LECTURETTE #7: MODULES I have already mentioned the shift in attention in REST from transforma- tional operations to constraints on those operations. Most if not all of these constraints are grouped into what are often referred to collective- ly as 'modules', typically semi-autonomous systems each of which has its own set of basic principles and constraints based on those principles and which is relevant at particular levels of a derivation. The full deriva- tion of a grammatical string necessarily involves the interaction of these different modules; different modules will approve or disapprove of different constructions, focussing on different details, but in general in order to qualify as grammatical a string has to be approved by all modules. In each of the next few lecturettes i will be discussing at some length one of the major modules postulated in REST. Here i am merely going to give a quick overview of what they are supposed to be and do. 1. X-Bar Theory is the theory of phrase structure. It identifies commona- lities in the internal organization of different types of phrases and in the relations they bear to each other. It is relevant primarily at DS. 2. Theta-Theory is concerned with the proper assignment of semantic roles to NPs and the satisfaction of subcategorization or valency requirements. It too is relevant primarily at DS, although a formalization of the SPC known as the 'Extended Projection Principle' guarantees that it is also, by extension, relevant at all 'subsequent' levels. 3. Government Theory defines certain formal relations between constitu- ents in a constituent-structure tree-diagram and the nodes they occupy. As shall become clear in subsequent discussion, Government Theory serves primarily as a resource for other modules, providing them with the tools they need to carry out their own work, rather than as an autonomous mod- ule itself. 4. Case Theory operates at PF and defines when a given NP is or is not in a legitimate position in constituent structure. 5. Binding Theory is concerned with the co-indexing or coreferentiality of distinct syntactic nodes or constituents. In part it is concerned with overt coreferential elements such as reflexives and pronominals; in part it is concerned with 'empty categories' that behave in similar ways but have no phonological manifestation. Its constraints are to some ex- tent relevant at SS, but primarily at LF. 6. Bounding Theory (the similarity of name with Binding Theory is occa- sionally irritating, and i have been happy to be able to avoid Bounding Theory in a lot of my own work) is concerned with limits on the distance that a constituent can move. Many of you will probably have come across the term 'Subjacency'; this is where it belongs. Bounding Theory is re- levant primarily at SS. 7. Control Theory, like Binding Theory, has to do with the relations be- tween co-referential constituents. The difference, to the extent there is one, is that Binding Theory is concerned with relations that can be perspicaciously defined in strictly syntactic, i.e. structural terms; Control Theory is concerned with relations that appear to be just as real but which are nebulous from the formal point of view, and may in fact be more pragmatic than syntactic. So much for the overview; we'll be getting down to the details soon. Best, Steven --------------------- Dr. Steven Schaufele 712 West Washington Urbana, IL 61801 217-344-8240 fcosws@prairienet.org **** O syntagmata linguarum liberemini humanarum! *** *** Nihil vestris privari nisi obicibus potestis! ***