Fluid intransitivity in Middle Welsh: Gradience, typology and ‘unaccusativity’

Abstract

Abstract Welsh philologists have long noted that non-finite clauses headed by a verbal noun display considerable variation in the marking of the notional subject of intransitive verbs. This argument is sometimes marked like a transitive agent, sometimes like a transitive object. This paper demonstrates that two systems of independent variables, the intrinsic denotational content of the intransitive subject NP and the aktionsart and control characteristics of the lexical verb, serve to constrain this variation to only a small residue. These variables are shown to be articulated one to another hierarchically, rather than paradigmatically. The resulting casemarking system is thus typologically a ‘fluid’ rather than ‘split’ intransitive system in the terminology of Dixon (1979). It is argued that the particular case-marking model employed (‘the coding view’) allows partially variable ‘fluid’ systems like the one described here to be compared extensionally (rather than intensionally) with other partially variable ‘fluid’ and categorical ‘split’ systems (including Tsova-Tush and Georgian, respectively). A particular approach to the theory and typology of grammatical variation in case-marking systems is advocated in terms of this model. It is further argued that the proposed typological distinction between ‘split’ and ‘fluid’ intransitivity is nothing other than a more general, orthogonal variable of ‘categorical’ versus ‘partially variable’ case-marking splits readily observable in other case-marking domains and not specific to intransitive subject case-marking.