WO2002017122A1 - A natural language type system and method - Google Patents

Abstract

In an embodiment of the present invention a natural language database forming method is provided. Text information having a plurality of words (110) is provided and a type for a word in the text information is determined (220). Next an object having the type information is formed and stored (410). In one embodiment the type has quale information, which may have one or more of the following: formal (412), agentive (416), telic (414), or constitutive (418).

Description

A NATURAL LANGUAGE TYPE SYSTEM AND METHOD

CROSS-REFERENCES TO RELATED APPLICATIONS

This application is a continuation-in-part of the following commonly assigned non-provisional patent application, the disclosure of which is herein incorporated by reference in its entirety:

U.S. Patent Application No. 09/433,630 in the names of James D.

Pustejovsky, et al. titled, "A Natural Knowledge Acquisition Method," filed November 16, 1999.

This application claims priority from the following commonly assigned provisional patent application, the disclosure of which is herein incorporated by reference in its entirety:

U.S. Provisional Patent Application No. 60/228,616 in the names of James D. Pustejovsky, et al. titled," Answering User Queries Using A Natural Language Method And

System," filed August 28, 2000 (Attorney Docket No. 019497-000150US).

This application is related to U.S. Patent Application No. in the names of James D. Pustejovsky, et al. titled, " Answering User Queries Using A Natural

Language Method And System,", filed September 14, 2000 (Attorney Docket No. 019497- 000151US); U.S. Patent Application No. 09/449,845 in the names of James D. Pustejovsky, et al. titled, "A Natural Knowledge Acquisition System,", filed November 26, 1999; and to

U.S. Patent Application No. 09/433,630 in the names of James D. Pustejovsky, et al. titled,

"A Natural Knowledge Acquisition Method," filed November 26, 1999; and to U.S. Patent

Application No. 09/449,848 in the names of James D. Pustejovsky, et al. titled, "A Natural Knowledge Acquisition System Computer Code," filed November 26, 1999; and to U.S.

Provisional Patent Application No. 60/163,345 in the names of James D. Pustejovsky, et al. titled," A Method For Using A Knowledge Acquisition System," filed November 3, 1999; and to U.S. Provisional Patent Application No. 60/191,883 in the names of James D.

Pustejovsky, titled," Returning Dynamic Categories in Search and Question- Answer Systems," filed March 23, 2000; and to U.S. Provisional Patent Application No. 60/ in the names of James D. Pustejovsky, et. al. titled, "Type Construction And The Logic Of Concepts," filed August 18, 2000 (Attorney Docket No. 019497-002200US). Each of the applications listed above are assigned to Lexeme, Inc., the assignee of the present invention and each of the above-referenced applications are hereby incorporated by reference.

BACKGROUND OF THE INVENTION

This invention generally relates to the field of natural language understanding. More particularly, the present invention provides techniques for determining lexical semantic information in one or more natural language words.

Traditionally natural language compilation of a textual input, for example, a sentence or sentence fragment, included lexical analysis, syntactic analysis, and semantic analysis. The result being a representation of the meaning of the textual input. For context- free grammars, for example, some computer languages, this approach has been relatively successful. For context-sensitive grammars, for example, English, in which the meaning of a word is dependent upon the sentence or sentence fragment which contains the word, this traditional approach has encountered numerous problems.

Recent research has shown that better results are achieved for natural language interpretation, when semantic information is incorporated during lexical analysis. The open question it is what information should be associated with the lexical unit, e.g., word, during lexical analysis in order to give good results. U.S. Patent Application No. 09/433,630 in the names of James D. Pustejovsky, et al. titled, "A Natural Knowledge Acquisition Method," filed November 26, 1999 gives embodiments which associates a type property in the type system with each lexical entry. Thus, for example, a word after lexical analysis includes a part-of-speech, for example, noun, verb, adverb or adjective, and an aspect of meaning (semantic information), for example quale, associated with it. While the above application gives the basic structure of the semantics that may be associated with a lexical element, the complexities of the English language create a need for improvements and modifications to this basic structure in order to capture the various aspects or senses of a word depending on the context. In addition there is a need for a technique to automatically combine a plurality of words into a single lexical element. Thus creating a lexical element with a more complex semantics, and hence improving the natural language interpretation process. SUMMARY OF THE INVENTION According to the present invention, a technique including a system for acquiring semantic information is provided. In a specific embodiment, the present invention provides a method using a combination of syntactic and semantic information objects. In one embodiment of the present invention a natural language database forming method is provided. Text information having a plurality of words is provided and a type for a word in the text information is determined. Next an object having the type information is formed and placed into an object oriented database. In an embodiment the type has quale information, which may have one or more of the following: formal, agentive, telic, or constitutive. Each of these quale may be direct or indirect.

A second embodiment of the present invention provides a data structure stored in a computer system for determining one or more meanings of a natural language word. The data structure includes: a first entry representing a type in a type structure; and a second entry representing a characteristic of the natural language word, where the characteristic may be a quale.

Another embodiment provides a method for determining a lexical semantic structure used in a computer system. A simple type is selected and a composite entity is generated by combining the simple type with at least one quale. And the result is stored in memory, for example a database. Yet another embodiment provides a method for generating a lexical semantic structure for a plurality of words. First a quale is determined for one of the words and a qualia rule is selected based on the quale selected. Then a lexical semantic structure is generated for the plurality of words using the qualia rule. In an embodiment the plurality of words may include a compound or a preposition.

These and other embodiments of the present invention are described in more detail in conjunction with the text below and attached figures.

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 shows a simplified overview of the natural language system of an embodiment of the present invention. Fig. to 2 illustrates an expanded view of the engine 112 and knowledge resources 114 of an embodiment of the present invention.

Fig. 3 shows a simplified type structure of one embodiment of the present invention. Fig. 4 illustrates the major types of an embodiment of the present invention.

Fig. 5 illustrates an example of a complex entity of an embodiment of the present invention.

Fig. 6 illustrates an example of a simple event, for example verb, of an embodiment of the present invention. Fig. 7 illustrates a multiple meaning for a word for one embodiment of the present invention.

Figs. 8A., 8B, 8C, and 8D show the other senses of the word in a specific embodiment of the invention.

