Case Conflict in German Free Relative Constructions: An Optimality Theoretic Treatment

Abstract

Languages differ as to how big a case conflict must be in a free relative (FR) construction to cause ungrammaticality. While English requires true categorial matching, German allows the suppression of structural cases if assigned by the matrix verb. There are also different types of non-matching languages. Paradigmatic examples are Gothic and Modern Greek. Earlier generative syntactic accounts mainly propose a distinction only between matching and non-matching languages. This is not fine-grained enough to capture the typological findings. An optimality theoretic treatment permits a richer, but not infinite, typology, and it allows constraint violation (which obviously happens in FR constructions). The proposed account makes use of the optimality theoretic conception of correspondence. The assumed constraints are on input-output correspondence (input-LF as well as input-PF), and also on PF-LF correspondence. Free Relative ConstructionsA considerable amount of attention has been paid to free relative constructions in generative grammar in the last thirty years. Some interesting and puzzling properties make this construction worth examining. Free relatives have the somewhat paradoxical property of being clauses that stand for non-clausal constituents. The following example contains five such relative clauses:(1)Wer sich mit freien Relativsätzen beschäftigt, verwendet, was von ihm dafür gehalten wird, sooft er kann, wann immer sich ihm dafür eine Gelegenheit bietet und wo immer er sich befindet.≈ "Whoever deals with free relative clauses, uses what he considers to be one, as often as he can, whenever he has the opportunity and wherever he is" This paper will not deal with adverbial clauses, such as the last three highlighted clauses in (1). I will concentrate on free relatives (FRs) that realise Note: This is a pre-final draft version of the published version. It is nearly, not fully identical to it! Vogel, Ralf (2001). Case Conflict in German Free Relative Constructions.

Degraded Acceptability and Markedness in Syntax, and the Stochastic Interpretation of Optimality Theory

“…An example is the pair in (1) Experiments usually test for contrasts between minimally different expressions. In our example, the theory of case matching in argument free relative clauses (Groos and van Riemsdijk, 1981;Pittner, 1991;Vogel, 2001) is confirmed if (1-b) is judged as grammatical less often than (1-a) to a statistically significant degree. This is indeed the result of a speeded grammaticality judgement experiment by Boethke (in preparation).…”

“…Simplifying my own account (Vogel, 2001(Vogel, , 2002(Vogel, , 2003b, we can assume the following constraints to distinguish the four structures:…”

“…(1-b) * * * (1) and (2) Groos and van Riemsdijk (1981); Pittner (1991) and Vogel (2001) offer three different approaches to case conflicts in German FRs. Interestingly, these authors also differ in the grammaticality judgements they present.…”

On the nature of the input in optimality theory

“…The universe of syntactic structures is in turn generated by an input-free system, for instance a context-free grammar. Such a system (see Vogel 2001a, Wilson 2001 does without the additional generator Σ and is therefore not susceptible to the objection just made. However, given that input and output are not in any way derivationally related in such an approach, the distinction between input and output in the classical sense becomes blurred, and it is not quite clear anymore to what extent the notion of input continues to play a crucial role in the theory.…”

“…A classical violation of UNIFORMITY in phonology is coalescence. In syntax the analysis of free relative constructions proposed in Vogel (2001a) is a case in point. Vogel assumes that the input is a fully speciÞed syntactic structure.…”