In Czech there are no articles; determination (definiteness and indefiniteness), a linguistic universal, is known to be submerged in the wider domains of discourse context, sentence intonation, word order, and lexical quantification. The following study examines the interaction of these domains in the expression of indefiniteness, with spe cial attention to word order, speaker-knowledge entailments, and the four quantifiers nejaky, jakysi, jeden, and jakykoliv. It is shown that in some environments nejaky 'a; some' approaches the status of an indefinite article. In certain word-order contexts this quantifier is obligatory with animate subjects.

Derived nominals in Russian and many other languages come in types with different interpretations, different relations to the underlying verb and different argument realizations. This paper argues for an approach to these distinctions along the following lines:

• In order to make sense of the nominalization data the first task is to distinguish between result nominals, Simple Event Nominals and Complex Event Nominals (CENs, Grimshaw 1990). This includes a discussion of the argument structure of the underlying verbs, of the derived nominals and their various argument realizations;
• The second task is to clarify the origin of different readings of CENs, like 'manner of action', 'fact', 'event'. I contest the idea that argument realization is crucial in determining the various readings (Paduceva 1980, 1984). Instead, my claim (following Vendler 1967) is that context is the only factor relevant to the interpretation;
• It is the status of the nominal as a CEN or result nominal that will in turn determine how the various arguments may be realized.

The context determination of the available reading extends to al possible noun phrases in a particular context. It turns out that in any context (i.e. reading) a CEN may occur in any argument realization.

In this paper I demonstrate the semantic, morphological, and syntactic restrictions on possessive formation in Czech. Referring to the distinctions between possessive (poss) and genitive (gen) I argue that poss are nps while gens are dps in Czech and their complementary distribution is evidence for a syntactic movement which I call Possessive Movement. I propose that the potential of n to take an argument is encoded as a weak subcategorization feature +a of n. Referring to the Unlike Feature Condition I propose that if a cannot be satisfied at LF within the smallest np domain, it is transferred as strong to the N's functional projection d. The checking of the feature a of n takes place

a) at LF at the n level by an n-complement of form dp,

b) in syntax at the d level by spec-head relation between the d and np in spec(dp).