DESCRIPTION OF THE SPECIFIC EMBODIMENTS

Fig. 1 shows a simplified overview of the natural language system of an embodiment of the present invention. A customer provides a corpus 110, e.g., a database of text, usually a large number of smaller documents, each of which has a similar structure. A corpus may range in size, for example, from about 5 MBytes to about 2 GBytes. The customer corpus 110 is input into the natural language engine 112 and is used to create the customer database 116 using the Knowledge Resources 114. Once the customer database 116 has been created, the engine 112 is ready to receive and answer queries from customers. A user at user system 120a and enters a user query 122 which is communicated though a communication network, for example, the Internet 124a, to engine 112. Engine 112 receives the user query 122 and using knowledge resources 114 and customer database 116 returns through the though a communication network, for example, Internet 124b an answer to the user query 130 to user system 120b. Note user system 120a and 120b are the same system and are shown separated in Fig. 1 for illustration purposes only, i.e., to show the flow of user query 122 to user answer 130. Similarly Internet 124a and Internet 124b have been separated for illustration purposes only.

Fig. 2 illustrates an expanded view of the engine 112 and knowledge resources processor of text and can recognize old concepts and phrases and understand new concepts and phrases in questions and then construct customized answers. The engine includes a tokenizer 210, a tagger 212, a stemmer 214, and an interpreter 220. The engine 112 through its interpreter 220 receives information from the knowledge resources 114. The interpreter includes a lexical look-up 222 and a syntactic-semantic composition rules/module 224. The knowledge resources include a lexicon 230 interacting with a type system 232, and grammar rules and roles 234.

The tokenizer 210 takes a text stream composed of punctuation, words, and numbers from a user query coming from 126 or a customer corpus 110 and creates tokenized elements. The tokenizer performs this procedure by first dividing the text into subparts of orthographic words which are unbroken sequences of alphanumeric characters delimited by white space; next, grouping the orthographic words into sentences; and then separating punctuation from words, except where the punctuation should remain part of the word like in abbreviations. The tagger 212 then attaches to each tokenized element a grammatical category or part of speech label based on the Brill ruled-based tagging algorithm. The tagger 212 uses a tag dictionary which has a master list of words with tags. The lexical rules provide a means for the tagger 212 to guess a word and contextual rules provide a means to interpret words and tags according to context. Next the stemmer 214 provides a system name to be used for retrieval for each labeled/tokenized element. The stemmer 212 creates a root form and assigns a numeric offset designating the position in the original text. The stemmer 214 uses a stem dictionary which is a master list of stems.

The interpreter 220 translates the part of speech labels of the tagger 212 into fully specified syntactic categories and uses these new categories with the lexical lookup form of the stemmer 214 to see if the stem already exists in the knowledge resources 114. If the stem exists, the syntactic and semantic information in the lexical entry, for example word, is added to the syntactic category. If the stem is unknown, the interpreter adds default information. The lexical lookup form using, for example, the word's stem, is done by the lexical lookup 222 which interacts with a lexicon 230 and a type system 232. The lexicon 230 includes syntactic concepts (the words in the language) and includes a file for each part of speech. The type system 232 describes semantic concepts. The interpreter 220 also parses (assembles syntactic compositions out of) these categories by applying the grammar rules to combine them into larger syntactic constituents. By applying the grammar rules and the grammar roles 234, and the output of the lexical lookup 222, the interpreter makes a syntactic-semantic composition 224 as it parses. The resulting syntactic-semantic composition 224 is the meaning of the input text stream. An answer based on this syntactic-semantic object is output from the engine 112 at node B 128.

The system described in Fig. 2 is covered in detail in U.S. Patent Application No. 09/449,845 in the names of James D. Pustejovsky, et al. titled, "A Natural Knowledge Acquisition System,", filed November 26, 1999, which is herein incorporated by reference in its entirety.

Fig. 3 shows a simplified type structure of one embodiment of the present invention. This bipartite type structure has a root of "T" 310, which represents the TopType. The first level under root 310 includes entity 312, for example, nouns and adjectives, and event 314, for example, verbs, nouns and adjectives. Entity type 312 then has simple types 318 and complex types 320. Event type 314 has simple types 322 and complex types 324.

Fig. 4 illustrates the major types of an embodiment of the present invention. The root for the type system 232 is given by class GLType 410 (Abstract Class). The root class instance is GLTopType 435. There are two sub classes: GLEntity 440 for entities, which includes nouns and adjectives, and GLEvent 460 for events, which includes nouns, verbs and adjectives. The subclasses GLEntity 440 and GLEvent 460 inherit characteristics, for example, member and member functions, from the parent class, GLType 410. Inheritance as used in object oriented programming is used throughout the type structure.

GLType 410 provides the system template for an abstract characterization of meanings of words, and it includes the following: A. Formal 412 (required): an array. The formal provides a unique identity and is a required field. The formal establishes the type/subtype relation between two types and provides the key for doing inheritance.

B. The following are optional and may or may not be present in GLType:

1. Telic (GLType ) gives the purpose or function. What do I do? What am I for?

2. Agentive (GLType) gives creative factors: How do I come about?

3. Constitutive (GLType) gives a relationship to parts: a. hasElementOf: What specific subparts do I have. b. isElementof : I am a part of another.

4. Entries (dictionary) 420 are words in the lexicon associated with this type in the Lexicon 230. 5. LocalQualia (set) and otherQualia (dictionary) 422 are qualia in addition to formal, constitutive, agentive, and Telic and are an open-ended possibility.

6. name 424: string name of category of type

7. types 428 (dictionary): system generated

8. subtypes 430 (array): system generated list of children In one embodiment, for each GLEntity, there may be one or more of the above qualia (formal is required) but only one of each kind.

In a specific embodiment GLEntity 440 may include zero or more of the following qualia relations:

1. directTelic 442: (GL Event) ~ external argument ~ What do I do? The , "subject" of the GL Event is the one being defined. For example: 'musician' has Formal

[[Human]] and directTelic [[Play Music Activity]].

2. indirectTelic 444: (GL Event) ~ internal argument — What do you do to me?. The "object" of the GL Event is the one being defined. For example: 'trumpet' has formal: [[Musical Instrument]] and indirectTelic: [[Play Music Activity]]. 3. instrumentTelic 446: (GL Event) What am I useful for? For example,

'knife' has instrumentTelic: [[Cut Activity]].

4. Constitutive hasElement 448: (GL Entity) I have a part of which a group is made. For example, 'flock' (birds) hasElement [[Bird]].

5. Constitutive isElementof 450: (GL Entity) I am an inherent part of another. For example of 'ally': isElementof: [[Alliance]].

