Comparing the Use of Feature Structures in Nativism and in Database Semantics
Roland Hausser · 2007
Linguistics has always been a field with a great diversity of schools and sub-schools. This has naturally led to the question of whether different grammatical analyses of the same sentence are in fact equivalent or not. With the formalization of grammars as generative rule systems, beginning with the “Chomsky revolution ” in the late nineteen fifties, it became possible to answer such questions in those fortunate instances in which the competing analyses were sufficiently formalized. An early example is the comparison of Context-Free Phrase Structure Grammar (CF-PSG) and Bidirectional Categorial Grammar (BCG), which were shown to be weakly equivalent by Gaifman 1961. More recently, the question arose with respect to the lan-guage classes and the complexity hierarchies of Phrase Structure Grammar (PS-grammar) and of Left-Associative Grammar (LA-grammar), which were shown to be orthogonal to each other (TCS’92). Here we apply the question to the use of feature structures in contemporary schools of Nativism on the one hand, and in Database Semantics (DBS) on the other. The practical purpose is to determine whether or not the grammatical analyses of Nativism based on constituent structure can be used in Database Semantics. 1