AN INDIRECT APPROACH TO TYPES OP SPEECH ACTS.
Jeremy Ellman
Department of Psychology
U n i v e r s i t y of Warwick
Coventry, U.K.*
Abstract
Consideration
o f the s i m i l a r i t y between d i r e c t
and i n d i r e c t speech act u n d e r s t a n d i n g g i v e r i s e
to the notion that taxonomies of speech acts
may not be h e l p f u l in modelling language understanding. A computer model which t r e a t s represent a t i o n s o f d i r e c t and i n d i r e c t speech a c t s
s i m i l a r l y and s u c c e s f u l l y has
been implemented
without any such taxonomy and without an e x p l i c i t
r e p r e s e n t a t i o n of the d i f f e r e n c e between d i r e c t
and i n d i r e c t speech acts.
Introduction.
Speech acts
theory is of great i n t e r e s t
t o many w o r k e r s f r o m v a r i o u s d i s c i p l i n e s
i n t e r e s t e d in n a t u r a l language.
One of the
problems that t h e i r work has h i g h l i g h t e d is that
of the i n d i r e c t speech act: An utterance
which
the speaker
makes c a r r y an i l l o c u t l o n a r y
force
more
important than,
and in a d d i t i o n
t o , the i l l o c u t l o n a r y f o r c e d e r i v a b l e from the
l i t e r a l meaning of that utterance (Searle 1975).
Another problem which has r e c e i v e d c o n s i d e r a b l e a t t e n t i o n (eg A u s t i n 1962,
Searle
1976,
Dore
1977,
Katz 1977, V e n d l e r 1972
Hancher 1979, S t i l e s 1981 and many o t h e r s ) is
the
classification
of
speech
acts i n t o
types. Here there are two questions:
1) What types of speech acts are there?
2 ) What s h o u l d t h e i r c r i t e r i a
for
c l a s s i f i c a t i o n be?
I am g o i n g to suggest
in the f o l l o w i n g t h a t
these q u e s t i o n s may be avoided as the n o t i o n of
speech a c t types i s not e s s e n t i a l f o r n a t u r a l
language understanding (NLU). F u r t h e r m o r e I am
going to reach t h i s p o s i t i o n by considering the
understanding of a few d i r e c t and i n d i r e c t speech
a c t s and by a r g u i n g t h a t d i r e c t and i n d i r e c t
s p e e c h a c t s may be u n d e r s t o o d by a s i n g l e
mechanism.
For b r e v i t y
the
discussion
is
r e s t r i c t e d to REQUEST and INFORM type speech
acts.
An Example.
Consider such utterances as:1) I t s cold in here.
2) Turn on the h e a t e r .
3) Shut the window.
Speech a c t
t h e o r i s t s (eg A u s t i n 1962,
Searle 1969) t e l l us that (1) is an INFORM type
( d i r e c t ) speech a c t ,
w h i l s t (2) and (3) are
DIRECTIVE
or REQUEST type ( d i r e c t ) speech acts.
However i t I s easy t o v i s u a l i z e a s i t u a t i o n i n
which a speaker (Mary,
say) u t t e r s (1) and
her hearer (John, say) i n f e r s t h a t she a c t u a l l y
meant to convey s o m e t h i n g in a d d i t i o n to t h a t
(ie (2),
or (3),
o r (2) and ( 3 ) ) .
Utteri n g (1) w o u l d t h e n be termed
making
an
"indirect"
REQUEST.
Let
us
call
this
s i t u a t i o n case A.
S i m i l a r l y , one could u t t e r (2) or (3) and
simultaneously convey the INFORM meaning of (1).
However it is not usual to c a l l e i t h e r (2) or
(3) " i n d i r e c t INFORMs" as the INFORM is not t h e i r
primary f u n c t i o n . Let us c a l l t h i s case B.
A d d i t i o n a l l y , if Mary enters a cool room and
says (3) to John, as she b e l i e v e s the window to
be open, it would seem very s t r a n g e to c a l l
(3) an i n d i r e c t way of u t t e r i n g the REQUEST
(2) If John were to turn on the heater (as he
knows that the window is already closed and that
there is a heater which is turned o f f ) .
(case
C).
If it
is
accepted
that
case A above
is a " c l a s s i c " i n d i r e c t speech a c t , how do we
d e s c r i b e cases B and C? Case B cannot be d e s cribed as a t y p i c a l i n d i r e c t speech act as it is
l e s s i m p o r t a n t t o Mary t h a t John i n f e r (1) than
that he
confirm with
her
literal
request
( 3 ) . (That I s , i f one accepts t h a t the r e q u e s t i s
the more primary meaning in case A.) Possibly
one could avoid the problem by describing case B
as an I n s t a n c e of a "background
entailment"
( S m i t h and W i l s o n 1979).
and c o n s e q u e n t l y
claim that
a d i f f e r e n t mechanism is used
t o u n d e r s t a n d (1) f r o m the u t t e r a n c e o f (2) i n
case B than the mechanism
used to u n d e r s t a n d (2) f r o m the u t t e r a n c e of (1) In case A.