6. DirectAgentive 452: (GL Event) an external argument of the event specified — To what activity do I give rise? Example: 'pedestrian' : directAgentive: [[Walk activity]].

7. I-ndirectAgentive 454: (GL Event) — What activity gives rise to me? For example: 'book' : indirect Agentive: [[Write activity]].

8. constitutiveRelation 456: (GL Event) -What is the relationship between the stuff I am made of and me? 9. genre (not shown): a grouping of things that have something in common like dept. in a store, types of books, a category in a music store.

In a specific embodiment GLEvent 460 includes one or more of the following: 1. argumentStructure 462: (dictionary) This is a required field that describes the semantic roles of the word and answers the question "Where can I be found in the sentence?"

2. inferredEvents 466 : (dictionary) For example in the phrase: I give the book to Mary, "give" has the inferred event of possession. The argument structure 462 deals with the semantic roles of a word made available by its type by answering the question: "Where will you find each role in the sentence structure?" In one embodiment there are two categories of roles: Semantic roles that go into the Type System 232 and argument roles are properties of the lexical entry. Semantic roles include, but are not limited to, the following: 1. externalArgument: [[Entity]]: who does the action?

2. theme: [[Entity]]: who does it get done to?

3. goal: [[Entity]]: where does the theme go?

4. locative: [[Area]]: where does the action take place?

5. Inferred Events Argument roles indicate where the lexical entry, for example, word, is used in a phrase. In the lexicon, the words include the following mappings to the argument structure: a. subjectRole: maps one of the argument to the subject of the sentence, or a noun to an adjective that modifies it b. objectRole: maps one of the arguments to the object of the sentence c. ppHeadl is one preposition "to", "for", etc., where "pp" stands for prepositional phrase d. ppRolel describes the assignment role that the object of the preposition or clause plays. This is a required field associated with ppHeadl for it tells the system where the roles are when the word is used. For example: "to John" where John is the goal e. clauseRole tells the system how to map the phrase in a sentence. f. clausalComp is an optional field which says to the system, "I need this clause at the same time."

An example of a simple entity 318 of Fig. 3 is the noun "book." It is a subtype of "Readable Representational Artifact" indicated by its formal qualia. The simple entity structure for book may look as follows:

Book (Books)

"a Simple GLEntity" formal: #([[Readable Representational Artifact]]) indirectAgentive: [[Write Activity]] directTelic: [[Describe Relation]] indirectTelic: [[Read Activity]] location: [[Locative Relation]] genre: [[Genre]] medium: [[Communication Medium]]

Fig. 5 illustrates an example of a complex entity of an embodiment of the present invention. Fig. 5 shows a window 510 of the TS and Lexicon Browser tool. There are three sub-windows of interest. A type tree window 512 showing the GLType tree, a lexical entry window 540 showing the lexical entry "mutual fund" 538, and a detailed type window 550 showing a complex type for [[Mutual Fund]] 542. From the type tree window, "Mutual Fund" 514 is a subtype of "Financial Instrument" 516 which is a subtype of Individuated Instrumental Entity 520, which is subtype of Individuated Entity 522, which is a subtype of Entity 524, and which is a subtype of TopType 526. While typically the lexical unit is a single word, it can be more than one word as in this case where the lexical unit is "mutual fund." Note mutual fund is more than concatenation of two meanings "mutual" and "fund," but its meaning includes an investment company performing some function. The values of formal 552 in the detailed type window 550 show that [[Mutual Fund]] has two supertypes [[Company]], which is the priority supertype and [[Financial Instrument]] which is the default supertype.

Fig. 6 illustrates an example of a simple event, for example verb, of an embodiment of the present invention. Fig. 6 shows a window 510 of the TS and Lexicon Browser tool. There are three sub-windows of interest. A type tree window 612 showing the GLType tree and the "Invest Activity" selection 614, a lexical entry window 630 showing the lexical entry "invest" 620 and a detailed type window 640 showing a simple type for [[Invest Activity]] 635. The formal qualia 642 in the detailed type window 640 show a supertype of Business Activity 644a which corresponds to the entry Business Activity 644b in the type tree window 612.

J-n another embodiment of the present invention the types represent one dimension of description of a word. Orthogonal to the type dimension are the qualia dimensions. The qualia (see "The Generative Lexicon," by James Pustejovsky, MIT Press, 1995, which is herein incorporated by reference in its entirety) are:

1. Formal - "that which distinguishes the object within a larger domain;" formal holds the type, which represents the "semantic head." It may have multiple values as an array, for example Complex Types.

2. Constitutive - "the relation between an object and its constituents, or proper parts;" This includes constitutive group, which is a constitutive with group like properties. In another embodiment constitutive is replaced by a) directConstitutive: a relational role which holds a GLEvent, generally a [[Constitutive Relation]] or similar relation. The entity for which

#hasElement was previously defined is the #externalArgument of this relation. b) indirectConstitutive: is a relational role which holds a GLEvent, generally a [[Constitutive Relation]] or similar relation. The entity for which #hasElement was previously defined is the #theme of this relation.

3. Telic - (GLEvent) "purpose and function of the object;" This may include subcategories, for example: a) indirectTelic : the entity being defined is the internal argument of the event specified. b) directTelic : the entity being defined is the external argument of the event specified. c) instrumentTelic: for entities whose Telic is a three place relation and which are allowed to appear in subject position. d) purposeTelic: for entities with a Telic expressing the semantics of

"purpose" but which do not occupy an argument of the event specified. 4. Agentive - (GLEvent) "factors involved in the origin or 'bringing about' of an object." This may include subcategories, for example: a) indirectAgentive : the entity being defined is the internal argument of the event specified. b) directAgentive : the entity being defined is the external argument of the event specified.

