Lecture Notes in
Computer Science
Edited by G. Goos and .I. Hartmanis
119
II
I
Graeme Hirst
Anaphora in Natural
Language Understanding:
A Survey
Springer-Verlag
Berlin Heidelberg NewYork 1981
Editorial Board
W. Brauer P, Brinch Hansen D. Gries C. Moler G. Seegm(Jller
J. Stoer N. Wirth
Author
Graeme Hirst
Department of Computer Science, Brown University
Providence, RI 02912, USA
A M S Subject Classifications (19 70): 68 F 20, 68 F 30, 68 G 99
CR Subject Classifications (1974): 3.42, 3.60
ISBN 3-540-10858-0 Springer-Verlag Berlin Heidelberg New York
lSBN 0-387-10858-0 Springer-Verlag New York Heidelberg Berlin
This work is subject to copyright. All rights are reserved, whether the whole or part
of the material is concerned, specifically those of translation, reprinting, re-use of
illustrations, broadcasting, reproduction by photocopying machine or similar means,
and storage in data banks. Under £354 of the German Copyright Law where copies
are made for other than private use, a fee is payable to "Verwertungsgesellschaft
Wort", Munich.
© by Springer-Verlag Berlin Heidelberg 1981
Printed in Germany
Printing and binding: Beltz Offsetdruck, Hemsbach/Bergstr.
2145/3140-543210
CONTENTS
PREFACE .......................................................................................................
How to read this report
ix
Notation x
PREFACE
TO THE
ACKNOWLEDGEMENTS
SPRINGER
EDITION
ix
...........................................................
xi
.................................................................................
xiii
INTRODUCTION
...............................................................................................
1
i.i, Natural language understanding
1
1.2. Reference and anaphora
2
2.
ANAPHORA
.....................................................................................................
4
2.1. What is anaphora?
4
2.2. Anaphors
as references to entities in consciousness
7
2.3. Varieties of anaphora
I0
2.3.1. Pronominal
reference
i0
2,3.2. Pronominal
noun phrases: Surface count anaphors
12
2.3.3. Pronominal
noun phrases: Epithets 13
2.3.4. Prosentential reference
13
2.3.5. Strained anaphora
14
2.3.6. Difficult indefinite uses of o~e 16
2.3.7. Non-referential pronouns
17
2.3.8. Pro-verbs
19
2.3.9. Proactions
19
2, 3. i0. Proadjectives 20
2.3.1 i. Temporal
references 21
2.3.12. Locative references 21
2.3.13. Ellipsis: The ultimate anaphor!
22
2.3.14, An awkward
miscellany
22
2.3.15. Summary
of anaphors
24
2.4. Where does anaphora
end? 34
2.4.1. Paraphrase
24
2.4.2. Definite reference
26
2.5. Types of reference
28
2,6. Ambiguity
in anaphora
and default antecedents
29
2.7. Summary
and discussion 32
CONTENTS iii
IV
3, TRADITIONAL
APPROACHES
TO ANAFHORA
...................................................
