Local Constraints vs. Economy (review)

This monograph critically investigates the assumption that human language is a perfect computational system, underlying the (1995 version of the) minimalist program (MP; cf. N. Chomsky, The minimalist program, Cambridge: MIT Press, 1995); Johnson and Lappin argue that language design is not optimal along the lines presented in MP and related work. It particularly subjects the notion(s) of economy employed in MP to scrutiny; J & L argue that MP does not present persuasive evidence for the implementation, not even for the concept of either local or global economy (of derivation). The ‘Introduction’ (1–7) lays out the aim and scope of the book and sketches the main theoretical ideas of MP.

The chapter ‘Global economy’ (8–45) presents versions of (global) economy discussed in the minimalist literature (‘procrastinate’, ‘smallest derivation’, and ‘have an effect on output’). J & L replace such global conditions by local constraints (or eliminate them altogether, such as the last one). They argue specifically from the perspective of computational complexity and criticize MP for not being able to deal with these issues. (On the relevance of such complexity issues in a minimalist setting, however, see J. C. Castillo, J. Drury, and K. K. Grohmann, ‘Merge over move and the extended projection principle’, University of Maryland Working Papers in Linguistics 8. 63–103, 1999.) Ch. 3, ‘Local economy’ (46–78), then critically examines MP’s desire to evaluate derivations locally. J & L discard some local economy conditions (such as ‘last resort’ or ‘minimality’) as not being real economy conditions at all and replace them by local constraints, or eliminate them altogether (on the basis of lacking empirical support and/or being ad hoc devices). They also argue against the global ‘look-ahead’ character of some recent minimalist models that tackle complexity issues and computation. The single-point interface architecture of MP is also replaced by an ‘incremental interface correspondence model’ which processes the syntactic, semantic, and phonological structures ‘incrementally and in parallel’ (65). (Incidentally, some recent minimalist work aims at opening up the classic single-point ‘interface model’, see, for example, a paper that has been circulated as far back as 1995: J. Uriagereka, ‘Multiple Spell-Out’, Working minimalism, ed. by S. D. Epstein and N. Hornstein, 251–82. Cambridge: MIT Press 1999, or the reviewer’s 2000 University of Maryland doctoral dissertation, Prolific peripheries: A radical view from the left.)

As the chapter heading suggests, ‘Wh-questions: Local constraints vs. feature checking’ (79–123) criticizes checking theory and compares a feature-checking-and-movement approach with a declarative, typed feature structure employing local constraints. Needless to say, J & L argue in favor of the latter. For illustration, they adopt a revised version of head-driven phrase structure grammar (HPSG) within a multiple inheritance hierarchy for (phrase and clause) construction types; as J & L point out, however, alternative local constraint-based approaches allow similar analyses. In the final chapter, ‘The concept of optimal design in the theory of grammar’ (124–33), J & L return to questions of optimal design and the hypothesis that grammar is a perfect computational system. Again, they propose replacing the MP-approach of economy (of derivation) and feature-checking with a local constraint-based framework, of which a current version of HPSG is a possible candidate (as presented in Chs. 3 and 4).

This book offers an interesting discussion of economy and optimization in grammar, posing some potential challenges to the minimalist endeavor. How far the arguments go through or can be dealt with in continuing refinements of MP should be left to the target audience of the book, economy- and constraint-proponents alike. What can be hoped for is an ensuing lively discussion between both factions.