One embodiment uses the above qualia to define, for example, 18 qualia relations

1. Constitutive

2. Constitutive Group (this is a constitutive with group like properties)

3. Direct Telic

4. Indirect Telic

5. Direct Agentive

6. Indirect Agentive

7. Direct Telic, Direct Agentive (this qualia indicates there is both a direct Telic and a direct agentive relation)

8. Indirect Telic, Direct Agentive

9. Direct Telic, Direct Agentive

10. Indirect Telic, Indirect Agentive

11. Constitutive, Direct Telic, Direct Agentive

12. Constitutive, Indirect Telic, Direct Agentive

13. Constitutive, Direct Telic, Direct Agentive

14. Constitutive, Indirect Telic, Indirect Agentive

15. Constitutive Group, Direct Telic, Direct Agentive

16. Constitutive Group, Indirect Telic, Direct Agentive

17. Constitutive Group, Direct Telic, Direct Agentive

18. Constitutive Group, Indirect Telic, Indirect Agentive

An alternative embodiment may include additional qualia, for example: a) #genre b) #medium c) #location (this has been turned into an Event quale already) d) #institution e) #madeOf

In addition to and independent from the qualia given above, the type structure of Fig. 3 in one embodiment of the present invention may be expanded to include simple (or natural) types and complex types for entities, and simple (or natural) types and complex types for events. An example of a type hierarchy for Simple entities is: Physical Entity

1.1. Individuated Entity

1.1.1. Organic Entities

1.1.1.1. Animate Living Entity

1.1.1.1.1. Human

1.1.1.1.1.1. Female Person

1.1.1.1.1.2. Male Person

1.1.1.1.1.3. Young Person

1.1.1.1.2. Animal

1.1.1.2. Inanimate Living Entity 1.1.1.2.1. Plant

1.1.2. Inorganic Entity

1.1.2.1. Material Object

1.1.2.2. Virtual Entity

1.2. Geographical Area

1.2.1. Area

1.2.2. Country

1.3. Mass Entity

1.3.1. Organic Substance

1.3.2. Material (viz. Inorganic Substance)

1.3.2.1. Liquid Substance

1.3.2.2. Gas Substance

1.3.2.3. Solid Substance

2. Abstract

2.1. Mental

2.2. Psychological

2.3. Proposition

2.4. Social

3. Property

3.1. Measure

3.1.1. Weight

3.1.2. Time %distinguished by constitutive

3.1.2.1. Period

3.1.2.1.1. Month

3.1.2.1.2. Day

3.1.2.1.3. Year

3.1.2.1.4. Season (??Functional?)

3.1.2.1.5. Decade

3.1.2.2. Point

3.1.2.2.1. Date

3.1.3. Nationality

4. Sound Complex entities represent "dot objects" (see U.S. Provisional Patent Application No. 60/ in the names of James D. Pustejovsky, et. al. titled, "Type Construction And The Logic Of Concepts," filed August 18, 2000 (Attorney Docket No.

019497-002200US), which is herein incorporated by reference in its entirety. This typically means that the formal quale is a logical pairing of the senses denoted by the individual types in the complex type. For example in Fig. 5, the formal for "mutuaLfunds" has both a super type of Company and Financial Instrument. From an object oriented perspective this is similar to multiple inheritance. Examples of Complex types are physical_object.information, e.g., "book," "records"; eventevent, e.g., "construction;" eventquestion, e.g. "exam," animakrational, e.g. "person." An example of a type hierarchy for complex entities is:

1. Event.Time 2. Even Physical

3. Even Representational Artifacts

3.1. Event . Topic/Information Content

3.1.1. News

3.1.1.1. Interview 3.1.2. Event . Visual Components

3.1.2.1. Scene

3.1.3. Event . Sound/Audible Components 3.1.3.1. Music Event

3.1.4. Artifact . Text Components 3.1.4.1. Printed media

4. ArtifactRepresentation

4.1. Artifact . Topic/Information Content

4.1.1. Art 4.1.2. History 4.1.3. Biography

4.1.4. Artifact . Visual Components 4.1.4.1. Visual Art

4.1.4.1.1. Map

4.1.4.1.2. Video 4.1.5. Artifact . Sound/ Audible Components

4.1.5.1. Music 4.1.6. Artifact . Text Components

4.1.6.1. Printed Media

4.1.6.2. Book //Expand by subject 4.1.6.2.1. Art Book

4.1.6.2.2. TextBook

4.1.6.3. Directory 4.1.6.3.1. Catalog

5. Artifact.Time 5.1. Calendar

6. Organization. Architectural Object

6.1. Organization

6.1.1. Music 6.1.2. Medicine

6.1.3. Arts 6.1.3.1. Gallery

6.1.4. Business

6.1.4.1. Shop//Expand by product sold 6.1.4.1.1. Bookstore

6.1.5. Sports

6.1.6. Education 6.1.6.1. University

An example of a type hierarchy for simple events is:

1.1. State

1.1.1. Existence Predicate

1.1.2. Physical State 1.1.3. Mental State

1.2. Relational State

1.2.1. Possessive State

1.2.2. Locative State

1.2.3. Relational Physical State 1.2.4. Relational Mental State

1.3. Process %%one place, subtyped according to external argument 1.3.1. Physical Process

1.3.1.1. Motion Activity

1.3.1.2. Communicate Activity (talk, speak, chat) 1.3.2. Mental Process

1.3.3. Event Process

1.3.4. Time Process

1.4. Relational Process

1.4.1. Cause 1.4.1.1. Transfer

1.4.2. Change

1.5. Aspectual

1.5.1. Begin Activity

1.5.2. Continue Activity 1.5.3. Terminate Activity

In an embodiment complex events are generated similar to complex entities, using the dot constructor. In another embodiment complex events are generated similar to the complex event structure of page 69 of "The Generative Lexicon," by James Pustejovsky, MIT Press, 1995.

In order to determine a "lexical semantic structure," i.e., sense or meaning, for each lexical element, e.g., word, each simple type or complex type is combined with one or more qualia. For simple types combined with one or more qualia, these are called composite types. Appendix A gives an example of entity and event composite types for an embodiment of the present invention. For event types, however, qualia roles are "distributed" somewhat differently that entity types, in that they substitute the argument structure representation, such that for a given event each qualia role expresses the "role" that an entity plays in that event. Fig. 7 illustrates a multiple meaning for a word for one embodiment of the present invention. In Fig. 7 the TS and Lexicon browser 510 shows five different senses or meanings for the word "stock:" 712, 714,716, 718, and 720. The highlighted selection for "stock" 712 is shown in the lexical entry window 721. In the lexical entry window 721 "stock" is a "verb entry" 722 and has a type "supply activity" 723. In the type window 725 the type "supply activity" is a simple GLEvent 726 with a super type or formal "give Activity" 727. Since "supply activity" is an event, it has an argument structure 728.