33
3.1. Some traditional systems
33
3, I,i. STUDENT
34
3.1,2. SHRDLU
35
3,1.3. LSNLIS
36
3, 1.4. MARGIE
and SAM 38
3.1.5. A case-driven parser 38
3,1,6. Parse tree searching
39
3.1.7. Preference semantics
40
3,1.8. Summary
40
3.2. Abstraction of traditional approaches
41
3.2. I. A formalization of the problem
42
3.2,2. Syntax methods 43
3.2.3. The heuristic approach 43
3.2.4. The case g r a m m a r approach 45
3.2,5. Analysis by synthesis 47
3,2,6. Resolving anaphors by inference 48
3.2,7. S u m m a r y arld discussion 49
THEME
IN ANAPHORA
RESOLUTION
4. THE NEED FOR DISCOURSE
51
4.1. Discourse theme
4. i,i. The linguistic approach
52
4, 1.2. The psycholinguistic approach
53
4.1,3. Lacunae abounding
54
4.2. Why focus and theme are needed in anaphor resolution
59
4.3. Can focusing be tamed?
...................50
54
6. DISCOURSE-ORIENTED
ANAPHORA
SYSTEMS
AND THEORIES
.......................60
5,1. Concept activatedness
60
5, I, I. Kantor's thesis 60
5.1.2. The implications of Kantor's work 62
5.2. Focus of attention in task-oriented dialogues 63
5,2,1. Motivation 63
5.2.2. Representing
and searching focus 64
5.2.3. Maintaining focus 65
5.3. Focus in the PAL system and Sidner's theory 67
5.3.1. PAL's approach to discourse 67
5.3.2. The frame as focus 68
5.3,3. Focus selection 69
5,3,4. Sidner's general theory 71
5.3.5. Conclusions
72
5,4. Webber's formalism
73
5.4.1. Definite pronouns
73
5.4.2. O~e-anaphors
75
5,4,3. Verb phrase ellipsis 78
5.4,4. Conclusions
79
5.5, Discourse-cohesion
approaches
to anaphora
resolution 80
5.5,1. Coherence
relations 61
5.5.2. An e x a m p l e of a n a p h o r r e s o l u t i o n u s i n g a c o h e r e n c e r e l a t i o n 83
5.5.3. L o e k m a n ' s c o n t e x t u a l r e f e r e n c e r e s o l u t i o n a l g o r i t h m 84
5.5.4. C o n c l u s i o n 86
5.6. N o n - n o u n - p h r a s e f o c u s i n g 86
5.6. i. The f o c u s of t e m p o r a l a n a p h o r s 87
5.6.2. The f o c u s of l o c a t i v e a n a p h o r s 90
6. CONSTRAINTS
AND DEFAULTS
IN ANAPHOR
RESOLUTION
6.1. Gender and number 92
6.2. Syntactic constraints 92
6,3. Inference and world knowledge
93
6,4. Parallelism
93
6.5. The preferred antecedent
and plausibility 93
6.6. Implicit verb causality 95
6.7. Semantic
distance 96
.............................92
......................................................................................
97
7. THE LAST CHAPTER
in spoken language
97
7,1. Anaphora
in computer
language generation
98
7.2. Anaphora
7.2.1. Introduction
98
7.2.2. Structural constraints on subsequent
reference
99
7.2.3. Conclusion
99
judgements
i00
7.3. Well-formedness
7.4. Research problems 10l
7.5. Conclusion 105
REFERENCES
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
107
INDEX OF NAMES ........................................................................................ 123
SUBJECT
INDEX
..........................................................................................
126
In. loving m e m o r y of m y F a t h e r
PREFACE
I was a victim
-
of a series of accidents.
Kurt Vonnegut Jr 1
This r e p o r t was s t a r t e d in t h e b o r e a l s u m m e r of 1976, m a k i n g i t s f i r s t a p p e a r a n c e as H i r s t (1976b), a n d was c o m p l e t e d a l m o s t t h r e e y e a r s l a t e r , a f t e r a
n u m b e r of l a p s e s a n d r e l a p s e s . Like a c h i n c h i l l a o n e is t r y i n g t o p h o t o g r a p h ,
t h e field I was t r y i n g t o d e s c r i b e w o u l d n o t s i t still. T h e r e f o r e , w h i l e I h a v e
t r i e d t o i n c o r p o r a t e all t h e c h a n g e s t h a t o c c u r r e d in t h o s e y e a r s , t h e r e m a y b e
some blurring at the edges.
I have tried to m a k e this survey c o m p r e h e n s i b l e both to the c o m p u t e r
s c i e n t i s t who h a s no g r o u n d i n g in l i n g u i s t i c s , a n d to t h e l i n g u i s t who k n o w s
n o t h i n g of c o m p u t e r s . H o w e v e r , it h a s b e e n n e c e s s a r y to p r e s u m e s o m e i n f o r m a t i o n , s i n c e d i g r e s s i o n s to e x p l a i n t r a n s f o r m a t i o n a l g r a m m a r o r F i l l m o r e ' s
case theory, for example, were clearly impractical. (Readers not familiar with
these m a y wish to read an introductory text on transformational g r a m m a r s
such as Jacobsen (1977), Akmajian and H e n y (1975) or Grinder and Elgin (1973),
and F-illmore's (1968) introduction to cases. The reader not familiar with
artificial intelligence will find Winston (1977), B o d e n (1977) or B u n d y (1979) useful i n t r o d u c t i o n s . )
I t i s to be n o t e d , t h a t w h e n a n y
a p p e a r s dulL, t h e r e i s a d e s i g n i n i t .
