fido.linguistic
Class NounResolver

java.lang.Object
  |
  +--fido.linguistic.NounResolver

public class NounResolver

extends java.lang.Object

Resolve a NounPhrase to one instance in the ObjectTable.

See Also:

ObjectTable

Constructor Summary

NounResolver(Discourse discourse)
Creates a new NounResolver instance.

Method Summary

ResolvedNoun

resolve(NounPhrase np) Searches for an instance in the ObjectTable that the NounPhrase representes.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

NounResolver

public NounResolver(Discourse discourse)

Creates a new NounResolver instance.

Parameters:

discourse - Discourse object used in resolving nouns

Method Detail

resolve

public ResolvedNoun resolve(NounPhrase np)
                     throws FidoException

Searches for an instance in the ObjectTable that the NounPhrase representes. Each part of the NounPhrase changes the search:

• Noun - The noun class will have the name of the class in the Noun hierarchy that was assigned by the Grammar module from the Dictionary table. This is the starting point. This is a class, never an instance.
• Article - The two types of articles:
  • Definite - Signifies the noun should be resolved to an instance that is already defined in the discourse. For example, the dog is outside refers to a specific instance of a dog, not just any dog, but one the user and the computer has referred to before.
  • Indefinte - Signifies a new instance. For example, A dog bit me refers to an actual dog, but one that the user and computer has not spoken about previously. This will create a new instance of dog within the hierarchy.
  • Along with the two above rules, there is an exception called generics. Generics allow a noun to refer to a whole class of nouns. For example:
    • Lions are dangerous.
    • The lion is a dangerous animal.
    • A lion can be dangerous.
    In the above, lion refers to all lions, not just a specific one. See the section below called Sentence Patterns.
• Adjectives - If the article was a definite article, adjectives can narrow the number of matching instances by applying constraints. For example, if the system has a couple of dog instances in the hierarchy, the phrase the white dog narrows the number of dogs by putting a constraint on the fur color. Indefinite articles causes adjectives to be added as attributes to newly created instances.
• Prepositional Phrases - Prepositional phrases perform the same function as adjectives, but establish a relationship between the main noun and the noun object of the preposition. For example the dog with the blue collar establishes a HASA relationship between the dog and a blue collar. The system will search for any dog instance with a HASA relationship with the blue collar noun instance.

Sentence Patterns

Sentence Pattern	N1 Type	N2 Type	Example
the N1 AV the N2	EI	EI	The shark eat the seal
the N1 AV a N2	EI	CI	The man bought a car
a N1 AV the N2	CI	EI	A boy broke the window
a N1 AV a N2	CI	CI	A lion killed a gazel
the N1 LV the N2	EI	EI	The problem is the process
the N1 LV a N2	EI	C	The dog is a Poodle
a N1 LV a N2	C	C	A lion is a cat
PN1 LV PN2	C	C	Wolves are hunters
PN1 AV PN2	C	C	Sharks attack seals

Legend: LV = linking verb, AV = action verb, PN = plural noun, EI = existing instance, CI = create instance, C = class

Parameters:

np - NounPhrase to resolve

Returns:

A ResolvedNoun class that the NounPhrase represents

Throws:

FidoException - General exception thrown if the NounPhrase could not be processed. The exception will contain a wrapped exception that is the root cause.