We now want to deal with how semantic networks can be used to represent knowledge. It has already become clear that the term "semantic network" does not stand for a clearly delimitable set of representations with precisely defined properties, but is a rather unspecific designation. In principle, all representations based on a graph formalism (Dörfler/Mühlbacher 73) can be conceived as semantic networks. For better illustration, we will illustrate semantic nets graphically only. Linear notations do not make the characteristic of semantic networks so clear, but are of course possible without further ado.
For example, edges can be represented by two-digit predicates; see also (Hendrix 79) and (Sowa 84).
The assumption of name uniqueness introduced in chapter 2.2.2 also applies to semantic networks:
**Agreement: Assumption of Name Uniqueness** In a semantic network, two nodes of different labels are considered different.
DOT FROM lambda-browsing
It follows that two nodes which are not different have the same Labels. Since both in such a case must be connected by the same edges to the same other nodes, the following agreement makes sense:
**Agreement: Uniqueness of Nodes** In a semantic network, no two nodes of the same label occur.
Finally, we also make the following agreement:
**Agreement: Implicit Conjunction** All statements represented in a semantic network are implicitly conjunctively linked.
⇒ Features
~
Labels: give Name: remove Name: remove Distinctions:
Region: define Boundary:
//wiki.ralfbarkow.ch/assets/scripts/freeform-schema.html HEIGHT 300