[Next] [Up] [Previous]
Next: A Description of Semantic Up: Reasoning with Semantic Nets Previous: Reasoning with Semantic Nets

What Are Semantic Networks? A Little Light History

The concept of a semantic network is now fairly old in the literature of cognitive science and artificial intelligence, and has been developed in so many ways and for so many purposes in its 20-year history that in many instances the strongest connection between recent systems based on networks is their common ancestry. The term `semantic network' as it is used now might therefore best be thought of as the name for a family of representational schemes rather than a single formalism. A little light history will clarify how the network we shall use in our Automated Tourist Guide is related to other networks you may come across in your reading.

The term dates back to Ross Quillian's Ph.D. thesis (1968), in which he first introduced it as a way of talking about the organization of human semantic memory, or memory for word concepts. The idea of a semantic network -- that is, of a network of associatively linked concepts -- is very much older: Anderson and Bower (1973, p. 9), for example, claim to be able to trace it all the way back to Aristotle. Specifically, semantic networks were conceived as a ``representational format [that would] permit the `meanings' of words to be stored, so that humanlike use of these meanings is possible'' (Quillian, 1968, p. 216), and, as with almost all mainline research in semantic nets since Quillian's original proposal, they were intended to represent the non-emotive, so-called `objective' part of meaning: the properties of things, rather than of the way we may feel about them.

To get some feel for semantic nets, think of a common, but evocative, word, say, `castle'. Write it down in the middle of a sheet of paper. Now think of some words related to castle, say, `king', or `battlement'. Write down these words in a ring around `castle', and join each of them with a line to `castle'. Try and give each line a label that describes the relationship between the two words -- for example, the line linking `king' and `castle' might be labelled `lives in'. Continue outwards, writing down words relating to `king', words relating to `battlement', and so on. What you are constructing is, roughly, a semantic net.

Quillian's basic assumption was that the meaning of a word could be represented by the set of its verbal associations. To see what this means, imagine that, in the course of reading a novel, you come across the word `dugong' and the context does not make clear what the word refers to. So you look up the word in a dictionary, and there you find, not the object or the property or the action itself, but rather a definition made up of other words -- in the present case,

DUGONG: a herbivorous marine mammal of tropical coastal waters of the Old World, having flipperlike forelimbs and a deeply notched tail fin.

You still have no clear idea of what a `dugong' is, so you then look up each of the words making up the definition, and in turn each of the words making up the definition of each word in the definition of the original word, and so on, learning that `herbivorous' means `feeding on plants; plant-eating', that a `flipperlike forelimb' is `a wide, flat limb, as of a seal, adapted especially for swimming', that `marine' means `native to or formed by the sea', but that nonetheless it is not a fish but a `mammal' which is `a member of the class Mammalia', in turn `a class of vertebrate animals ...distinguished by self-regulating body-temperature, hair, and, in the female, milk-producing mammae', and so on and so forth. As you follow through all the cross-references, so you build up a complex picture of the concept named by the word and of its relation to other concepts, say, that of manatee, whale, mammal, animal, life form.

Clearly, such a mental representation exceeds the mere dictionary definitions of the words you have looked up: semantic networks, as do dictionaries more indirectly, reflect the complex manner in which human knowledge is structured, every concept being defined in terms of its place in a web of relationships between concepts. We might picture a person's knowledge as a map, with points or nodes representing individual concepts and labelled links (called arcs or pointers in some texts) connecting those nodes together. Just as we know where Trafalgar Square is because we know how to get there from Picadilly Circus, or from Charing Cross, or from St. James Park, so too we know now what, for example, a dugong is because we `know how to get there' from a tail fin, a herbivore, a flipper, and a mammal.

The foregoing `map' analogy should not be pushed too far. In Quillian's original semantic networks, a relation between two words might be shown to exist if, in an unguided breadth-first search from each word, there could be found a point of intersection of their respective verbal associations. We would not, by contrast, wish to find a route from Tower Bridge to Trafalgar Square by blindly sending out search parties in all directions from each location in the hope that they might eventually meet! While Quillian's early nets might have appeared to be an attractive psychological model for the architecture of human semantic knowledge, they did not provide an adequate account of our ability to reason with that knowledge.

A couple of years later a psychologist, Allan Collins, conducted a series of experiments along with Quillian to test the psychological plausibility of semantic networks as models, both of the organization of memory and of human inferencing. The networks they used, such as that in figure 6.1, now gave far greater prominence than before to the hierarchical organisation of knowledge. (Don't worry too much on the first reading about trying to understand what the network in figure 6.1 means; it will become clearer on a second reading, after the network formalism has been explained.)

  [IMAGE ]
Figure 6.1: A taxonomic tree. From Collins and Quillian (1969).