However,
t h i s p o s s i b i l i t y does not help in
the analysis of case C.
J. Ellman 601
In case
C
John
responds cooperatively
(Grice 1975, Perrault Allen and Cohen 1978,) to
a request t h a t Mary
did
not i n t e n t i o n a l l y
make.
Note t h a t t h i s case does not seem to be
accomodated by the work of A l l e n and P e r r a u l t
who e x p l i c i t l y f o l l o w speech act theory in assumi n g t h a t the speaker makes h i s u t t e r a n c e s w i t h
the i n t e n t i o n t h a t the hearer
recognizes h i s
i n t e n t i o n s ( A l l e n 1979 p l 0 7 , A l l e n and P e r r a u l t
1980 pl70). However, to accomodate t h i s case are
we to say t h a t (2) is understood from (3) by a
d i f f e r e n t mechanism a g a i n than t h a t used
in cases A or B? Or are we to look instead for a
single mechanism which can deal w i t h a l l three
example cases and for which " I n d i r e c t speech
acts" are merely a special case?
A s i n g l e mechanism is c e r t a i n l y the most
elegant s o l u t i o n , and i t s presence is suggested
by an e x a m i n a t i o n of Mary's p r i m a r y
aim
(or
superordinate
goal,)
in making e i t h e r
u t t e r a n c e in cases A, B or C: She wishes to be
warm.
Recognizing Mary's
a c t i o n helped Mary to
even when he t u r n e d on
obey her l i t e r a l request
( i e case C).
aim
t e l l s us John's
a c c o m p l i s h her g o a l
the h e a t e r and d i d not
to shut
the
window
From the above it would seem t h a t John (or
any h e a r e r , ) is capable of i n f e r r i n g at l e a s t
some of Mary (or any speaker's,) superordinate
goal(s) in making an utterance. However the
hearer of an utterance can not t e l l in advance
which of the speaker's utterances w i l l I n d i c a t e
higher u n s a t i s f i e d goals ( l e be " i n d i r e c t speech
acts").
Therefore
it seems that the hearer
must
continually
examine
each of
the
speaker's u t t e r a n c e s f o r such a l l u s i o n s .
This
must be so even f o r so c a l l e d d i r e c t REQUESTS, as
In case C. Consequently the hearer must process
d i r e c t and i n d i r e c t speech acts
identically
and use the same inference procedures in order to
understand both f u l l y .
What t h e n i s t h e f u n c t i o n o f S e a r l e ' s
REQUEST speech act type? As h e a r e r s may choose
not to acceed to REQUESTS, or decide to cooperate
in
other
ways (case C), it would seem t h a t
REQUESTS only suggest, or INFORM, the hearer
as to the speaker's d e s i r e s .
Perhaps then
REQUESTS may be l i t t l e more than INFORMS in which
the speaker expresses p r e c i s e l y the course of
a c t i o n she wishes the hearer to take. That i s ,
f r o m the p o i n t of view of the u n d e r s t a n d e r of
speech acts it may not be relevant that u t t e r ances (2) and (3) are REQUESTS, as long as the
hearer can d e t e c t
from
them the speaker's
superordinate
goal(s).
It therefore
seems
that the notion of d i f f e r e n t speech act types is
not necessary f o r NLU**
A d d i t i o n a l l y , i t I s c l e a r from the l i t e r ature that the c l a s s i f i c a t i o n of speech acts i n t o
types is p r i m a r i l y f o r a n a l y t i c purposes (Austin
1962, S t i l e s 1981 e t c ) and i t i s not s t a t e d
anywhere that that t h i s c l a s s i f i c a t i o n is essen-
t i a l to the understanding process. This may seem
strange considering the a t t e n t i o n that typologies
of speech a c t s have r e c e i v e d but I b e l i e v e t h i s
to be more a r e f l e c t i o n of the d i f f e r i n g aims of
the workers from the various d i s c i p l i n e s Involved
in the area ( i e Philosophy, L i n g u i s t i c s , Psychology etc) and those of Computational L i n g u i s t i c s .
Two suggestions have been made so far then,
1) t h a t the n o t i o n of a speech act type is
not essential
f o r NLU.
2) that d i r e c t and I n d i r e c t speech acts may
be understood by the same mechanism.
A Computer m o d e l .
