A case for a Weak Case Contiguity hypothesis-a reply to Caha.

The purpose of this paper is to discuss some counterexamples to the Strong Case Contiguity Hypothesis ( scch) argued for by Caha (2009, 2013) from West Nordic, which were absent from Caha's sample of languages. The languages to be discussed are Old Norse, Modern Icelandic and Modern Faroese. These three languages show a great deal of syncretism between dative and accusative to the exclusion of genitive, which also fulfill Caha's criteria for systematic syncretism. This pattern is not predicted to be possible under Caha's theory. Under closer inspection, however, the data from West Nordic turns out to be compatible with a weaker version of Case Contiguity. Syncretism between accusative and dative appear in a variety of contexts whereas syncretism between accusative and genitive to the exclusion of dative occurs only in a highly restricted context and should be considered accidental under Caha's criteria. Syncretism thus appears to target continuous regions on a slightly different case hierarchy. Hence although scch is wrong a weaker version is still feasible. I also propose a draft of a feature-based account for the observed variation in case hierarchies, providing a restrictive medium between scch and an unrestricted variation in hierarchies. [ABSTRACT FROM AUTHOR]