Figs. 8A., 8B, 8C, and 8D show the other senses of the word in a specific embodiment of the invention. In these figures "stock" 714, 716, 718, and 720 are a "Noun Entry" (730, 740, 750, 760) and a "Simple GLEntity" type. Stock 714 has a type "stocking" 732. Stock 716 has a type "financial instrument "742. Stock 718 has a type "soup" 752. And stock 720 has a type "merchandise" 762. Thus the four other senses for stock are stocking, e.g. as in socks, financial instrument, e.g., such as a stock certificate, soup, e.g., such as chicken stock and stock, e.g., as is in putting a product on the shelf for selling. The lexical semantic structure for "stock" includes the plurality of different senses for this specific word. Which sense is picked may depend on the context in which the word is used. For example in the context of "Recipes for soup," "stock" 718 as used in Fig. 8C may be chosen.

In a specific embodiment qualia rules generate one lexical semantic structure for the plurality of words in a compound or preposition. This may, for example, bind a qualia of the modifier to the head or may use two lexical semantic structures to generate a new lexical semantic structure. The rules for compounds and prepositions permit the treatment of a great number of nouns relationally. They represent a way to:

1. make explicit the event information that is implicitly carried by a nominal;

2. relate a query expressed in a compound form to a query expressed in other relational forms.

In the case of compounds, the semantic rules typically reference the qualia of the head and/or modifier. In the case of prepositions, there are semantic rules for modification of LexLFs without qualia (including argument binding of functionLexLFs) that make no reference to qualia.

In one embodiment the following rules are examples of qualia rules. Note MOD stands for modifier. And it is assumed that satisfaction of a type restriction includes being a subtype of the restriction, and that satisfaction of a qualia restriction includes being a subtype of the event specified in the qualia restriction. Telic Related Rules

BindThemeOfTelic ('Direct Telic Rule'): The rule binds the modifier noun to the theme of the #directTelic of the head. Compounds:

Examples: "piano store," "geology book" Conditions:

MOD: [[Entity]] type HEAD: [[Entity]] type with #directTelic Prepositions:

Examples: "book about geology," "pictures of Africa" Conditions

HEAD: Any specified [[Entity]] with #directTelic OBJECT: any specified [[Entity]]

BindlndirectTelic (Indirect Telic Rule'): The rule binds the modifier noun to the #externalArgument of the #indirectTelic of the head (E.g. children's books, adolescent literature). Very rare for compounds. Compounds: Example: "adolescent literature"

Conditions

MOD: [[Entity]] type HEAD: [[Entity]] type with #indirectTelic Prepositions: Example: "books for children"

NOTE: this may be lexically specified. Conditions HEAD: Any specified [[Entity]] with #indirectTelic OBJECT: any specified [[Entity]] BindThemeOflndTelic/ SubtvpeThemeOflndirectTelic

Compound: The rule binds the modifier noun to the theme of the #indirectTelic of the head (E.g. book supply, book purchases). It is equivalent to replacing/subtyping the already implicitly bound theme. Conditions

MOD: [[Entity]] type as long as it is a subtype of the #theme selected by the relation in the #indirectTelicHEAD: [[Entity]] type with #indirectTelic

Preposition: The rule binds the object of the preposition to the theme of the #directTelic of the governor/head (E.g. book about geology, pictures of Africa).

Conditions

HEAD: Any specified [[Entity]] with #directTelic OBJECT: any specified [[Entity]]

InstrumentTelic Rule

Co-composition: The modifier has #instrumentTelic which contains a value that is the same as the #indirectAgentive or of #instrumentTelic of the Head. Bind the modifier into the instrument role of either the #indirect Agentive or of the #instrumentTelic of the head. Example:, oven pizza, microwave recipes. Conditions

MOD: Any [[Entity]] type with #instrumentTelic HEAD: Any type with #instrumentTelic or #indirectAgentive Qualia setting

Specialize Telic -> Purpose Telic: Create new #purposeTelic for the Head: e.g., learning tools. For argument binding copy any other bound arguments to their equivalent roles in the event expressed by the #purposeTelic.

For example to generate one lexical semantic structure for "hunting rifle," the lexical semantic structures are first obtained from the system.

Next the Specialize Telic -> Purpose Telic rule is applied to combine "rifle" and "hunting" to "hunting rifle:"

Other examples:

E.g., recording studio, production studio, trading center, catering service Hunting rifle, toys for learning; Conditions

MOD: Event nominal HEAD: any [[Entity]] type (may need restrictions)

AGENTIVE RELATED RULES

BindExternalArgAgentive Rule

Compounds

Bind Modifier to the external argument of the #agentive

Example: "community event," "witch forum"

Conditions

MOD: Entity HEAD : Entity with #indirect Agentive Prepositions

Examples: "books by Frank," "the paintings of Emil Nolde"

Conditions:

MOD: Entity HEAD: Entity or Event Nominal with #indirect Agentive

BindThemeOfAgentive Rule: Bind Modifier to the theme of the #agentive Examples: "apple tree"

(Note that with agentive nominals in "-er" the role of the directTelic is usually the same event)

Conditions

MOD: Material Object OR Individuated Entity

HEAD: Individuated Entity indirectAgentive

Bind the pre-modifier to the external Argument of the relation in the #indirectAgentive

Examples: restaurant menu, store catalog, library resource Conditions MOD: Institution

HEAD: Any Type OTHER QUALIA

Location Rule: Bind GeoPolical Area to #location For example: Boise Artist, area band, Cambridge shop

Conditions

MOD: Geopolitical Area

HEAD: Any type with #location quale Genre Rule: Bind the modifier to the #genre quale of the head Conditions

MOD: Genre

HEAD: Any type with #genre quale Example: Jazz Music

Made of Rule

Bind Modifier to the type in the #madeOf quale of the head

E.g., wood product, glass item, seafood dish Conditions

MOD: Material Entity

HEAD: Any type with #madeOf quale Institution Rule: Bind Modifier to the type in the #institution quale of the head

For example: MIT professor Conditions

MOD: Institution

HEAD: Any type with institution quale

Level Rule

Bind Modifier to the type in the #level quale of the head

E.g., elementary school

Conditions

MOD: Education Category HEAD: Any type with #level quale Example: Elementary School

Elementary School

In one embodiment there are certain ordering constraints on the application of the rules of compounding. The reason is to filter out the results of the composition due to possible overgeneration. The constraints on the compound rules may include one or more of the following:

1. If a location rule or a genre rule applies then do not try any other rule.

2. If a modifier is 'Prize' and head is a subtype of Human, bind the head to the externalArgument of the indirect Telic of 'Prize'

3. If modifier is 'Legal Entity' or 'Human' and the head has a Telic with a three place relation (e.g. program -> 'Plan Activity' externalArgument: 'Legal Entity' theme: 'Event', goal 'Legal Entity'), bind the modifier to the goal.

