Control of Mixed-Initiative Discourse
Through Meta-Locutionary Acts:
A Computational Model
David Graham Novick
CIS-TR-88-18
December 1988
This technical report is a slightly
abbreviated version of the author’s dissertation presented to the Department of
Computer and Information Science and the
© 1988 David Graham Novick
ABSTRACT
CONTROL OF MIXED-INITIATIVE DISCOURSE THROUGH
META-LOCUTIONARY ACTS: A
COMPUTATIONAL MODEL
Human-computer interaction typically
displays single-initiative interaction in which either the computer or the
human controls the conversation. The
interaction is largely preplanned and depends on well-formed language. In contrast, human-human conversations are
characterized by unpredictability, ungrammatical utterances, non-verbal
expression, and mixed-initiative control in which the conversants take
independent actions. Traditional
natural-language systems are largely unable to handle these aspects of
"feral" language. Yet
human-human interaction is coherent for the participants; the conversants take
turns, make interruptions, detect and cure misunderstandings, and resolve
ambiguous references. How can these
processes of control be modeled formally in a manner sufficient for use in
computers?
Non-sentential aspects of conversation
such as nods, fragmentary utterances, and correction can be seen reflecting
control information for interaction.
Such actions by the conversants, based on the context of their
interaction, determine the form of the conversation. In this view ungrammaticality, for example,
is not a problem but a guide to these "meta" acts. This dissertation develops a theory of
"meta-locutionary" acts that explains these control processes. The theory extends speech-act theory to
real-world conversational control and encompasses a taxonomy of
meta-locutionary acts.
The theory of meta-locutionary acts was
refined and validated by a protocol study and computational simulation. In the protocol study, subjects were given a
cošperative problem-solving task. The
conversants' interaction, both verbal and non-verbal, was transcribed as
illocutionary and meta-locutionary acts.
The computational model was developed using a rule-based system written
in Prolog. The system represents the
independent conversational knowledge of both conversants simultaneously, and
can simulate their simultaneous action.
Simulations of the protocol conversations using the computational model
showed that meta-locutionary acts are capable of providing control of
mixed-initiative discourse. The model
agents can, for example, take and give turns.
A single agent can simultaneously take multiple acts of differing
control. The simulations also confirmed
that conversations need not be strictly planned. Rather, mixed-initiative interaction can be
plausibly controlled by contextually determined operators.
This research has application to natural
language processing, user interface design and multiple-agent artificial
intelligence systems. The theory of
meta-locutionary acts will integrate well with existing speech-act-based
natural-language systems.
ACKNOWLEDGMENTS
I wish to express appreciation to
Professor Sarah A. Douglas for her guidance, inspiration, faith and
perserverance. I thank Professors Scott
C. DeLancey, Stephen F. Fickas, and Kent A. Stevens for their constructive,
diligent work as members of the dissertation committee. Professor Russell S. Tomlin was very helpful
with respect to discourse issues.
Barbara J. Grosz graciously reviewed my dissertation proposal and
provided many detailed comments; I am grateful for her time and assistance. I received helpful comments from the mentors
and members of the ACM SIGCHI CHI'88 Doctoral Consortium. I also wish to thank audiences at Xerox PARC,
Information Sciences Institute, and the
TABLE OF CONTENTS
Chapter
I. INTRODUCTION 1
Speech-Act
Theory 3 Mutuality of Knowledge 5
II. SURVEY OF RELATED WORK 9
The
Integration Problem 9
Characteristics
of Conversation 10
Evidence
for Shared Models of Conversations 11
Repair-Based
Conversational Interaction 15
Intention,
Action, and Language 21
Summary 25
III. A THEORY
OF META-LOCUTIONARY ACTS 27
Approaches
to the Integration Problem 27
A
Computational Model of Meta-Locutionary Acts 29
Summary 35
IV. METHODOLOGY 37
Methodology
in Cognitive Science 37
Effects
to be Studied 38
Protocol
Analysis of Conversational Interaction 44
Summary 44
V. RESULTS OF THE
PROTOCOL STUDY 47
Modeling
the Conversation 47
Application
of the Model 52
VI. THE SIMULATION STUDY 61
Representation 62
The
Rule-Based System 65
Implementation
of the Model as Rules 68
Results
of the Simulation 71
Limitations
and Extensions 74
VI. CONCLUSION 77
Summary 77
Open
Issues 78
Humans
and Computers 82
APPENDIX
A. Transcript of Protocol 85
B. PREDICATE REPRESENTATIONS 93
C. CONVERSATIONAL
MODELS 95
D. CONVERSATIONAL OPERATORS 99
E. SHORT
TRACE OF SIMULATED CONVERSATION 103
F. FULL
TRACE OF SIMULATED CONVERSATION 107
G. RULE-BASED SYSTEM: PROGRAM 131
H. RULE-BASED
SYSTEM: META-LOCUTIONARY OPERATORS 145
I. RULE-BASED
SYSTEM: DOMAIN OPERATORS 153
J. RULE-BASED
SYSTEM: DOMAIN-SPECIFIC CLAUSES 163
K. RULE-BASED
SYSTEM: INITIAL STATE 165
BIBLIOGRAPHY 167
LIST OF FIGURES
Figure
1. Differences Between Interaction Modes 1
2. Possible Levels of Conversational Interaction 17
3. Taxonomy of Meta-Locutionary Acts for
Turn-Taking 34
4. Taxonomy of Meta-Locutionary Acts for Repair
of Mutual Models 35
5. Taxonomy of Meta-Locutionary Acts for
Information 35
6.
Taxonomy of Meta-Locutionary Acts for Attention 35
7. Example of Random-Letter Sequences Used in
Protocols 42
8. Example of Non-Identical Sequences Used in
Third Task 43
9. Overhead View of the Experimental Set-Up for
the Letter-Sequence Protocols 43
10. Illocutionary Interpretation of the Protocol 49
11. Operator Repeat-Act-1 51
12. Application of Operator Repeat-Act-1 After
Act Bi6 51
13. Partial Transcript of Experimental Protocol 53
14. Form of Rule Representing Meta-Locutionary
Operator 64
15.
16. Meta‑Locutionary Rule
Assertion_Received_1 69
17. Domain-Rule Confirmed_Next_Letter 70
18. Trace of Simulation of Protocol 72