AN INTRODUCTION TO SNEPS
Original Publication Date: 1976-Dec-31
Included in the Prior Art Database: 2007-Mar-30
Software Patent Institute
Shapiro, Stuart C.: AUTHOR [+2]
Stuart C . Shapiro
AN INTRODUCTION TO SNePS
Stuart C . Shapiro
Computer Science Department Indiana University Bloomington, Indiana 47401
An Introduction t o SNePS
Stuart C. Shapiro
Computer Science Department Indiana UniverJs i t y
Bloomington, 1ndia.na 47401
SNePS (Semantic Network Processing System) is a system for build- ing directed graphs with labelled nodes and edges and locating nodes i n such graphs according t o graph patterns. Rather then being a general system for processing labelled digraphs, SNePS is restricted i n certain ways, appropriate for its intended use--to model "seman- tic" or "cognitive" structures. SNePS may be used interactively
by a human t o explore various approaches t o semantic representation, or it may be used as a collection of functions by a more complete natural language understanding program. This paper gives a user- oriented introduction t o SNePS.
SNePS (Semantic Network Processii~g System) is a system for building directed graphs with labelled nodes and edges and loca- ting nodes in such graphs according to graph patterns. Rather than being a general system for processing labelled digraphs, SNePS is restricted in certain ways, appropriate for its intended use -- t o model "semantic" or "cognitive" structures. SNePS, a revised ver- sin of MENTAL [Shapiro 1971a, 1971b1, is written In LISP 1.6 and runs on a DEC-system-10.
Edge labels represent binary semantic relations which are used to structure the network and about which no information can be stored in the network. For example, the cases of Fillmore, 1968, might be such labels. The user of SNePS is free to choose and declare his own set of labels. There are two kinds of edges, regular edges and auxiliary edges. Regular edges come i n pairs, one repre- senting a descending relation, the other the ascending converse of the first. Auxiliary edges do not come in pairs and do not have converses. If apath of descending edges goes from node n to node
m, we w i l l say that node n dominates node m.
There are three kinds of nodes in the network: constant, non-
constant, and auxiliary nodes. Auxiliary nodes are connected t o each other and/or to other nodes only by auxiliary edges. Constant nodes represent semantic concepts, including anything about which information may be stored in the network. Nodes which dominate
no other node are called atomic nodes. Atomic constants are called base nodes and atomic non-constants are called variable nodes. Non-atomic nodes are called molecular nodes. A molecular node which dominates any variable node is called a pattern....