-
part
of
this paper
Richard Steele 2
How t o r e a d t h i s r e p o r t
This is a long r e p o r t , b u t few p e o p l e will n e e d t o r e a d it all. The C h a p t e r o u t lines below will help you find the sections of greatest interest to you.
Chapter 1 introduces and motivates work on natural language understanding and in particular anaphora. If you are already motivated, skip to Chapter 2.
1From: The s/zens ojr P~tan. London: Coronet, 1967, page 161.
2In: The taller, number 38, Thursday 7 July 1709. Reprinted in: The
2rat/ons. Edinburgh: Robert Martin, I845, volume 1, page 236.
taller,
qzri2hnotes and illus-
PREFACE ix
C h a p t e r 3 d e f i n e s a n a p h o r a f o r m a l l y , a n d m o t i v a t e s t h e i d e a of " c o n s c i o u s n e s s " as a r e p o s i t o r y for a n t e c e d e n t s . S e c t i o n 2.3 is a n e x p o s i t i o n of t h e v a r i ous t y p e s of a n a p h o r a . I s u g g e s t t h a t r e a d e r s f a m i l i a r with a n a p h o r a n e v e r t h e less a t l e a s t s k i m t h i s s e c t i o n , as I have i n c l u d e d a n u m b e r of u n u s u a l e x a m p l e s
and c o u n t e r e x a m p l e s which are often ignored but which should be c o n s i d e r e d
b y a n y o n e c l a i m i n g t o have a c o m p l e t e a n a p h o r - h a n d l i n g s y s t e m or t h e o r y .
C h a p t e r 3 r e v i e w s t r a d i t i o n a l a p p r o a c h e s to a n a p h o r a r e s o l u t i o n , a n d shows
why t h e y a r e i n a d e q u a t e . S e c t i o n 3.1 d i s c u s s e s t h e w o r k of Bobrow, Winograd,
Woods a n d his a s s o c i a t e s , S c h a n k a n d his s t u d e n t s , Taylor, H o b b s a n d Wilks.
T h e n in s e c t i o n 3.2 I a b s t r a c t a n d e v a l u a t e t h e a p p r o a c h e s t h e s e p e o p l e took.
In C h a p t e r 4, I show t h e i m p o r t a n c e of d i s c o u r s e t h e m e a n d a n a p h o r i c
f o c u s in r e f e r e n c e r e s o l u t i o n .
]n C h a p t e r 5 I review five c u r r e n t d i s c o u r s e - o r i e n t e d a p p r o a c h e s to a n a p h o r a - t h o s e of K a n t o r , Grosz, S i d n e r , Webber, a n d t h e d i s c o u r s e c o h e s i o n
a p p r o a c h of L o c k m a n a n d o t h e r s . A p p r o a c h e s to n o n - N P a n a p h o r a a r e also o u t lined here.
C h a p t e r 6 d e s c r i b e s t h e role of a n a p h o r - s p e c i f i c i n f o r m a t i o n i n r e s o l u t i o n ,
a n d i n t e g r a t e s t h e o r i e s of c a u s a l v a l e n c e i n t o a m o r e g e n e r a l f r a m e w o r k .
C h a p t e r 7 d i s c u s s e s s o m e i s s u e s r a i s e d i n e a r l i e r c h a p t e r s , s u c h as p s y c h o l i n g u i s t i c t e s t i n g , a n d also t h e p r o b l e m s of a n a p h o r a i n l a n g u a g e g e n e r a t i o n .
The r e p o r t c o n c l u d e s with a review of o u t s t a n d i n g p r o b l e m s .
Copious b i b l i o g r a p h i c r e f e r e n c e s will k e e p y o u b u s y i n t h e l i b r a r y for h o u r s ,
a n d a n i n d e x of n a m e s will h e l p y o u find o u t w h e r e i n t h i s work y o u r f a v o r i t e
work is d i s c u s s e d . A s u b j e c t i n d e x is also p r o v i d e d .
