LECTURETTE #10 (#3 on X-Bar Theory): THE DEMISE OF PS-RULES I mentioned earlier (back in lecturette #8) that the development of X-Bar Theory was motivated by the recognition that there were generalizations that could be made across different category-types. One type of generali- zation that i haven't so far mentioned has to do with constituent order. It was noticed that, if in an NP like the one in (1) the PP 'on the ta- ble' is identified as the 'complement' but the adjectives 'big' and 'blue' are not (perhaps constituting rather one specifier AP), then in Modern Standard English (MSE) head-complement order is the same in NPs as in VPs or PPs. Thus, it began to appear that a grammar of MSE didn't need to specify that PP complements follow head nouns while NP comple- ments follow head verbs or prepositions. All that needed to be said was that complements follow heads. (1) The big blue notebook on the table It became increasingly obvious as well that it shouldn't be necessary to state explicitly what categories were acceptable as specifiers or comple- ments. It ought to be possible to say that prepositions, merely by vir- tue of being prepositions, take NP complements; that verbs in general can take either NP or PP complements and (assuming the validity of the Inter- nal Subject Hypothesis) NP specifiers; and that nouns take PP complements and specifiers of appropriate types. More refined specifications (e.g., that certain verbs take no complements, that some take NP complements but not PP complements, that some take CP complements, etc.) should be speci- fiable as part of the head's lexical entry and generalizations of Theta Theory, which i shall be discussing soon. By the early '80's, some REST researchers were working intensively on the logical implications of this sort of generalization. Perhaps the most important single work in this direction was Tim Stowell's 1981 MIT dissertation Elements of Phrase Structure; subsequent works of nearly equal importance are Lisa Travis' 1984 MIT dissertation Parameters and Effects of Word Order Variation and Peggy Speas' book already mentioned in lecturette #8. The goal towards which all this work has been directed has been, to put it very briefly and bluntly, the complete elimination of phrase-structure rules as they existed in the Aspects model and in most other pre-1980 ge- nerative theories. It was increasingly felt that it should be possible, in describing any language, merely to state generally what the relative orders of heads, specifiers, and complements are, leaving all other de- tails up to either stipulations of Universal Grammar (UG) or independent constraints. Brief digression into somewhat more philosophical reaches of theory. The attitude described above is an example of the reason why this particular framework is called the 'Principles & Parameters Approach'. X-Bar Theo- ry, with its generalizations about structure and the relations between heads, specifiers, and complements, is a fundamental *principle* of UG. We presumably don't need to learn it; it's hard-wired into our brains. The issue of whether complements precede or follow their governing heads is presumably variable from language to language or (as we shall see) from category-type to category-type. It is thus a *parameter*, meaning something on which variation is allowed by UG, and which the language learner has to learn on the basis of empirical evidence. The child grow- ing up in an English-speaking household, on the bais of input like the strings in (2), concludes that in this language the order is specifier- head-complement, and 'sets' the relevant parameter accordingly. (2) a. Put it on the table. b. Eat your broccoli. c. That's a big dog. d. See the red squirrel in the tree? e. The table has four legs. f. The dog is chasing the squirrel. Imagine the internal grammar as a piece of complex electronic equipment with several switches on the outside. Most of its operation is predeter- mined on the basis of the internal circuitry, but the user has a number of options represented by the switches. Throwing a given switch one way sets the grammar to function in a certain specific fashion; throwing the same switch another way results in a somewhat different operation. The internal circuitry represents the principles of UG, the switches on the outside represent the parameters. The goal of most theoretical research in PPA has been to identify and minimize the number of parameters, while maintaining descriptive adequacy over the thousands of attested human lan- guages. As we shall see, this implies that each parameter-setting is likely to have multiple repercussions in the overt grammar. You may be wondering if in every language head-complement order can real- ly be generalized across categories. Scepticism in this area is well- founded. Indeed, if i'm not mistaken, it is necessary to say that in Chinese NPs exhibit complement-head order while VPs exhibit head-comple- ment order. (Or maybe it's the other way around; i'm having trouble find- ing my copy of Huang's dissertation where this matter is discussed in pas- sing.) But in any case, there are languages in which the overt ordering of heads and complements varies depending on the category-type of the head. Largely successful attempts have been made in REST to account for these variations on the basis of details of grammar outside X-Bar Theory per se. Many of these attempts appeal to Case Theory and will be dis- cussed in further detail under that heading. For the moment i'll just say that in a given language a post-head complement may be able to satis- fy Case Theory constraints just in case the head is a verb, but that pre- head position is necessary for it if the head is a noun. Since Case Theo- ry is critical at PF but not at DS, that means that at DS all complements can be generated on the same side of their respective heads, but some of them have to move by PF to the opposite side. So the goal of a lot of work in REST on X-Bar Theory and phrase structure has been the elimination of any need for explicit phrase-structure rules and the demonstration that all the facts that such rules were originally designed to express are derivable from more general principles of gram- mar. I'm not sure that the effort has been universally successful, but i will say that it has been very fruitful. We may, in subsequent discus- sion, come across some of the interesting insights it has provided. 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! ***