Getting Computers to Talk Like You and Me
Rachel Reichman, 1985

Artificial Intelligence is about building computer programs to model or simulate intelligent behavior. This book is about one such task, being able to sit down at the computer and tell it what you would like to have done. The computer will need to understand our language and to have a vast amount of knowledge that we use in everyday thought.

In the sixties and seventies, linguists studied how sentences were constructed. Computational linguists built programs to parse sentences and to generate sentences. But when people sit down to talk about something, the conversation spans many sentences. What is needed is a system of rules for dealing with discourses, sequences of many sentences working together to accomplish a given conversational task.

This book will explore some of the complexities which occur when a person tries to communicate with a machine or with another person. A program will be presented which embodies some of the rules needed to carry out that communication.

Chapter 1. The Need for a Theory of Discourse

Is a Linguistic Theory of Discourse Necessary?

A Preliminary Sketch

Chapter 2. Relations between Discourse Elements

Hierarchical Structure of Discourse Relations

The Basic Discourse Constituent: The Context Space

Focused Processing

Discourse Expectations

The Formal Specification of Context Space Components

Chapter 3. Conversational Moves

Because/Like/Like When: Support

So: Restatment and/or Conclusion of Point Being Supported

Incidently/By the Way: Interruption

Anyway/In Any Case: Return

Yes/Right But: Indirect Challenge

(No) But: Direct Challenge

All Right/Okay But: Subargument Concession

Support Challenges

But Look/Listen/You See: Prior Logical Abstraction

Now: Further Development

Methods of Development: Analogy

The Succession of Conversational Moves

Chapter 4. The Discourse Grammar: Context Space Constituents

Basic Components

Types of Context Spaces

Updating the Discourse Model

Chapter 5. Surface Linguistic Phenomena

The Idea of Focus in Recent Theories of Reference

The Context Space Theory of Reference

Focus Level Assignments

Focus Level Assignments During Activation

Examples and Analysis During Activation

Focus Level Reassignment in Suspensions and the Consequences in Resumptions

Chapter 6. The Grammar: An Abstract Process Module

The Augmented Transition Network Paradigm

Alternate Representations of the Discourse Grammar

Discourse Modeling

Updating the Discourse Model: The Registers

Flow Description for Discourse Processing

Violation of Tests

Modeling Discourse Engagement

Appendix 6A. Some Transition Tests

Appendix 6B. Program Representation

Chapter 7. Context Space Suspensions and Resumptions

The Abstract Process Module

Resumption after a Subargument

Conceding a Subargument

Joining Forces

States Accessed in the Simulation

The Trace

Features Underlying the Modeling of Discourse Resumptions

Appendix 7A. Portion of the Grammar Accessed in the Trace

Appendix 7B. Trace Lines 41-46 of the Genetics-Environment Debate

Chapter 8. Cognitive Processing and the Context Space Theory

Hierarchical Structuring of Information

Focused Processing: Selective Attention and Frames of Reference

Discourse Processing and Frame of Reference Identification

Formal Logic and Spontaneous Discourse

Analogies in Spontaneous Discourse

Chapter 9. The Structure of Discourse: On the Generalizability of Discourse Theory

Does Genre Dictate Structure?

Abstract Structure in Other Theories of Discourse

The Context Space's Characterization of Abstract Structure in Discourse

Chapter 10. Linguistic Theory and World Knowledge: An Integration

Communicative Function versus Speaker Intent

The Issue of Text Linguistics

The Integration of Speech Act/World Knowledge and Linguistic Theory

Top of Page | GCTYM Opinion | Sort by Topic | Sort by Title | Sort by Author