Allomorphy without context specification: a case study of Czech adjectival stems

This paper investigates stem-marker allomorphy in Czech adjectives. It shows that an analysis based on the frequently used context-sensitive rules comes at the expense of having to postulate widespread accidental homophony or disjunctive rules. The paper further demonstrates that the allomorphy can be accounted for within an approach based on portmanteau realisation of features, specifically the version of Nanosyntax proposed in Starke (2018), although alternative implementations are conceivable. Along the way, we explore a fine-grained decomposition of adjectival meaning and we also discuss the implications of these observations for the general issues surrounding context-sensitive rules compared to other systems of dealing with allomorphy.