Microvariation in Catalan existential constructions

"We cannot learn what we seek to know about that which is

invariant in grammar without studying that which is variable in it. We need linguistic diversity". Ken Hale (1997: 72)

The aim of this paper is to show that certain parametric properties allow us to differentiate between two main varieties of Catalan. Interestingly the values of the parameter studied here are active in other Romance languages: Some dialects of Catalan are similar in this respect to certain dialects of Spanish and Portuguese, whereas others may be grouped with Occitan and French.

Catalan existential constructions follow two different strategies, as shown in (1) and (2).

(1) a. Hi ha una cadira Ribagorçan Catalan

cl. has a chair

'There is a chair'

b. Hi ha tres cadires

cl. has three chairs

'There are three chairs'

c. Mos caleva una cadira

to-us was-necessary a chair

'We needed a chair'

d. Mos caleva tres cadires

to-us was-necessary three chairs

'We needed three chairs'

(2) a. Hi ha una cadira Central Catalan

cl. has a chair

'There is a chair'

b. Hi han tres cadires

cl. have three chairs

'There are three chairs'

c. Ens calia una cadira

to-us was-necessary a chair

'We needed a chair'

d. Ens calien tres cadires

to-us were-necessary three chairs

'We needed three chairs'

In Ribagorçan, no agreement is shown between the existential verb and its NP object. However Central Catalan shows agreement between the verb and its NP object in existential constructions (see Rigau (1997, 1999). This agreement is in number not in person. No nominative first, second or third person is allowed in (2): *Hi ha ella; *Ens calíeu vosaltres.

Because of their argument structure, existential verbs coappear with an oblique (locative or dative) clitic as a (quirky) subject. We will be able to account for the impersosnalizer function of quirky clitic subjects in Catalan assuming that oblique clitics have the L(ocative) property.

In parametric terms, we assume that the functional category I(nflection) can manifest the L property in its person feature or in both person and number features. As pointed out by Hale & Keyser (1998: 11f), many languages -e.g. Navajo- have 'locative' or 'areal' agreement in addition to conventional person and number agreement.

The choice of L-number or singular/plural number in impersonal (i.e. L-person) existential constructions account for the bifurcation between Catalan dialectal grammars:

Features of I: [L-person, L-number] Ribagorçan Catalan
Features of I : [L-person, singular/plural-number] Central Catalan

Our analysis of existential sentences can be extended to constructions with the impersonalizer clitic se.

(5) a. Es pot tancar les finestres Ribagorçan Catalan

cl. can/may(sing) close the windows

'The windows can be closed'

b. Es poden tancar les finestres Central Catalan

cl. can/may(plural) close the windows

'The windows can be closed'

References:

Hale, K. (1997) "Some observations on the contributions of local languages to linguistic

science", Lingua 100, 71-89.

Hale, K. & J. Keyser (1998) "There-Insertion Unaccusatives and Other Complex

Intransitives", ms. MIT, Cambridge, Mass.

Rigau, G. (1997) "Locative Sentences and Related Constructions in Catalan: ésser/haver

Alternation", in Uribe-Etxebarria, M. & A. Mendikoetxea (eds.) (1997),

Theoretical Issues at the Morphology-Syntax Interface, Bilbao/Donostia: Universidad del País Vasco, 395-421.

Rigau, G. (1999) "Relativized Impersonality: Deontic Sentences in Catalan", in Treviño,

E. & J. Lema (eds.) (1999), Semantic Issues in Romance Syntax, Amsterdam:

John Benjamin, 193-230.