Against the Feature Bundle Theory of Case

In virtually all syntactic theories, case is taken to be a bundle of features, and case agreement as identity of case features. There is a large variety of facts which this view cannot explain. Suffixaufnahme 1 is just one example. Based on evidence from Finnish, Hungarian and German, I shall argue here that cases are sequences of morphemes. These morphemes are proper signs of the language, and therefore also have a meaning, which is also called case function. When a sign combines with a case marker, it has basically two options: it can combine with it via function application or via function composition, or it can put the case marker onto its case stack. This case marker can later either be popped off the stack, or the entire stack can be selected by a higher head. In the latter instance we say that the head selects its argument in the case named by the stack. 2 1Literally translated it means ‘taking up of suffixes’. For the history of this term as well a survey of the languages with Suffixaufnahme see Plank (1995) 2This paper is based on a talk held at the workshop on case at the DGfS Meeting in Marburg. I am indebted to the organizers, Ellen Brandner and Heike Zinsmeister, for giving me the oppor-tunity to present this material, and to the audience of the workshop, in particular Miriam Butt and Barbara Stiebels for useful comments, which have given rise to serious improvements. Parts of this material have also been presented at the GGS 2000 in Potsdam. I am grateful to Josef Bayer, Regine Eckhart and Werner Frey for discussing important issues with me. Thanks to Monika Budde, Ilse Zimmermann and two anonymous referees for reading an earlier version of this paper and providing me with numerous comments. Thanks finally to Olli Valkonen for his judgements on Finnish, and to Johanna Domokos for her judgements on Hungarian. I take full responsibility for all omissions and errors. 1