Browse Prior Art Database

ON INTERPRETING SEMANTIC NETWORK FORMALISMS

IP.com Disclosure Number: IPCOM000128812D
Original Publication Date: 1982-Dec-31
Included in the Prior Art Database: 2005-Sep-19
Document File: 23 page(s) / 78K

Publishing Venue

Software Patent Institute

Related People

David J. Israel: AUTHOR [+3]

Abstract

In a recent paper, Reiter and Criscuolo remark "that (semantic) networks are notational. variants of logical formulae is by now a truism in Artificial Intelligence circles" [24]. Shamelessly exploiting the foregoing quote as a pretext, I attempt to sketch adequate semantic accounts for at least two (kinds of) semantic network formalisms: one, based on the notion of inheritance; one, not. A crucial condition of adequacy to be satisfied is fidelity to some of the intuitions of the creators ~.1 of the formalisms.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 5% of the total text.

Page 1 of 23

THIS DOCUMENT IS AN APPROXIMATE REPRESENTATION OF THE ORIGINAL.

ON INTERPRETING SEMANTIC NETWORK FORMALISMS*

. David J. Israel September 1982

Prepared by: Bolt Beranek and Newman Inc. lO Moulton Street Cambridge, MA 02238

Prepared for: The Office of Naval Research

This research was supported by the office of Naval Research under Contract No. N00014-77-C- 0371. The views and conclusions contained in this document are those of the-author and should not be interpreted as necessarily representing the official policies, either expressed or implied, of the Office of Naval Research or the U.S. Government. ~A version of this report is to appear in a special issue on Computational Linguistics of the 'International Journal of Computers and- Mathematics. Unclassified SECURITY CLASSIFICATION OF THIS PAGE (When Date RaterwO ,~", Report No. 5117 Bolt Beranek. and Newman Inc.

ABSTRACT

In a recent paper, Reiter and Criscuolo remark "that (semantic) networks are notational. variants of logical formulae is by now a truism in Artificial Intelligence circles" [24]. Shamelessly exploiting the foregoing quote as a pretext, I attempt to sketch adequate semantic accounts for at least two (kinds of) semantic network formalisms: one, based on the notion of inheritance; one, not. A crucial condition of adequacy to be satisfied is fidelity to some of the intuitions of the creators ~.1 of the formalisms.

`~` DOCTRINAL PREAMBLE

One often hears that modal. (or some other) logic is pointless because it can be translated into some simpler ' ~ language in a first-order way. Take no notice of such arguments. There is no weight to the claim that the original system must therefore be replaced by the new one. What is essential is to single out important concepts and to investigate their properties. The fact that the real numbers can be defined in terms of sets is no argument for being interested in arbitrary sets. One must look among the sets for the significant ones and cannot be censured if one finds the intrinsic properties of the reals more interesting than any of their formulations in set theory. Of course if we can argue that set theory provides other significant concepts, then we may find some reason for going beyond the real numbers (and it is not hard to find the reasons i). But clearly this discussion cannot proceed on purely formal grounds alone. (From [261).

AMEN!

Report No. 5117 Bolt Beranek and Newman Inc.

1. INTRODUCTION

Bolt, Beranek & Newman, Inc. Page 1 Dec 31, 1982

Page 2 of 23

ON INTERPRETING SEMANTIC NETWORK FORMALISMS

In a recent paper, Ray Reiter and Giovanni Criscuolo remark that "the fact that networks are rotational variants of logical formulae is by now a truism in Artificial Intelligence circles" [241. Let us put aside the empirical, sociological claim -about which I am more than willing to defer to Messrs. Reiter and Criscuolo. Let us look rather at the content of the truism itself.

When we do so, we notice a certain ambiguity. Perhaps, R...