1 SUBJECTHOOD OF QUIRKY SUBJECTS AND GF-SPLIT KATARZYNA MIECHOWICZ-MATHIASEN

In this paper we show that most of the adduced intuitions about subjects and subjecthood are true for Nominative subjects and depend exactly on this Case, i.e. a subject is bound to have all, or most, of the subject properties only when it stands in Nominative. The problems with subjecthood properties arise, however, once the potential subject does not bear this apparently special Case, but is inherently Dative or Accusative, i.e. "Quirky". We argue that Quirky Subjects undeniably bear subjecthood and we support this view with relevant data. We also propose parameter settings for the analysed languages and eventually a definition of subject based on Ura's (2000) analysis of Grammatical Function Split phenomena (hence GF-Split). 1. Defining subject and subjecthood. Even though subjecthood has quite a long history in linguistics, it cannot really boast a satisfying definition. The existing definitions concentrate on the following three areas: form (formal, or morphological, features of a subject like agreement features and Case); meaning, i.e. theta-role; syntactic position, i.e. where the subject usually occurs in sentences. Falk (2003: 1-2), following Keenan (1976) and Andrews (1985), enumerates various standardly assumed subjecthood properties; we present them below. If a verb has an Agent argument, in the active voice the Agent is realised as subject: (1a) John ate an apple