The primary goal of this paper is to provide an analysis of the demonstrative to in Serbo-Croatian when used to refer to events. To is argued to be an event pronominal with three basic functions, also exhibited by regular pronouns: deictic, anaphoric and bound-variable function. In its deictic use, to is argued to head a distinct functional projection, which projection is associated with quantification over events. In its anaphoric use, to refers to a previously mentioned event. In its bound-variable use, to is proposed to be the spell-out of the bound event pronominal, which constitutes a syntactic reflex of the semantic analysis of adverbials as predicates of events (see Davidson 1967). It is a virtue of this analysis that it can unify the three uses of the event pronominal to, explaining both the similarities and the differences. To the extent to which this is the only way to unify the three uses of to, the paper provides indirect support not only for the underlying quantification over events/states, but also for the syntacticization of certain aspects of this quantification.

When languages come into contact, they influence, not only each other's vocabularies through borrowing, but they may also influence each other's grammatical structures. In cases of "high contact" one normally observes morphological simplification, i.e., "koineization" (Trudgill's term). Conversely, one expects "complication", to accompany "low contact" or peripheralization of a dialect (Andersen 1988). The present work concerns a case of "high contact" and resulting borrowing, not between different languages, but between different styles of a single language (see Bartsch 1987: 196f). As Trudgill predicted, "high-contact situations come in many different forms, and we will not expect to find simplification in those (very many) contact situations where childhood bilingualism and second-variety acquisition are the norm. In these situations, on the contrary, we are liable, although not certain, to find intensive borrowing and interpenetration of linguistic systems, with possible resulting complication. (Trudgill 1989: 232)" [Italics are mine, RS] This, along with reinterpretation of function, is what we observe as a result of inter-stylistic borrowing of derivational models in Russian. In this case the resulting "complication" is not only quantitative (Style Y takes on a new affix from Style X ); but it is qualitative as well.

The article studies a contrast in the aspectual marking of telicity in English and Slavic languages (most examples are from Bulgarian). A solution based on a syntactic decomposition of eventive verbs into a causal subevent and a resultative state subevent is proposed. A template approach to aspectual composition is outlined. The differences in English and Slavic aspectual usage are argued to be due to the null versus overt character of the telic morpheme and its phrase structure position. An experimental study, based on this parametric difference, and investigating the competence of Slavic native speakers acquiring English is presented. Results are interpreted in the light of current theories of second language learners' access to Universal grammar.

We examine one of the traditional Russian Aktionsarten ('sposoby dejstvija'), the attenuative PRI- verb (e.g. priotkryt´ 'to open a little'). This is universally claimed to be a type of quantitative Aktionsart. However, we advance morphological, syntactic and semantic arguments against this assumption. PRI- verbs readily give secondary imperfectives, which is uncommon with true Aktionsarten. They also permit 'unselected objects' as in prisypat´ jamu 'to partly fill in a pit' (cf. *sypat´ jamu). This is never found with genuine Aktionsarten but is characteristic of a type of lexical derivation based on lexical subordination at the level of semantic representation. Finally, a careful investigation of the meaning of PRI- verbs show that they do not express quantification over their objects but instead quantify over a resultant state. This fits in well with an analysis as lexical derivation but is incompatible with current thinking on the semantics of quantitative Aktionsart.