Modelling Control Theory

The main goal of the project is to develop a comprehensive syntactic theory of control which is based on the phase model as underlying framework. Since one basic insight in recent discussions on control has been that languages differ in whether they allow backward control or not, and since it has subsequently been claimed that the existence of backward control provides strong support for the MTC, the project has two pillars: One main focus lies on a close investigation of BC. In fact, BC raises two issues. First of all, the phenomenon as such has to be critically scrutinised: Are all structures that are commonly subsumed under BC really control structures? Which languages exhibit BC, and which criteria are reliable to identify BC? Once these questions are settled, BC can be used as an indicator of the adequacy of different theories of control – the question arises as to whether a given theory can handle the data, and since the discussion of BC is a relatively new one, it might play a decisive role in determining the validity of control theories. The hypothesis we will explore is that BC can be reduced to LDA. If this turns out to be correct, then the picture is as follows: There are on the one hand Greek-type languages (which we label Agree languages), for which LDA is involved also in control structures. This is a hypothesis which will have to be tested for more languages, and if true, it would be an appealing account of BC within the phase model. Moreover, this would imply that the phenomenon of BC does not provide a direct argument for the MTC, after all. On the other hand, there are languages which lack BC (like English), and what is therefore still missing is a new theory of control for these languages. This is another major goal of this project: to develop a theory of control which combines the advantages of both the PRO-based theories of control and the MTC. Although a semantic impact on control is undeniable (as discussed before), it is indisputable that control is also governed by syntactic constraints; cf., for instance, the parallelism between OC PRO and anaphors as far as the syntactic restrictions on their antecedents are concerned (cf. Hornstein 1999:73ff.). The focus of this project lies on the syntactic aspects of control: How is control syntactically constrained and licensed? Conceptually, the idea is to come up with a local derivational theory which is again compatible with recent minimalist assumptions, i.e., a theory which is based on the phase model. Such an approach is missing in the recent debate on control and would finally complete the overall picture of a comprehensive phase-based theory of control.