4. If head is 'Group' or 'Club' and the modifier has an indirect Telic, assign the value of the Telic of the modifier to the directTelic of the head.

5. If head has a qualia role #medium and the pre-modifier is of type 'Communication Medium', bind pre-modifier to the qualia.

6. If pre-modifier is an individuated Event and the head is 'Institution' or 'Functional Location' apply subtype Telic first.

7. If the head has an instrumentTelic, bind the pre-modifier to the role that matches the event in the Telic. Conclusion

Although the above functionality has generally been described in terms of specific hardware and software, it would be recognized that the invention has a much broader range of applicability. For example, the software functionality can be further combined or even separated. Similarly, the hardware functionality can be further combined, or even separated. The software functionality can be implemented in terms of hardware or a combination of hardware and software. Similarly, the hardware functionality can be implemented in software or a combination of hardware and software. Any number of different combinations can occur depending upon the application. Many modifications and variations of the present invention are possible in light of the above teachings. Therefore, it is to be understood that within the scope of the appended claims, the invention may be practiced otherwise than as specifically described.

APPENDIX A

This appendix illustrates an example of a composite entity type structure and a composite event type structure.

1. COMPOSITE ENTITY TYPES : are derived by combining one or more qualia with one of the simple entity types.

1.1. Natural Type Constitutive 1.1.1. Part

1.2. Natural Type Constitutive Group 1.2.1. Group

1.3. Natural Type Dir Telic 1.3.1. Tool

1.4. Natural Type Ind Telic 1.4.1. Modal

1.5. #([[Natural Type]] [[Dir Agentive]]) 1.5.1. Causer

1.6. Natural Type Ind Agentive 1.6.1. Product

1.7. Top 2-level_indirect_telic_indirect_agentive

1.7.1.1. Business

1.7.1.1.1. Merchandise

1.8. Physical- 1 -level_direct_telic 1.8.1. Medium

1.9. Group l-level_direct_telic

1.10. Human Group

1.10.1. Human Group Modal

1.10.1.1. Human Group Music

1.10.1.1.1. A Cappella Group

1.10.1.1.2. Ska Band

1.10.1.1.3. Punk Band

1.10.1.1.4. Rock Band

1.10.1.1.5. Crank Band

1.10.1.2. Human Group Education 1.10.1.2.1. Faculty

1.11. Human Constitutive [[ Human ]] [[Constitutive]]

1.11.1. Family Role %tlιis accounts for all relational nouns: sister of, father of

1.11.2. Group Member

1.12. Human Functional Stage [[ Human ]] [[ Dir Telic ]] % expand by DOMAIN

1.12.1. Instruction Stage (2^nd grader, beginner, student)

1.12.2. Discourse Role

1.13. Human Creator [[ Human ]] [[ Dir Telic ]]

1.13.1. Art Creator

1.13.1.1. Music Creator

1.13.1.2. Visual Art Creator

1.13.1.3. Text Info Creator

1.13.2. Crafts Creator

1.14. Human Doer [[ Human ]] [[ Dir Telic ]] 1.14.1. Music Doer 1.14.1.1. Player

1.14.2. Textual Info Doer

1.14.3. Discourse Role (speaker, hearer)

1.14.4. Business Role 1.14.4.1. Buyer

1.14.5. Expert Role

1.14.5.1. Appraiser

1.14.6. Habit (e.g. smoker)

1.14.7. Occupation

1.15. Human [[ Human ]] [[ Ind Telic ]] %this accounts for nouns: such as "slave"

1.16. Human [[ Human ]] [[ Dir Agentive ]] %aspectual nouns: such as "pedestrian"

1.16.1. Progressive (pedestrian)

1.16.2. Perfective (murderer)

1.17. Animal l-level_indirect_telic 1.17.1. Draft Animal

1.18. Substance l-level_direct_telic

1.19. Substance l-level_indirect_telic

1.20. Substance l-level_direct_agentive

1.21. Substance l-level_indirect_agentive

1.22. Substance 2-level_direct_telic_direct_agentive

1.23. Substance 2-level_direct_telic_indirect_agentive

1.24. Substance 2-level_indirect_telic_direct_agentive

1.25. Substance 2-level_indirect_telic_indirect_agentive

1.26. Substance 3-level_constitutive_direct_telic_direct_agentive

1.27. Substance 3-level_ constitutive_direct_telic_indirect_agentive

1.28. Substance 3-level_ constitutive_indirect_telic_direct_agentive

1.29. Substance 3-level_ constitutive _indirect_telic_indirect_agentive

1.30. Material Object Ind Telic 1.30.1. Natural_tools (e.g. lever)

1.31. Artifact [[Material Object]] [[Ind Agentive]]

1.32. Artifact Tool [[Artifact]] [[Tool]]

1.32.1. Musical Instrument

1.32.2. Music Artifact

1.32.3. Pharmaceutical %%for curing diseases

1.32.4. Transportation Objects %%vehicles of all sorts

1.32.5. Food Object %% edible objects

1.32.6. Clothing Object %% wearable objects

1.32.7. Art Object

1.32.7.1. Installation

1.32.7.2. Mask

1.33. Artifact Tool Group [[Artifact Tool]] [[Group]]

1.34. Communication Medium [[Medium]] [[Representational Artifact]]

1.34.1. Television

1.34.2. Radio

1.35. Artifact Tool Part [[Artifact Tool]] [[Part]] 1.35.1. Artifact Tool Part Music

1.36. Architectural Object

1.36.1. Building

1.36.2. Artifact Medium

1.36.2.1. Music nedium

1.36.3. Artifact_Container

1.37. Obj ect 3 -le vel_constitutive_direct_telic_indirect_agenti ve

1.38. Obj ect 3 -level_constitutive _indirect_telic_indirect_agentive

1.38.1. Artifact Part %%these are parts of things which have a correspondence in the Artifact object type, (carburetor, computer keyboard, amplifier, staircase)