Notation
In t h e s a m p l e t e x t s i n t h i s r e p o r t , I u s e u n d e r l i n i n g to i n d i c a t e t h e a n a p h o r ( s )
of i n t e r e s t , t h e s y m b o l ¢ t o e x p l i c i t l y m a r k t h e p l a c e w h e r e a n ellipsis
occurred, and small capitals to indicate words t h a t are s t r e s s e d when the sent e n c e is s p o k e n . S u p e r s c r i p t n u m b e r s i n p a r e n t h e s e s a r e s o m e t i m e s u s e d to
e x p l i c i t l y l a b e l d i f f e r e n t o c c u r r e n c e s of t h e s a m e word i n a text. V a r i a n t r e a d ings of a t e x t a r e e n c l o s e d i n b r a c e s , with t h e v a r i a t i o n s s e p a r a t e d b y a v e r t i c a l
bar. A s e n t e n c e w h i c h is g r a m m a t i c a l b u t u n a c c e p t a b l e i n t h e g i v e n c o n t e x t is
d e n o t e d b y " # " . As u s u a l , "*" a n d " ? " d e n o t e t e x t w h i c h is i l l - f o r m e d a n d of
questionable welI-formedness, respectively.
In t h e m a i n b o d y of t h e r e p o r t , I u s e s m a l l c a p i t a l s for e m p h a s i s or to i n d i c a t e t h a t a new t e r m is b e i n g defined. I t a l i c s h a v e t h e i r u s u a l m e t a l i n g u i s t i c
role i n r e f e r r i n g t o w o r d s a n d p h r a s e s . N P a n d VP s t a n d for n o u n p h r a s e a n d
verb phrase.
By L I m e a n m y s e l f , G r a e m e Hirst, t h e w r i t e r of t h i s d o c u m e n t , a n d b y we, I
m e a n you, t h e r e a d e r , a n d m e t o g e t h e r . So, for e x a m p l e , w h e n I say I
think . . . . I am expressing a personal opinion; whereas when I say we see .... t
am pointing out s o m e t h i n g a b o u t which the r e a d e r and I u n d o u b t e d l y agree a n d we d o n ' t , t h e f a u l t is p r o b a b l y i n t h e r e a d e r .
Notation
PREFACE TO THE SPRINGER EDITION
I o r i g i n a l l y w r o t e t h i s r e p o r t as a t h e s i s for t h e M a s t e r of S c i e n c e d e g r e e i n t h e
D e p a r t m e n t of E n g i n e e r i n g P h y s i c s , A u s t r a l i a n N a t i o n a l U n i v e r s i t y . The t h e s i s
was also p u b l i s h e d as t e c h n i c a l r e p o r t 79-2 (May 1979) by t h e D e p a r t m e n t of
C o m p u t e r S c i e n c e , U n i v e r s i t y of B r i t i s h C o l u m b i a ,
In t h e y e a r or so p r e c e d i n g t h e first p u b l i c a t i o n of t h e r e p o r t , t h e s t u d y of
a n a p h o r a i n n a t u r a l l a n g u a g e u n d e r s t a n d i n g was a v e r y c u r r e n t topic, with t h e
p u b l i c a t i o n of s e v e r a l i m p o r t a n t d o c t o r a l t h e s e s (which are r e v i e w e d i n C h a p t e r
5). t h a d o r i g i n a l l y b e l i e v e d t h a t t h e field was c h a n g i n g so f a s t t h a t t h e s u r v e y
would be s u b s t a n t i a l l y o u t of d a t e w i t h i n a y e a r . This h a s n o t p r o v e d to be t h e
case; r a t h e r , work i n t h e a r e a h a s slowed, as r e s e a r c h e r s p a u s e to e v a l u a t e a n d
r e c o n s i d e r t h e a p p r o a c h e s t a k e n . I now t h i n k t h a t t h i s r e p o r t h a s a l o n g e r than-anticipated life expectancy, and that it will continue to be helpful to those
constructing natural language understanding systems.
