Advertisement

Abstract

Over the years a number of schemes of knowledge representation (KR) have been developed that purportedly possess expressive power comparable to first- order logic (hereafter FOL). Frames (Minsky, 1975) and semantic networks (e.g. Schubert, 1976; Hendrix, 1979; Shapiro, 1979) have been employed success- fully in a number of systems and are well known in AI circles. Whether or not these systems are merely notational variants of FOL has been the subject of much debate (e.g. Israel and Brachman, 1984). Tenacious loyalty to the predicate cal- culus by some AI researchers is no doubt the result of its deductive strength and familiarity (e.g. Nilsson, 1971, 1980; McDermott, 1987). It has become popular to employ the linear notation of the predicate calculus to build databases that become intelligent through implementation of rules of inference and proof stra- tegies (e.g. Robinson, 1965; Bundy, 1983; Wos et al, 1984; Genesereth and Nils- son, 1987). The viability of first-order languages as KR systems is being increased by the development of methods to handle quantification over events, say through individuation of those events (see e.g. Davidson, 1967), and to handle knowledge and belief, say by allowing possible worlds to be admitted into the domain of quantification (see e.g. Moore, 1984). At the other end of the spectrum is the view that logic is inadequate from both psychological and linguistic per- spectives. Some researchers who hold this view have concentrated their efforts on the development of what has been described as informal methods for common- sense reasoning (e.g. see Schank and Nash-Webber, 1975). The polarized views on the role of logic in AI have prompted notable efforts to produce a common core of principles that would have a unifying effect on AI (e.g. Sowa, 1984). As yet, these efforts have failed to unify the field, and the aforementioned polarity in AI is still quite appreciable.

Keywords

Inference Rule Truth Predicate Existential Quantifier Standard Logic Fictitious World 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Kluwer Academic Publishers 1991

Authors and Affiliations

  • Cary G. deBessonet
    • 1
    • 2
  1. 1.AI ProjectLouisiana State Law InstituteUSA
  2. 2.Southern UniversityUSA

Personalised recommendations