the group of American customers bought 74.000 BMWs????). Chierchia (1998:379) discusses a similar example (??The tiger is roaring in the zoo), where he comes up with a context to save the example.

3 The prediction of Chierchia’s analysis with respect to bare plurals are not discussed in this paper.

4 There is a nice parallel in Russian for this type of examples. Consider the contrast in (i):
(i) a.
U nih svinoj gripp
at them swine.sg flue.sg
‘They have swine flue’
b. *U nih svinye grippy
at them swine.pl flue.pl
The interpretation of (ia) is that for each individual referred to by the pronoun, the property of having swine flue applies. Normally, this distributive interpretation requires plural marking, as shown in (ii). However, since svinoj gripp denotes a property of kinds but not of individual objects, plural marking in (ib) is not possible.
(ii) a. U nih novaja mashina
at them new.sg car.sg
‘They have a new car.’ (one for all, collective interpretation)
b. U nih novye mashiny
at them new.pl car.pl
‘They have new cars.’ (one car per person, i.e. distributive interpretation)

5 As will soon become clear, the prediction we make here is that the plural counterpart of (15a) can only get a subkind interpretation.

6 As an anonymous reviewer pointed out, it is not obvious that the definite is not ambiguous if it can apply to both properties and proper names. However, in Partee’s (1987) model, proper names can be type-shifted to property-type expressions before the ι operator applies.

7 Note that (23b) is ambiguous between the definite kind reading, associated with a DP structure that does not project Number, and a definite singular plus a taxonomic reading, the latter two associated with a DP structure that projects Number. We will come back to the taxonomic reading below.

8 See also Carslon (1977/1980:26) for a specific reference to English names of kinds of things.

9 If the subject of this sentence is modified by a relative clause, the example is grammatical with a subkind interpretation, just like (26a-b). This is just an additional piece of evidence that shows that relative clauses, like numerals and demonstratives, need more functional structure above the noun.

10 See also Aguilar-Guevara and Zwarts (2010) for a kind lifting rule that shifts transitive verbs (or verb-preposition combinations) from an object-level meaning to a kind-level meaning, although we do not share with these authors the hypothesis that weak definites refer to kinds.

11 It seems that there is an additional puzzle that appears to be related to the presence of a demonstrative. In examples like (29) an additional interpretation is available: the one that refers to representative instantiations of the subkind. This observation is also true for Russian demonstrative determiners, discussed in the next section.

Haut de page

Table des illustrations

URL http://rlv.revues.org/docannexe/image/2104/img-1.jpg
Fichier image/jpeg, 12k
URL http://rlv.revues.org/docannexe/image/2104/img-2.jpg
Fichier image/jpeg, 8,0k
URL http://rlv.revues.org/docannexe/image/2104/img-3.jpg
Fichier image/jpeg, 16k
Haut de page

Pour citer cet article

Référence électronique

Olga Borik et M. Teresa Espinal, « On definite kinds », Recherches linguistiques de Vincennes [En ligne], 41 | 2012, mis en ligne le 01 septembre 2014, consulté le 13 février 2016. URL : http://rlv.revues.org/2104

Haut de page

Auteurs

Olga Borik

Departament de Filologia Catalana / Centre de Lingüística Teòrica Universitat Autònoma de Barcelona

M. Teresa Espinal

Departament de Filologia Catalana / Centre de Lingüística Teòrica Universitat Autònoma de Barcelona

Haut de page

Droits d’auteur

© Presses universitaires de Vincennes

Haut de page