1.38.1.1. Pharmaceutical Part %%for curing diseases

1.38.1.2. Transportation Objects Part %%vehicles of all sorts

1.38.1.3. Food Object Part %% edible objects

1.38.1.4. Clothing Object Part %% wearable objects

1.38.1.5. Music Object Part %% things that allow people to listen to music "records"

1.38.2. Architectural Object Part

1.39. Object - group 3-level_constitutive _indirect_telic_indirect_agentive 1.39.1. Craft

1.39.1.1. Jewelry

1.40. Virtual Object 1-level _const 2-level_direct_telic_indirect_agentive 1.40.1. Website

1.41. Virtual Object 3-level_constitutive_direct_telic_indirect_agentive 1.41.1. Webpage

1.42. Geographical Area l-level_indirect_telic 1.42.1. Functional Location

1.43. Geographical Area 2-level_indirect_telic_indirect_agentive 1.43.1. Architectural Functional Location

1.44. Abstract 3-level_constitutive_direct_telic_indirect_agentive 1.44.1. Organization Part

1.45. Mental 1 -level _const

1.46. Mental l-level_direct_telic

1.47. Mental l-level_indirect_telic

1.48. Mental 1 -level_direct_agentive

1.49. Mental l-level_indirect_agentive

1.50. Mental 2-level_direct_telic_direct_agentive

1.51. Mental 2-level_direct_telic_indirect_agentive 1.51.1. Mental_artifacts (e.g. idea, though)

1.52. Mental 2-level_indirect_telic_direct_agentive

1.53. Mental 2-level_indirect_telic_indirect_agentive

1.54. Mental 3-level_constitutive_direct_telic_direct_agentive

1.55. Mental 3 -level_ constitutive_direct_telic_indirect_agentive

1.56. Mental 3-level_ constitutive_indirect_telic_direct_agentive

1.57. Mental 3-level_ constitutive _indirect_telic_indirect_agentive

1.58. Psychological 1 -level _const

1.59. Psychological l-level_direct_telic 1.60. Psychological 1 -level_indirect_telic 1.61. Psychological 1 -level_direct_agentive 1.62. Psychological 1 -level_indirect_agentive

1.62.1. Feelings

1.63. Psychological 2-level_direct_telic_direct_agentive 1.64. Psychological 2-level_direct_telic_indirect_agentive 1.65. Psychological 2-level_indirect_telic_direct_agentive 1.66. Psychological 2-level_indirect_telic_indirect_agentive 1.67. Representational Artifact [[ Representation ]] [[ Artifact Tool]]

1.67.1. Sample

1.68. Collective Representation [[Representational Artifact]] [[Group]]

1.68.1. Art

1.68.2. History

1.69. Readable Representation Artifact

1.69.1. Book 1.70. Audible Representation Artifact

1.70.1. Music 1.71. Visual Representation Artifact

1.71.1. Photograph 1.72. Communicative Artifact

1.72.1. Communicative Artifact Event

1.72.2. Communicative Artifact Place

1.72.3. Communicative Artifact Time

1.72.4. Communicative Artifact People

1.72.5. Communicative Artifact Music

1.72.6. Communicative Artifact Object

1.73. Time l-level_direct_telic %%expand by domain

1.73.1. Season

1.73.2. Fiscal year

1.74. Measure 2-level_direct_telic_indirect_agentive

1.74.1. Monetary Value (e.g. price)

.75 Measure 2-level_indirect_telic_direct_agentive .76 Measure 2-level_indirect_telic_indirect_agentive .77 Measure 3 -level_constitutive_direct_telic_direct_agentive .78 Measure 3 -level_constitutive_direct_telic_indirect_agentive .79 Measure 3 -level_constituti ve_indirect_telic_direct_agentive .80 Measure 3 -level_constitutive_indirect_telic_indirect_agentive .81 Social Product [[ Social ]] [[ Product ]] .82 Social Product [[ Social Product ]] [[ Modal ]]

1.82.1. Organization Place [[ Social Product Modal ]] [[ Human Group Modal

]]

1.83. Company [[Organization Place]] [[ Architectural Object ]]

1.83.1.1. Company

1.83.1.2. Store

1.83.1.3. Art Place

1.83.1.4. Entertainment Place COMPOSITE EVENT TYPES: are derived by combining one or more qualia with one of the simple entity types:

2.1. event l-level_constitutive

2.2. event 1 -level _direct_telic

2.3. event 1 -level _indirect_telic

2.4. event 1-level _direct_agentive 2.4.1. Cause Change Activity

2.5. event 1-level _indirect_agentive

2.6. event 2-level_direct_telic_direct_agentive

2.7. event 2-level_direct_telic_indirect_agentive

2.8. event 2-level_indirect_telic_direct_agentive

2.9. event 2-level_indirect_telic_indirect_agentive

2.10. event 3-level_constitutive_direct_telic_direct_agentive

2.11. event 3-level_constitutive_direct_telic_indirect_agentive

2.12. event 3-level_constitutive_indirect_telic_direct_agentive

2.13. event 3 -level_constitutive_indirect_telic_indirect_agentive

2.14. State

2.14.1. Existence Predicate

2.14.1.1. existence predicate 1 -level_constitutive

2.14.1.2. existence predicate 1 -level _direct_telic

2.14.1.3. existence predicate 1 -level _indirect_telic

2.14.1.4. existence predicate 1 -level _direct_agentive

2.14.1.4.1. Self Creation Activity

2.14.1.4.2. Change of State

2.14.1.5. existence predicate 1 -level _indirect_agentive 2.14.1.5.1. Create Activity

2.14.1.6. existence predicate 2-level_direct_telic_direct_agentive

2.14.1.7. existence predicate 2- level_direct_telic_indirect_agentive

2.14.1.8. existence predicate 2- level_indirect_telic_direct_agentive

2.14.1.9. existence predicate 2- level_indirect_telic_indirect_agentive

2.14.1.10. existence predicate 3- level_constitutive_direct_telic_direct_agentive

2.14.1.11. existence predicate 3- level_constitutive_direct_telic_indirect_agentive

2.14.1.12. existence predicate 3- level_constitutive_indirect_telic_direct_agentive

2.14.1.13. existence predicate 3- level_constitutive_indirect_telic_indirect_agentive

2.14.2. Physical State

2.14.2.1. physical state 1 -level_constitutive

2.14.2.2. physical state 1-level _direct_telic

2.14.2.3. physical state 1-level _indirect_telic

2.14.2.4. physical state 1-level _direct_agentive 2.14.2.4.1. Change Physical State