A computer model has been programmed using
the above ideas which is capable of understanding
some I n d i r e c t speech acts. The model copes w i t h
representations of cases A, B and C in i t s form
of E n g l i s h . This i n v o l v e s o n l y two types of
mental predicates, ( f o l l o w i n g Perrault A l l e n and
Cohen 1978, Cohen 1978, A l l e n 1979: eg SPEAKER
INTENDS, HEARER KNOWS e t c . ) and u n i n t e r p r e t e d
propositions (eg: that A HAS THE SALT). A d d i t i o n a l l y , the model uses a n o t i o n of mutual b e l i e f
(MB) (see S h i f f e r 1972, Smith 1982). M u t u a l bel i e f is necessary for the f o l l o w i n g reason: If a
speaker makes some u t t e r a n c e s u c c e s s f u l l y , he
can reason t h a t the hearer now knows the c o n t e n t s o f t h a t u t t e r a n c e . S i m i l a r l y the speaker
knows t h a t the hearer knows e t c . W h i l s t such
processing may be l i m i t e d in practice (Clark and
Marshall 1981), it is necessary in t h i s impleme n t a t i o n of the theory f o r knowledge states to
b e d e r i v a b l e t o any d e p t h f r o m a m u t u a l l y
believed item.
The two p a r t i c i p a n t s in the c o n v e r s a t i o n
are modelled as concurrently executing production
systems ( i e , the model is more s i m i l a r
to that
of Power(1979) than t h a t of A l l e n ( 1 9 7 9 ) . * * * ) .
They have independent agendas and b e l i e f s and
communicate by one special
r u l e which the p r o d u c t i o n i n t e r p r e t e r r e c o g n i z e s as an INFORM
speech a c t . This is the o n l y type of Speech Act
that the system knows about.
The r e s u l t of
t h i s act is to make the contents of the u t t e r ance m u t u a l l y b e l i e v e d . That i s , not o n l y does
the hearer know the c o n t e n t s of the u t t e r a n c e ,
but the speaker knows the hearer knows etc. This
d i f f e r s from A l l e n and P e r r a u l t ' s (1980) f o r m u l a t i o n of INFORM which only r e s u l t s in the hearer
knowing the utterance's contents.
The model operates as f o l l o w s : The p a r t i c i pants are g i v e n i n i t i a l b e l i e f s t a t e s .
One o f
them is then g i v e n an i n t e n t i o n to achieve some
goal.
If t h i s goal or i n t e n t i o n leads to the
subgoal of making an utterance, the second part i c i p a n t t r i e s t o i n f e r the f i r s t p a r t i c i p a n t ' s
superordinate
goal(s)
in
making j u s t
that
utterance. This i n f e r e n c e process e x p l o i t s the
f a c t t h a t the p r o d u c t i o n r u l e s used i n making
p l a n s are a p p l i c a b l e i n r e v e r s e , ( i e g i v e n
some r e s u l t s t a t e the p a r t i c i p a n t s may i n f e r the
goal s t a t e that gave r i s e to
that
result.)
602
J. Ellman
This method
contrasts
Allen(l979)
which
heuristics.***
with
t h a t used by
involves
several
Having i n f e r r e d such a superordinate i n t e n t i o n for the speaker, the hearer w i l l then f i n d
a r u l e which embodies Grice's(1975) Cooperative
p r i n c i p l e . H e w i l l apply t h i s r u l e and i n doing
so w i l l ( i n P e r r a u l t A l l e n and Cohen's (1978)
t e r m s , ) adopt the speaker's goal as h i s own.
T h i s means t h a t Mary's u t t e r i n g ( 1 ) , ( 2 ) , or (3)
w i l l lead John to i n f e r t h a t Mary wishes to be
warm. Therefore John wishes her to be warm. Consequently he w i l l shut the door, close the w i n dow, or do whatever e l s e is a p p r o p r i a t e .
His
a c t i o n s and u t t e r a n c e s are d e t e r m i n e d by h i s
knowledge of the s i t u a t i o n of which Mary's u t t e r ance (and h i s i n f e r e n c e s from i t , ) form j u s t a
part.
I t i s not i m p o r t a n t t o him
that
Mary
uttered a REQUEST or an INFORM speech act:
It
o n l y m a t t e r s t h a t h e i s able t o d e t e c t her
most p r i m a r y i n t e n t i o n (or g o a l ) i n making
that
speech act.
The i m p l i c a t i o n f o r NLU from t h i s work is
that not only is the notion of speech act
types
redundant, but so is the d i f f e r e n c e
between
" d i r e c t " and " I n d i r e c t " speech
acts. Indeed,
what is in c l a s s i c a l terras a d i r e c t REQUEST used
to convey i t s l i t e r a l meaning may lead the hearer
to disregard the speaker's s u p e r f i c i a l i n t e n t i o n
in favour of some action more consistent w i t h the
speaker's u l t i m a t e goal.
