Anna Malicka-Kleparska

One of the key operations in valency rearrangement is the formation of mono-argumental predicates from phonologically corresponding/identical bi-argumental predicates. It has been most recently revisited by Junghanns, Fermann and Lenertová (2011), who analyze decausatives in Slavic languages as cases of reflexivization of verbs with non-agentive causers, in the spirit of Koontz-Garboden (2009).We review these formations in Polish and find out that an alysis which is set against more extensive data gives no grounds for a reflexive analysis. We find the data in favor of decausatives showing the presence of the external argument through the appearance of the subject ‘by itself ’ anaphor misjudged as to their grammaticality. This claim is supported with examples from the National Corpus of the Polish Language contending against Jabłońska’s (2007) analysis. The overall picture of the morphological system and language behavior speak against setting apart decausatives with reflexive marking from other unaccusatives in Polish. We disregard the reflexive analysis and adopt the anticausative solution, where the formation of decausatives is not seen as identification of arguments. We see it as a subtraction of VoiceP. This solution assumes one of the structures for decausatives from Alexiadou (2010). The operation is seen as lexical, not syntactic, and in defiance of Koontz-Garboden’s Monotonicity Hypothesis.

The paper deals with a wider problem of the representation of causative structures in the root-based generative model of morphosyntax illustrated here with the Polish causativizing morpheme roz-. Following Koontz-Garboden’s (2009) analysis of anticausative verbs, we propose that the phenomenon of causation should be separated from the introduction of the additional causer argument brought in by the voice projection. In our analysis roz- is seen as the head of the active voice projection, as opposed to roz- się, the non-active voice head. Such an analysis allows us to account for the distributional properties of roz- versus roz- się in Polish. In the analysis of the typology of roots which can serve as bases for the causative structures taking the roz- voice heads, the typology of roots developed by Embick (2009) to account for the properties of states and stative passives has been adopted, as it seems to work in the case of the roots deriving causatives. The roots appropriate for the predicates of states cannot derive the roz- causatives in Polish, while these appropriate for the predicates of events form such causatives. The analysis ties in with recent proposals in root-based research into verbal valency, and contributes to the overall model of valencyrelated derivations in root-based approaches

In this text we consider properties of stative passive participles corresponding to roz- Object Experiencer verbs in Polish. They are viewed in the light of the distinction between Davidsonian states (Davidson 1967) and Kimian states (Kim 1976). Polish statives with roz- passive participles seem to show features of both Kimian and Davidsonian states. We will consider the results of various tests proposed in the literature to discover the properties of the relevant Polish structures and offer an explanation for the areas in which roz- structures diverge from the characteristics of Kimian states.

In this paper we will present a theory on the source of prefixation differences between Polish analytic and synthetic anticausatives. Analytic anticausatives are freely prefixed with superlexical, lexical and ‘pure perfectivizer’ prefixes, while synthetic anticausatives show propensity for ‘pure perfectivizers’, if indeed they are prefixed at all. We have looked for a source of this distinction in OCS anticausative morpho-syntax. We claim that OCS analytic anticausatives are formed within the limits of the voice system of OCS as middle voice word-forms. As such, they have the same rich prefix inventory as other verbal stems that have the same roots, with some of the prefixes introducing changes in verbal lexical meaning. On the other hand, synthetic anticausatives are already at this time members of the OCS lexicon, mostly without any related verbal forms, but sharing roots with nouns and adjectives (for which prefixation is not a frequent operation in Slavic languages). The prefixes appearing with synthetic anticausatives have the function of realizing the viewpoint aspect, rather than word-formational functions. We have traced the distinction between the two classes of anticausatives from the OCS times to Present-Day Polish, quoting also some intermediate stages in the history of these verbs.