How the reality can be compatible with the strong minimalist thesis : Interfacing, economy, and variation

This paper explores possibilities of implementing the Strong Minimalist Thesis (SMT) of Chomsky (2000, 2001, 2004) in the particular empirical domain of the computational system of human language (see (1)). Capitalizing on some specific approach to multiple spell-out and its interesting application to some concrete linguistic phenomena (Epstein et al., 1998; Lopez, 2003), it shows that the \u27invasion\u27 property of a number of external modules inter facing narrow syntax in the sense of Epstein et al. (1998; see also Lopez, 2003, and its modification in the present paper) may have certain restricting effects on one of the central assumptions of the cartography project (Rizzi, 1997; Cinque, 1999, Cinque (Ed.), 2002; Rizzi (Ed.), 2004; Belletti (Ed.), 2004) to the effect that "the interpretation is read off the syntactic configuration" (Belletti, 2004: 17), it being possible for linguistic features to enter narrow syntactic derivation at the time of spell-out, for example (Lopez, 2003). Note that the idea of interfaces invading the syntax and assigning it \u27interface\u27 features and of the syntax coping with this situation should definitely point to that of narrow syntax (language/FL) being usable in terms of the interfaces, thus explicitly representing one of the plausible and viable ways Chomsky\u27s (2000, 2001, 2004) SMT may be implemented within the domain involving the computational system of human language. Interfaces (PF, LF, pragmatics, and others) have now been called \u27components\u27 rather than \u27levels\u27 since they are assumed to exist interspersed throughout the computational component, due to the adoption of multiple approaches to spell-out (Chomsky, 2000, 2001, 2004; Epstein et al., 1998; Uriagereka, 1999). I further show mainly based on Lopez (2003) (see also Epstein et al., 1998) that interfaces should be considered to be \u27invasive\u27 (in the sense that they assign/attach linguistic features in some way to the syntactic structure under derivation at the time of spell-out; see also further development/modification of the notion throughout the present paper) rather than interpretive (the latter interpretation of interfaces having so far been standardly assumed). As for concrete analytical devices, I adhere to Chomsky\u27s (2001) original, rigid version of the Phase-Impenetrability Condition instead of its weaker version, and make an extensive use of Chomsky\u27s (2001) \u27guiding principle,\u27 which is called the Phase-Guiding Principle in the present paper. I also propose the standard X\u27-schema analysis of coordinate and adjunction structure, ConjP and AdjP constituting a strong phase, respectively. It will further be seen in the paper that the major part of parametric variation across languages should be attributed to optionality seen in optional rules/principles, the latter applying in accordance with economy, and that an \u27interfacing\u27 account of interactions between narrow syntax and various external (to it) modules/components is ensured by the SMT as a specific instance of its implementation in the domain of the computational system of human language