Notes
* Current address: Unilever Research, Port
Sunlight Laboratory, Quarry Road East, Bebington,
Merseyside L63 3JW, UK
* * This
idea has appeared p r e v i o u s l y i n
Ellman (1980) w i t h an
argument based upon the
f u n c t i o n versus c o n t e n t of several
types
of speech acts, and also in Clark and Carlson
(1982).
These a u t h o r s reached t h i s p o s i t i o n
from a discussion of Mutual
Belief.
* * * A f u l l comparison of t h i s work w i t h that
of Allen et a l . is done in Ellman (forthcoming).
Acknowledgement
I thank Mark Steedman f o r s u p e r v i s i o n , support
any many s t y l i s t i c changes to t h i s paper.
References
A l l e n J . F . ( 1 9 7 9 ) "A P l a n Based A p p r o a c h to
Speech Act R e c o g n i t i o n "
University
of
Toronto Ph.d Thesis (Technical Report 131/79)
A l l e n J.F. and P e r r a u l t C.R. (1980) " A n a l y z i n g
I n t e n t i o n in Utterances" A I , 15 ppl43-178
A u s t i n J . L . ( 1 9 6 2 ) "How
to
do t h i n g s w i t h
words". Oxford U n i v e r s i t y Press (OUP) London.
C l a r k H.H. and M a r s h a l l C.R. (1981) " D e f i n i t e
Reference and Mutual Knowledge" in "Elements
of D l s c o u r c e U n d e r s t a n d i n g " J o s h i A.K., Sag
I . , and Webber B (eds.) CUP
C l a r k H.H. and
C a r l s o n T.B.
(1982) "Speech
Acts and Hearer's B e l i e f s " in Smith N. (1982)
Cohen P.R.(1978) "On K n o w i n g What to Say:
P l a n n i n g Speech A c t s " U n i v e r s i t y of Toronto
Ph.d Thesis (Technical Report 118/78)
Cohen P.R. and P e r r a u l t C.R. (1979) "Elements
of a Plan Based Approach to speech Acts" Cog.
Sci. 177-212
Dore J.
(1977) " C h i l d r e n ' s i l l o c u t i o n a r y acts",
in "Advances in d i s c o u r s e processes 1" R.O.
F r e e d l e (ed.) Ablex: Norwood N.J.
E l l m a n , J. (1980) Speech A c t s :
Representing
Requesting and U n d e r s t a n d i n g M.Sc. t h e s i s ,
U n i v e r s i t y of Essex
Ellman J. (forthcoming) "On Understanding Speech
Acts I n d i r e c t l y " U n i v e r s i t y of Warwick Ph.d
Thesis.
Grice H. P. (1975) "Logic and C o n v e r s a t i o n "
in Syntax and Semantics 3.
Hancher M. (1979) "The c l a s s i f i c a t i o n of cooperat i v e i l l o c u t l o r f a r y acts" Lang. Soc. 8, p 1-1A
Katz J.J.
(1977) "Propositional s t r u c t u r e and
i l l o c u t i o n a r y force" Harvester press.
P e r r a u l t , A l l e n , and Cohen (1978) "Speech Acts as
a basis f o r Understanding Dialogue coherence"
in TINLAP2 Waltz D. (ed) ACM.
Power R.(1979) "The O r g a n i z a t i o n of P u r p o s e f u l
Dialogues" L i n g u i s t i c s 17, 107152
S c h i f f e r S.R. (1972) "Meaning" OUP London
Searle J.R.(1969) "Speech Acts" CUP London
Searle J.R.(1975) " I n d i r e c t Speech A c t s " in Syntax and Semantics 3 "Speech Acts" (1975) Cole
P. and Morgan J.L.(eds) AP NY.
Searle J. R.(1976) "A c l a s s i f i c a t i o n of I l l o c u t i o n a r y a c t s " Lang. Soc. 5, pp 1-23
S m i t h N . and W i l s o n D . ( 1 9 7 9 ) " M o d e r n
L i n g u i s t i c s : The r e s u l t s o f Chomsky's
r e v o l u t i o n " Penguin Books, Harraondswoth U.K.
Smith N.V. ed (1982) "Mutual Knowledge" AP NY
S t i l e s B.(1981) " C l a s s i f i c a t i o n o f I n t e r s u b j e c t i v e I l l o c u t i o n a r y Acts" Lang. Soc.10,227-249
Vendler Z.(1972) "Res C o g i t a n s : an essay in
Rational Psychology" Ithaca: Cornell Univers i t y press.