ELTE BTK Magyar Nyelvtudományi és Finnugor Intézet

Miljan, Merilin (Tartu)

nyomtatható változat

Underspecification and the pragmatics of Finnic grammatical cases

In Finnic, the syntactic arguments can be expressed by three different forms: nominative, genitive, and partitive. In Estonian, for example, the subject is normally signalled by the nominative and can also be marked by the partitive case. The object function can be expressed by three different forms: genitive, partitive, and nominative. The variation in cases is associated with partial affectedness (1a) and total affectedness (1b), or aspect. Partitive case is thus related to imperfectivity (1a) and ‘indeterminate quantity’ (2a), while genitive is seen to express perfectivity (1b) and ‘determinate quantity’ (2b).

  a.   Raul     ehitas               suvila-t.
      R.nom build.past.3sg cottage-part.sg
      ‘Raul was building a cottage.’

        b.   Raul    ehitas                suvila.
      R.nom build.past.3sg cottage.gen.sg
      ‘Raul built a cottage.’

(2)  a.   Ostsin            leiba.
     buy.past.1sg bread.part.sg
     ‘I bought (some) bread.’

       b.   Ostsin             leiva.
     buy.past.1sg bread.gen.sg
     ‘I bought a/the bread.’

Since grammatical case is traditionally seen as distinct from semantic case, the semantics which a particular alternating case is seen to express is derived from the construction, which then determines the assignment of a specific grammatical case. Yet a single case form (e.g. partitive) may yield other readings than those of imperfectivity and indefinite quantity in Estonian: depending on linguistic context, the interpretations may vary from generic reading and (in)alienable possession to temporality and distancing (also referred to as ‘negative polarity item’ in Kiparsky 1998; Kaiser 2003), not to mention the irregular unavailability of any (significant) meaning. The interpretation of a particular case-marked term may thus depend on many different things: the lexical semantics of the noun and its associated verb and the discourse context in which the term appears; as well as prosody and information structure in a sentence. This is hard to model using traditional or structural approach to case, and perhaps explains why these data has mostly been left unaccounted for.

In this paper I argue that it is in fact the case marker itself which provides both semantic and syntactic information, i.e. it has ‘inherent’ meaning, and the information provided by case is semantically and syntactically underspecified. Instead of treating morphological case as a surface realization of syntactic functions (or grammatical relations) and, in case of semantically conditioned alternations, matching case to the final interpretation of particular constructions, I suggest that morphological case itself functions as an aid to the establishment of the (final) interpretation. Case thus “instructs” how to interpret the noun it marks, but by providing only incomplete information (i.e. being underspecified) it needs pragmatic enrichment by linguistic context and perhaps also by non-linguistic context. The various case ‘meanings’ or no (significant) ‘meanings’ (i.e. optionality) then are the result of inferential effects.

The actual language of the intended paper is English.