The Semantics of Determiners

1.2 Types of Ds
Let us turn now to the DPs on the rightmost branch in Figure 1. Syntactically, their NP component is headed by a common noun, and they have a determiner in their D position. As mentioned above, crucial to the semantic fate of these DPs is the type of determiner they involve. There is a whole subfield of linguistic semantics and philosophy of language devoted to the study of the semantics of determiners. In this chapter we do not attempt to do justice to the many theoretical issues and frameworks advanced in this area but rather concentrate on minimal tools needed to describe the basic properties of determiners in Romanian.

As a starting point, let us consider the preliminary classification of determiners in Figure 2:

universal existential

fiecare ‘every’
orice ‘any’
definite indefinite

unmarked dem unmarked marked

-(u)l ‘the’ acest ‘this’ un ‘a’
partitive free choice negative …
un-ul ‘one of’ vreun ‘some’ nici un ‘no’

Figure 2: Typology of determiners

The first division, that between universal and existential determiners is one whose justification we will get to below. For now, it is enough to note that the examples in (3)

(3) a. Fiecare student a plecat.

every student has left
‘Every student left.’
b. Poţi invita orice student.
may.II invite any student
‘You may invite any student.’

have a universal flavor contributed by the underlined Ds in the sense that in order for each sentence to make a true claim, the properties of each student (within, most likely, a contextually restricted set) are relevant. This is not so in the case of the other Ds, where what matters as far as truth conditions are concerned is the existence of an appropriate entity. In (4), for instance, where an unmarked indefinite is used, all that is needed in order for this claim to be true is the existence of an entity with the right properties (in this case, an entity that is a student and that left early).

(4) Un student a plecat devreme.
a student has left early
‘A student left early.’

All the other existential determiners exemplified in Figure 2 can (and it will be argued here, should) be treated essentially as existentials with extra requirements.

A potential problem for the claim just made is the negative indefinite nici un, exemplified in (5):

(5) a. Nici un student nu a plecat devreme.

Neg a student not has left early
‘No student has left early.’
b. Nu există nici un cîine cu două cozi.
not exist Neg a dog with two tail.Pl
‘There are now dogs with two tails.’

One could argue that nici un ‘Neg. a’ is a universal since in order for (5a) to be true it must be the case that every student has the property of not having left. Note, however, that the claim (5a) makes can also be paraphrased by a statement that contains an existential ‘within the scope’ of negation: it is not the case that there is a student who left early. Given that this complex determiner is made up of a negative morpheme nici and the unmarked indefinite article un/-o, the second paraphrase appears more appropriate. Note also that it is appropriate to use nici un in case one denies the existence of a particular type of entity, as in (5b).

Before turning to the next subsection, in which we provide some theoretical tools, we deal with a final introductory issue, that of bare nominals. Romanian has nominal CN-headed arguments that may appear without any determiner whatsoever, as exemplified in (6):

(6) a. Ileana poartă mereu fustă.

Ileana wears always skirt
‘Ileana always wears a skirt.
b. Maria învaţă limbi străine.
Maria learns language.Pl foreign.Pl
‘Maria is learning foreign languages.’

Such nominals are called ‘bare’ and we see that Romanian, unlike English, allows not only bare plurals, exemplified in (6b), but bare singulars as well, exemplified in (6a). A current controversy concerning these types of nominals is whether they are in fact full-fledged DPs with a null D or whether they are what they appear to be, NPs with no D projections. (See, among many others, Chierchia 1998, Longobardi 2002, Dobrovie-Sorin and Laca 2002). On the empirical side, what is obvious is that ‘existential’ bare plurals in Romanian are similar to their counterparts in other languages in that their distribution is freer than that of bare singulars, and their interpretation is closer to their unmarked indefinite counterparts⁴. The plural indefinite article nişte (invariant in gender) is used much less commonly than the singular indefinite article un/o. The data from Romanian is consistent with data from other languages that have articles but allow both bare singulars and bare plurals (such as Hungarian) in that it points to the necessity of a three way distinction: (i) DPs with overt determiners; (ii) bare plurals, and (iii) bare singulars. In Farkas and de Swart (2003) it is suggested that these nominals differ at the syntactic level with respect to how much functional projection they involve, as illustrated in (7):

(7) a. DP b. DP

D NP D NumP

un nişte
a N some.Pl [Pl] NP
copil
child N
copii
child.Pl

c. NP d. NumP

N [Pl] NP

copil
child N
copii
child.Pl
Under this proposal, there is a two-way distinction between full DPs, whether singular or plural, (in 7a,b) and nominals that do not project a DP (in 7c, d). Within the latter, there is a further two way distinction between nominals without any functional projection (the bare singular in (7c)), and nominals with a functional number projection.⁵ Under this view at least some syntactic positions must not be sensitive to the finer distinctions between the nominals in (7) while some are. The syntactic distinctions, Farkas and de Swart suggest, are relevant to semantic and distributional restrictions that we will come back below. Alternatively, one may assume full functional projections with D nodes across the board for both (7c) and (7d) or only for (7d).

For the semanticist, the crucial question is how bare nominals differ in interpretation and distribution from DPs with the Ds in Figure 2, and how bare singulars differ from bare plurals in languages like Romanian, which allow both. In order to fully understand the issues involved one has to go into details concerning the interpretation of number, genericity and considerations governing the presence or absence of an article. Pursuing these matters would take us beyond the limits of this chapter and therefore we leave them outside its scope.