The network is displayed as a taxonomic tree or isa hierarchy (a term we introduced briefly in chapter 3). Each node (labelled by an underscored word) is connected upwards to its superset and downwards to its subset. A canary, in this schema, is a bird and, more generally, an animal. A shark, too, is shown to be an animal, but it is not a bird, as there is no link up from `shark' to `bird'. An ostrich is a bird because, like the canary, it is one of the children of the `bird' node. The links sideways from each node state properties that are true of the node -- that birds can fly, for example, or that canaries can sing -- and properties of higher nodes are inherited by the lower nodes to which they are connected unless there is a property attached to a lower node that explicitly overrides it. Thus, we may infer from the tree that canaries can fly because birds in general can fly, whereas we are inhibited from making the same inference for ostriches since there is the explicit statement at the ostrich node that it `can't fly'.

Collins and Quillian's experiments consisted in presenting subjects with sets of true and false sentences, and measuring their reaction time in deciding whether the sentences were true or false. Taking the number of links to be traversed between two nodes to be a measure of the semantic distance between concepts, they predicted that a person would require more time to decide, for example, that ``A canary is an animal'' or ``A canary has skin'' than to decide that ``A canary is a bird'' or ``A canary can sing'' since in the former cases the search for the relevant information requires rising through more links in the hierarchy. The experimental results met their predictions. Table 6.1 illustrates the kinds of stimulus sentences, with approximate reaction times, that were used in these experiments. While subsequent research on reaction times to sentences, such as that by Conrad (1972) and Smith, Shoben, and Rips (1974), raised doubts about the soundness of the model in the form it then had, Collins and Quillian's hierarchical nets were an important source of many good ideas for, and the direct forerunners of, more recent networks, particularly in the domain of language understanding.


Sentence True sentences Mean reaction
type time (msec)
P0 A canary can sing. 1.3
P1 A canary can fly. 1.38
P2 A canary has skin. 1.47
S0 A canary is a canary. 1.00
S1 A canary is a bird. 1.17
S2 A canary is an animal. 1.23
Table 6.1: Reaction times for sentences with differing semantic distances. From Collins and Quillian (1969).


One such network developed for the representation of sentence meaning was that of Robert Simmons and his co-researchers and successors. Quillian's hierarchical classification of world knowledge had no place in Simmons's networks, which were designed to capture the meanings of sentences by extending from a node representing the main verb a set of links to nodes representing the cases associated with the verb. Thus, to take a simple example, the meaning of the sentence ``John gave Mary an apple'' might be represented by a net such as

where the central node names the action, the remaining nodes are labelled by the names of the participants in the action, and the links between the nodes indicate the relationship of the participants to the action: it is John who does the giving, the apple which is given, and Mary who receives the apple. Furthermore, the arrows on the links indicate how the meaning should be read: that, for example, the apple is the object of the act of giving and not that giving is the object of apple.

Because, in any act of giving, the underlying case relationships remain the same -- there is always someone who gives, something given, and someone to whom it is given -- we might see the network above as the filling in, for the sentence ``John gave Mary an apple'', of a more abstract schema. That same schema also provides an interpretation for numerous other sentences: ``The boy gave his mother flowers,'' ``Charlie gave his son a blank cheque,'' ``Othello gave Desdemona a handkerchief,'' and so on. We might then say that the meaning of each of these sentences is `give', together with specific values filling the case slots, as follows:


agent: {john, the boy, charlie, othello}

recipient: {mary, his mother, his son, desdemona}

object: {apple, flowers, blank cheque, handkerchief}

Among the developments from this form of the semantic network which deserve mention, though we shall not discuss them further here, are Shapiro's distinction between different kinds of relations between nodes, Hendrix's `partitioned semantic networks' for dealing with quantification (as in ``Every boy gave his mother flowers''), and Schank's conceptual dependency representation which translates natural language sentences into their underlying conceptual forms, expressed as conceptual primitives. (For brief but illuminating descriptions of partitioned semantic nets and of conceptual dependency, you could look at Rich, 1983).

Around the same time as Collins and Quillian as well as Simmons and his co-researchers were working on their respective semantic net representations of word meanings and sentence meanings, others were using net-like structures to express types of complex knowledge, though without Quillian's emphasis on cognitive plausibility. Patrick Winston at MIT designed a program which, from net-like structural descriptions of physical structures such as an arch, could infer the concept of arch; Jaime Carbonell adapted Quillian's networks as a data structure for a program called SCHOLAR, which gave tuition on the geography of South America.

[Next] [Up] [Previous]
Next: A Description of Semantic Up: Reasoning with Semantic Nets Previous: Reasoning with Semantic Nets

Cogsweb Project: luisgh@cogs.susx.ac.uk