2.14.2.5. physical state 1-level _indirect_agentive 2.14.2.5.1. Create Physical State Activity

2.14.2.6. physical state 2-level_direct_telic_direct_agentive

2.14.2.7. physical state 2-level_direct_telic_indirect_agentive

2.14.2.8. physical state 2-level_indirect_telic_direct_agentive

2.14.2.9. physical state 2-level_indirect_telic_indirect_agentive

2.14.2.10. physical state 3- level_constitutive_direct_telic_direct_agentive

2.14.2.11. physical state 3- level_constitutive_direct_telic_indirect_agentive

2.14.2.12. physical state 3- level_constitutive_indirect_telic_direct_agentive

2.14.2.13. physical state 3- level_constitutive_indirect_telic_indirect_agentive

2.14.3. Mental State

2.14.3.1. mental state 1 -level_constitutive

2.14.3.2. mental state 1-level _direct_telic

2.14.3.3. mental state 1-level _indirect_telic

2.14.3.4. physical state 1-level _direct_agentive 2.14.3.4.1. Change Mental State

2.14.3.5. mental state 1-level _indirect_agentive 2.14.3.5.1. Create Mental Activity

2.14.3.6. mental state 2-level_direct_telic_direct_agentive

2.14.3.7. mental state 2-level_direct_telic_indirect_agentive

2.14.3.8. mental state 2-level_indirect_telic_direct_agentive

2.14.3.9. mental state 2-level_indirect_telic_indirect_agentive

2.14.3.10. mental state 3- level_constitutive_direct_telic_direct_agentive

2.14.3.11. mental state 3- level_constitutive_direct_telic_indirect_agentive

2.14.3.12. mental state 3- level_constitutive_indirect_telic_direct_agentive

2.14.3.13. mental state 3- level_constitutive_indirect_telic_indirect_agentive

2.15. Relational State

2.15.1. Possessive State

2.15.1.1. possessive state l-level_constitutive

2.15.1.2. possessive state 1 -level _direct_telic

2.15.1.3. possessive state 1-level _indirect_telic

2.15.1.4. possessive state 1-level _direct_agentive 2.15.1.4.1. Acquire Possessive State

2.15.1.5. possessive state 1-level _indirect_agentive 2.15.1.5.1. Receive Possessive State

2.15.1.6. possessive state 2-level_direct_telic_direct_agentive

2.15.1.7. possessive state 2-level_direct_telic_indirect_agentive

2.15.1.8. possessive state 2-level_indirect_telic_direct_agentive 2.15.1.9. possessive state 2-level_indirect_telic_indirect_agentive

2.15.1.10. possessive state 3- level_constitutive_direct_telic_direct_agentive

2.15.1.11. possessive state 3- level_constitutive_direct_telic_indirect_agentive

2.15.1.12. possessive state 3- level_constitutive_indirect_telic_direct_agentive

2.15.1.13. possessive state 3 - level_constitutive_indirect_telic_indirect_agentive

2.15.2. Locative State

2.15.2.1. locative state l-level_constitutive

2.15.2.2. locative state 1-level _direct_telic

2.15.2.3. locative state 1 -level _indirect_telic

2.15.2.4. locative state 1 -level _direct_agentive 2.15.2.4.1. Acquire Locative State

2.15.2.5. locative state 1-level _indirect_agentive 2.15.2.5.1. Receive Locative State

2.15.2.6. locative state 2-level_direct_telic_direct_agentive

2.15.2.7. locative state 2-level_direct_telic_indirect_agentive

2.15.2.8. locative state 2-level_indirect_telic_direct_agentive

2.15.2.9. locative state 2-level_indirect_telic_indirect_agentive

2.15.2.10. locative state 3- level_constitutive_direct_telic_direct_agentive

2.15.2.11. locative state 3- level_constitutive_direct_telic_indirect_agentive

2.15.2.12. locative state 3- level_constitutive_indirect_telic_direct_agentive

2.15.2.13. locative state 3 - level_constitutive_indirect_telic_indirect_agentive

2.15.3. Relational Physical State

2.16. Relational Mental State

2.17. Change of State

2.18. Functional Activity [formal, telic: process, agentive: cause] %distinguished by Telic

2.18.1. Service

2.19. Speech Acts

Claims

WHAT IS CLAIMED IS

1. A natural language database forming method, said method comprising: providing text information comprising a plurality of words; determining a type for a word in said text information; forming an object comprising said type information; and placing into an object oriented database said object.

2. The method of claim 1 wherein said type inherits type information from a parent type.

3. The method of claim 1 wherein said type comprises quale information.

4. The method of claim 3 wherein said quale comprises a formal quale.

5. The method of claim 1 further comprising: receiving a user query, comprising a plurality of words; and selecting an object using subsumption to achieve an answer to said user query.

6. A data structure stored in a computer system for determining one or more meanings of a natural language word, comprising: a first entry representing a type in a type structure; and a second entry representing a characteristic of the natural language word.

7. The data structure of 6 wherein said characteristic comprises a quale selected from a group consisting of formal, agentive, telic, or constitutive.

8. The data structure of 6 wherein said type structure comprises an entity type and an event type.

9. The data structure of 8 wherein said entity type includes a complex type.

10. The data structure of 9 wherein said complex type is generated using a dot object.

11. The data structure of 8 wherein said event type includes a complex type.

12. The data structure of 8 wherein said entity type includes a noun.

13. The data structure of 8 wherein said event type includes a verb.

14. The method of claim 8 wherein the quale comprises a selection from a group consisting of formal, directAgentive, indirectAgentive, directTelic, indirectTelic, directConstitutive, or indirectConstitutive.

15. A method for determining a lexical semantic structure used in a computer system, comprising: selecting a simple type; generating a composite entity by combining said simple type with at least one quale; and storing said composite entity in a memory system.

16. The method of claim 15 wherein a quale comprises a selection from a group consisting of formal, directAgentive, indirectAgentive, directTelic, indirectTelic, directConstitutive, or indirectConstitutive.

17. The method of claim 15 wherein said simple type includes a subtype of physical entity.

18. A method for generating a lexical semantic structure for a plurality of words comprising: determining a quale of one of the words; selecting a qualia rule based on said quale; and generating said lexical semantic structure for the plurality of words using said qualia rule.

19. The method of claim 18 wherein said plurality of words include a compound.

20. The method of claim 18 wherein said plurality of words include a preposition.