Recent llnguistic theories cast surface complexity cs the result of interactl. ng subsys- tems of constraints. For instance, thc ID/LP grammar fonntdism separa,tcs constreints on immediate dominance from those on linear order. Shitbet (1983) has shown how to carry out direct paring of ID/LP grammars. His algorithm uses ID mid LP constraints directly in langmrge processing, without expmidhg them into a context-free "object gra.mmar." This report exmnines the computational diculty of ID/LP parsing. ShicbeFs purported O(]GI . n ) runtime bound mderestlmates the difficulty of ID/LP parsing; the worst-case runtime of his algorithm is exponential in grammar size. A reduction of the vertex-cover problem proves that ID/LP parsing is NP-complete. The growth of intcrna.1 data struc- tures is the source of difficulty in Shieber's algorithm. The computational and linguistic implicatimxs of these results are discussed. Despite the potentid for combinatorial explo- sion, Shicber's algorithm roma.ins better than the alternative of pa'rs]ng an expanded object gramma.r.

This paper will not delve into this controversy, but will just show how some of the constraints proposed recently by Chomsky and others - constraints to which all human languages are thought to conform - can very easily be enforced in a parsing system that allows an elegant grammar notation. These grammars will be called restricted logic grammars (RLGs). Two well known logic grammar formalisms, definite clause grammars (DCGs) and extraposition grammars (XGs), will be briefly reviewed, and then RLGs will be introduced by showing how they differ from XGs. RLGs have a new type of rule ("switch rules") that is of particular value in the definition of natural languages, and the automatic enforcement of some of Chomsky's constraints makes RLG movement rules simpler than XGs'. We follow the work of Marcus (1981), Betwick (1980), Wehrli (1984) and others in pursuing this strategy of restricting the grammar formalism by enforcing Chomsky's constraints, but we use a simple nondeterministic top-down backtracking parsing method with lookahead, rather than Marcus's deterministic LR(k,t)-like parsing method. This approach to parsing, which has been developed in logic Copyright1987 by the Association for Computational Linguistics. Permission to copy without fee all or part of this material is granted provided that the copies are not made for direct commercial advantage and the eL reference and this copyright notice are included on the first page. To copy otherwise, or to republish, requires a fee and/or specific permission

This report describes research done at the Artificial Intelligence Laboratory of the Massachusetts Institute of Technology. Support for the laboratory's artificial intelligence research is provided in part by the Advanced Research Projects Agency of the Department of Defense under Office of Naval Research contract N00014-91-J-4038. The author was also supported by the National Science Foundation under grant NSF-ASC-92-9217041, and an Air Force Graduate Laborarory Fellowship. Acknowledgments Among the people who have provided valuable suggestions and criticism regarding this research, I would especially like to thank my advisor Bob Berwick for his support and prodding. Conversations with Dave Baggett and Eric Ristad helped clarify my positions on a variety of topics. This report describes research done at the Artificial Intelligence Laboratory of the Massachusetts Institute of Technology. Support for the laboratory's artificial intelligence research is provided in part by the Advanced Research Projects Agency of the Department of Defense under Office of Naval Research contract N00014-91-J-4038. The author was also supported by the National Science Foundation under grant NSF-ASC-92-9217041, and an Air Force Graduate Laborarory Fellowship. Chapter 1 Introduction Many search problems are commonly solved with simple combinatoric algorithms that unnecessarily duplicate and serialize work at considerable computational expense. There are a number of techniques available that can eliminate redundant computations and perform remaining operations in parallel, effectively reducing the branching factors of these algorithms. This thesis investigates the application of these techniques to the problem of parsing natural language into grammatical representations. The result is a use...

this paper have illustrated, context-free generarive power does not guarantee efficient parsability. Every ID/LP grammar technically generates a context-free language, but the potentially large size of the corresponding CFG means that we can't count on that fact to give us efficient parsing. Thus it is impossible to sustain this particular argument for the advantages of such formalisms as (early) GPSG over other linguistic theories; instead, GPSG and other modern theories seem to be (very roughly) in the same boat with respect to complexity. In such a situation, the linguistic merits of various theories are more important than complexity results. (See Berwick (1982), Berwick and Weinberg (1982, 1984), and Ristad (1985) for further discussion.) The reduction does not rule out the use of formalisms that decouple ID and LP constraints; note that Shieber's direct parsing algorithm wins out over the use of the object grammar. However, if we assume that natural languages are efficiently parsable (EP), then computational difficulties in parsing a formalism do indicate that the formalism itself does not tell the whole story. That is, they point out that the range of possible languages has been incorrectly characterized: the additional constraints that guarantee efficient parsability remain unstated. Since the general case of parsing ID/LP grammars is computationally difficult, if the linguistically relevant ID/LP grammars are to be efficiently parsable, there must be additional factors that guarantee a certain amount of constraint from some source. TM (Constraints beyond the bare ID/LP formalism are required on linguistic grounds as well.) Note that the subset principle of language acquisition (cf. Berwick and Weinberg 1984:233) would lead the language learner to initially hyp...

This collection of invited papers covers a lot of ground in its nearly 800 pages, so any review of reasonable length will necessarily be selective. However, there are a number of features that make the book as a whole a comparatively easy and thoroughly rewarding read. Multiauthor compendia of this kind are often disjointed, with very little uniformity from chapter to chapter in terms of breadth, depth, and format. Such is not the case here. Breadth and depth of treatment are surprisingly consistent, with coherent formats that often include both a little history of the field and some thoughts about the future. The volume has a very logical structure in which the chapters flow and follow on from each other in an orderly fashion. There are also many crossreferences between chapters, which allow the authors to build upon the foundation of one another’s work and eliminate redundancies. Specifically, the contents consist of 38 survey papers grouped into three parts: Fundamentals; Processes, Methods, and Resources; and Applications. Taken together, they provide both a comprehensive introduction to the field and a useful reference volume. In addition to the usual author and subject matter indices, there is a substantial

R(H:-G 1 )oe The Notation for Encoding Principles 250 phrase, it can make use of values already computed for its sub-constituents. 4 In fact, the macro expansion for interleaving is virtually identical to that for the noninterleaved case, except that every time the values for a constituent are requested, it tests to see whether those values have already been computed --- in which case, the saved values are used. Otherwise, it computes the values in the regular manner using !Conditions-1?. This is implemented by having an extra (initial) clause for !Operation-Name? that performs the memory retrieval function. Because of the similarity of the two forms, we omit the description of the second form. B.3 Type Composition Rules In this section, we define rules for propagating types in Prolog programs. For example, inference rules will be given for the conjunction, disjunction, and negatio