Sławomir Zdziebko

The paper proposes that the phonological make-up of segments is influenced by the activity of the constraint *Hydra, which penalizes the presence of more than one headed element per one phonological expression. *Hydra influences the shape of the inventories and the phonological behaviour of nasal vowels in languages such as French and Brazilian Portuguese. At the same time, the behaviour of nasal vowels in Yoruba shows that *Hydra a violable constraint. In Polish, the high ranking of *Hydra proves necessary to account for the absence of Surface Velar Palatalization before the front nasal vowel /ɛ/. It also allows us to formulate a unified account of the 1st Velar and Anterior Palatalization, which have very different structural descriptions but take place before the same set of derivational affixes.

Gussmann (2007) put forward a morpho-phonological analysis of Polish palatalizations which is an alternative to traditional re-write rules promoting abstractness in phonology. The aim of this article is to turn Gussmann’s descriptive tool into a coherent theory. In order to do that I propose a set of premises that regulate the working of the component of morpho-phonology. Among these premises the Minimalist Hypothesis (Kaye 1992) and the Locality Principle occupy the most prominent position. The former says that all morpho-phonological replacements work whenever their conditions are met (no (counter)bleeding or counterfeeding is possible), whereas the latter limits the scope of morpho-phonological replacements to nodes entering the relation of concatenation (Embick 2010). The last part of the article is devoted to the cases that apparently violate the Minimalist Hypothesis. These are the replacements that Gussmann (2007) subsumes under the label P(alatalization) R(eplacement) 7. The article shows that the problematic replacements presented as PR7 may be convincingly analysed as root-specific and thus do not constitute counterexamples to the Minimalist Hypothesis.

Polish adnominal participles accept a wide range of event modifiers except when they are additionally modified by focus or phase particles corresponding to still. The paper argues that the semantic contribution of still is incompatible with the change-of-state component of the meaning of participles. While still presupposes that the property denoted by the participle holds over the initial proper subinterval during which the focalized state holds, the measure-of-change function found in resultative participles entails that the relevant subinterval corresponds to the change of state over which the relevant property does not hold yet. The participles modifiable by still are argued to lack the change-of-state component.