The Semantics of Determiners

3.2.2 The co-varying determiner cîte
In Romanian, the unmarked indefinite determiner can be preceded by the morpheme cîte, whose effect is to turn the DP into one whose properties are identical to the reduplicated indefinites in Hungarian discussed in Farkas (1997) under the name of ‘dependent indefinites’. Simply put, the presence of this morpheme requires the discourse referent introduced by the DP to co-vary with another discourse referent. This DP therefore must have narrow scope relative to another expression that introduces the discourse referent with which it co-varies. In (56) for instance, the cîte-marked indefinite must be interpreted as co-varying with the discourse referent introduced by the universal:

(56) Fiecare copil a citit cîte o carte.

every child has read CO a book
‘Every child read a book.’

Unlike its English translation, the Romanian sentence lacks an interpretation where the indefinite DP has wide scope relative to the universal. As you recall, in the absence of cîte, the indefinite may be interpreted as having either wide or narrow scope relative to the universal in such cases. Note that it would not be enough to require the co-varying DP to have narrow scope relative to the universal, and have (56) be translated as (57):

(57) ∀x: child(x) ∃y: book(y) read(x,y)

If the co-varying marker is used, there must be actual co-variation between books and children and therefore there must be several books involved, while (57) allows but does not require co-variation.

The distributive DP may co-vary with a situation variable as well as an individual variable, as exemplified in (58):

(58) Din cînd în cînd, cîte un student punea o întrebare.

from when in when CO a student raised a question
‘From time to time, a student raised a question.’

The truth conditions of the Romanian sentence require it to have been several students raising questions (though not necessarily always a new student).

Note now that in the absence of cîte, a sentence like (58) is ambiguous, just like its English counterpart, between an interpretation which requires there to have been a single student asking a question multiple times, and an interpretation under which multiple students asked a question, and therefore where students co-vary with times. This shows that the existence of the marked co-varying indefinite does not block the co-varying interpretation of the unmarked indefinite in Romanian. The facts are parallel in Hungarian.

Interestingly, co-varying DPs in Romanian, just like their Hungarian counterparts, cannot co-vary with a world-variable. It has proven quite useful to analyze modals as quantifying over possible worlds, the way universal determiners quantifier over individuals and adverbs of quantification quantify over situations. In modal statements, however, co-varying DPs are ungrammatical (unless some other expression provides an appropriate variable for them to co-vary with):

(59) a. *Trebuie să citeşti cîte o carte.

must SUBJ read.II CO a book
‘You must read a book.’

b. *Cîte o pisică e deşteaptă.

CO a cat is smart

The impossibility of cîte indefinites as generic might be connected to the fact that the situations quantified over by the generic operator are intensional.

We conclude then that cîte contributes a feature, [CO], requiring the discourse referent introduced by the determiner to co-vary with a situational or individual variable. This co-variation results in a discourse-plural discourse referent in Brasoveanu (2007) terminology: for each child in (58) it is enough for the child to have read a single book, but there must be several books involved when you consider the child-book pairs. We cannot see cîte, however, as simply marking discourse plurality because being in the scope of a modal or negation is not enough to ensure the licensing of such a DP. Alternatively, the discourse plural nature of these DPs may be an implicature associated with their being marked for co-variation.

The constraint imposed by cîte then concerns the relationship between the discourse referent introduced by the DP and another discourse referent. Given this analysis, we correctly predict that the use of a cîte DP results in ungrammaticality in examples that do not involve an expression that introduces an appropriate discourse referent with which the co-varying DP can co-vary with. We have already seen such an example in (59). Another is given in (60):

(60) *Maria nu a citit cîte o carte.

M. nu has read Dist a book

Negative operators do not introduce an explicit discourse referent that an indefinite may co-vary with even though an indefinite within their scope receives varying values. Treating cîte as simply a mark of discourse plurality therefore is not enough.

The analysis also correctly predicts that in examples where cîte DPs are grammatical, such as (56) and (58), the sentences will be scopally unambiguous.

Finally, note that just like in the case of the partitive, the existence of a distributive indefinite in Romanian does not block the narrow scope interpretation of an unmarked indefinite. The Romanian example in (61) is scopally just as ambiguous as its English counterpart.

(61) Fiecare copil a citit o carte.

every child has read a book
‘Every child read a book.’

Here again then we see that a feature distinguishing a subtype of indefinite does not block the use of the unmarked indefinite.