Implications of a Revised Perspective on Minimalist Grammars
Jens Michaelis · 2002
ABSTRACT: The type of a minimalist grammar (MG) as introduced by Stabler [17] provides an attempt of a rigorous formalization of the perspectives adopted nowadays within the linguistic framework of transformational grammar. As shown in [11], MGs constitute a weakly equivalent subclass of linear context–free rewriting systems (LCFRSs) in the sense of Vijay–Shanker et al. [21]. Independent work of Harkema [5] and Michaelis [13] has proven the reverse to be true, as well. Hence, MGs as defined in [17] join to a series of formalism classes—among which there is e.g. the class of multicomponent tree adjoining grammars (MCTAGs) in their set–local variant of admitted adjunction (cf. [22])—all generating the same class of string languages. Inspired by current linguistic developments, a revised type of an MG as well as a certain type of a strict MG (SMG) has been proposed by Stabler [18]. Here we show that, in terms of derivable string languages, the revised MG–type as well as the SMG-type is not only subsumed by LCFRSs, but both also fall within a particular subclass of the latter: the righthand side of each rewriting rule of a corresponding LCFRS involves at most two nonterminals, and if two nonterminals appear on the righthand side then only simple strings of terminals are derivable from the first one. This result is in fact of specific interest, since conversely, in terms of weak equivalence, the corresponding LCFRS–subclass is provably subsumed by the class of revised MGs as well as the class of SMGs ([10]). Whether the inclusion of the respective classes of string languages derivable by the corresponding LCFRS–subclass and the class of all LCFRSs is proper or not seems to be an open problem. We briefly discuss what seems to constitute the crucial difference seen from the minimalist perspective at the end of this paper.