The present edition was typeset with a text-formatting system that is unfortunately typical of m a n y found in computer science departments, in that it was
designed by people w h o k n o w a lot about computers but not very m u c h about
typography or book design. I hope therefore that you will forgive the occasional footnote that runs onto a n e w page w h e n it shouldn't have, s o m e ludicrous hyphenation, the funny shape of certain letters, and the a w k w a r d widows
that turn up in a few places. The page n u m b e r s in the indexes should be
regarded as approximate only, especially where the reference is in a footnote
carried over to the next page.
Provi, de~ce, I May 1981
PREFACE TO THE SPRINGER EDITION xi
ACKNOWLEDGEMENTS
Who m a d e m e t h e g e n i u s I a m t o d a y -The m a t h e m a t i c i a n
that others all quote?
Who "s t h e p r o f e s s o r t h a t m a d e m e t h a t w a y ?
The g r e a t e s t t h a t e v e r g o t c h a l k o n h i s c o a t !
One m a n d e s e r v e s t h e c r e d i t , o n e m a n d e s e r v e s t h e
blame,
And Nicolai Ivanovich Lobachevsky is his name.
-
Tom Lehrer 1
I wrote this survey while a graduate student at the Department of C o m p u t e r
Science, University of British Columbia; m y supervisor was Richard Rosenberg.
P a r t s w e r e also w r i t t e n a t t h e D e p a r t m e n t of E n g i n e e r i n g P h y s i c s , R e s e a r c h
S c h o o l of P h y s i c a l S c i e n c e s , A u s t r a l i a n N a t i o n a l U n i v e r s i t y (ANU), u n d e r t h e
j o i n t s u p e r v i s i o n of S t e p h e n Kaneff a n d l a i n M a c l e o d .
Without the encouragement
been started.
of R o b i n S t a n t o n , t h i s w o r k would n e v e r h a v e
D i s c u s s i o n s w i t h a n d / o r c o m m e n t s a n d c r i t i c i s m f r o m t h e following p e o p l e
i m p r o v e d t h e q u a l i t y of t h i s r e p o r t : R o g e r Browse, W a l l a c e Chafe, J i m D a v i d s o n ,
B a r b a r a Grosz, M A K H a l l i d a y , Alan M a e k w o r t h , B o n n i e N a s h - W e b b e r (in 1977),
Doug T e e p l e a n d B o n n i e W e b b e r (in 1978). N e v e r t h e l e s s , t h e m i s t a k e s a r e m y
r e s p o n s i b i l i t y ( e x c e p t in a n y i n s t a n c e w h e r e , in p r e s e n t i n g t h e w o r k of a n o t h e r ,
I have been misled through the original author's inability to communicate
c o h e r e n t l y , in w h i c h c a s e t h e o r i g i n a l a u t h o r m u s t t a k e t h e b l a m e ) .
F i n a n c i a l s u p p o r t c a m e f r o m t h e A u s t r a l i a n D e p a r t m e n t of E d u c a t i o n a s a
C o m m o n w e a l t h P o s t g r a d u a t e R e s e a r c h (CPR) Award, f r o m ANU a s a CPR A w a r d
s u p p l e m e n t , a n d a s a s p e c i a l g r a n t f r o m t h e ANU D e p a r t m e n t of E n g i n e e r i n g
Physics.
Many s p e l l i n g e r r o r s a n d s t y l i s t i c g r o s s o s i t i e s h a v e b e e n e l i m i n a t e d t h r o u g h
t h e d e t e c t i v e w o r k of M a r k S c o t t J o h n s o n , N a d i a T a l e n t , l a i n M a c l e o d a n d
M W Peacock.
F i n a l l y , I w a n t t o e s p e c i a l l y a c k n o w l e d g e t h e h e l p of two p e o p l e : R i c h a r d
R o s e n b e r g , who f i r s t i n t e r e s t e d m e in t h i s w o r k a n d e n c o u r a g e d m e t o c o m p l e t e it, a n d N a d i a T a l e n t , w i t h o u t w h o m i t w o u l d all h a v e t u r n e d o u t q u i t e
differently.
tFrom: Nieolai Ivanovieh Lobaehevsky. On: Lehrer, Tom. Songs by Tor~ L~hTe~. LP recording,
Reprise RS6~16,
ACKNO~3~EDGEMENTS xiii