Michael B. Kac

Reconnaissance-attack parsing

In this paper wc will describe an approach to parsing one major component of which is a stragegy called RECONNAISSANCEATTACK. Under this strategy, no structure building is attempted until after completion of a preliminary phase designed to exploit low-level information to the fullest possible extent. This first pass then defines a set of constraints that restrict the set of available options when structure building proper begins. R-A parsing is in principle compatible with a variety of different views regarding the nature of syntactic representation, though it fits more comfortably with some than with others--a point to which we shall return.

Grammars and Grammaticality

At the outset, the goal of generative grammar was the explication of an intuitive concept grammaticality (Chomsky 1957:13). But psychological goals have become primary, referred to as “linguistic competence”, “language faculty”, or, more recently, “I-language”. Kac argues for the validity of the earlier goal of grammaticality and for a specific view of the relationship between the abstract, nonpsychological study of grammar and the investigation of the language faculty. The method of the book involves a formalization of traditional grammar, with emphasis on etiological analysis, that is, providing a “diagnosis” for any ungrammatical string of the type of ungrammaticality involved. Part I justifies this view and makes the logical foundations of etiological analysis explicit. Part II applies the theory to a diverse body of typically generativist data, among which are aspects of the English complement system and some problematic phenomena in coordinate structures. The volume includes pedagogical exercises and especially intriguing is a large analysis problem, originally constructed by Gerlad Sanders using data from Nama Hottentot, which exposes the reader to a syntax of extraordinary beauty.

Parsing without (much) phrase structure

Approaches to NL syntax conform in varying degrees to the older relational/dependency model, (essentially that assumed in traditional grammar), which treats a sentence as a group of words united by various relations, and the newer constituent model. Some modern approaches have nonetheless involved shifts away from essentially constituent-based models of the sort associated with Bloomfield and Chomsky to more relation-based ones (e.g. case grammar, relational grammar, daughter-dependency and word grammar, corepresentational grammar) while some others, notably lexical-functional grammar, have nonetheless continued to rely crucially on certain techniques inherited from constituency-based grammar, particularly context-free grammar. In computational linguistics there is a strong (if not universal) reliance on phrase structure as the medium via which to represent syntactic structure; call this the CONSENSUS VIEW. A significant amount of effort has accordingly been invested in techniques by which to build such a representation efficiently, which has in turn led to considerable work on the formal and computational properties of context-free gramamrs (or natural extensions of them) and of the associated languages. In its strongest form, the consensus view says that the recovery of a fully specified parse tree is an essential step in computational language processing, and would, if correct, provide important support for the constituent model. In this paper, we shall critically examine the rationale for this view, and will sketch (informally) an alternative view which we find more defensible. The actual position we shall take for this discussion, however, is conservative in that we will not argue that there is no place whatever for constituent analysis in parsing or in syntactic analysis generally. What we WILL argue is that phrase structure is at least partly redundant in that a direct leap to the composition of some semantic units is possible from a relatively underspecified syntactic representation (as opposed to a complete parse tree). However, see Rindflesch forthcoming for an approach to.parsing which entails a much stronger denial of the consensus view.

Theoretical implications of disordered syntactic comprehension

Abstract We reexamine data from Caplan and Hildebrandt 1988 within the context of a different set of background assumptions, concluding that where these can be clearly distinguished from those underlying GB theory the evidence favors a non-GB-based account. We attribute the deficits observed in the process of infinitival complement constructions to an inability on the part of the patients to access either or both of two data structures required to support our proposed parsing algorithm, and show that on this account it is unnecessary to posit a compensatory heuristic to account for the behavior observed. We also question some other cases where Caplan and Hildebrandt advert to compensatory heuristics.

A Simplified Theory of Boolean Semantic Types

The theory of semantic types in Keenan and Faltz (1985) is insufficiently constrained in the sense that it requires denumerable categories to be interpreted under certain conditions via nondenumerable algebras. An ontologically more austere version of the theory is proposed in which expressions are always interpreted in terms of finite algebras and it is shown how it is nonetheless possible to treat an infinite language by providing an inductively defined hierarchy of such algebras, each representing a stage of an expanding knowledge base. Some apparent obstacles are considered and disposed of and some advantages discussed, having to do with the alethic modalities and referential opacity induced by predicates of propositional attitude. Finally, it is shown that a weaker version of Keenan and Faltzs central mathematical result, the Justification Theorem, suffices for the revised system and a simple, intuitive proof