Demonstrative pronouns, also indexicals, are shown to access en- tities t h a t are not LCs because they lack discourse relevance or because they are not yet in the uni- verse of discour
Trang 1S O M E F A C T S A B O U T C E N T E R S , I N D E X I C A L S , A N D D E M O N S T R A T I V E S
Rebecca J Passonneau Columbia University
450 C o m p u t e r Science Bldg
New York, New York 10027, U.S.A becky@cs.columbia.edu
A B S T R A C T 1 Certain pronoun contexts are argued to establish a
local center (LC), i.e., a conventionalized indexical
similar to l s t / 2 n d pers pronouns Demonstrative
pronouns, also indexicals, are shown to access en-
tities t h a t are not LCs because they lack discourse
relevance or because they are not yet in the uni-
verse of discourse
Referring expressions in discourse are multifunc-
tional and dual-faced Besides functioning to spec-
ify referents, they also indicate the status of their
referents in the evolving discourse model, such as
the informational status of being given or new
[Pri81], or maintain the attentional status of be-
ing in focus [Sid83] [Gro77] T h e y are dual-faced
in t h a t the surface form of a referring expression
is constrained by the prior discourse context, and
then increments the context, serving to constrain
subsequent utterances [Isa75] As a consequence
of this latter property, the communicative effect of
many referring forms, especially pronouns, is rel-
ative to specific types of discourse contexts T h e
discourse reference functions of a few types of pro-
nouns are examined, taking into account their mul-
tifunctionality and their dual nature, in order to
clarify their processing requirements in dialogic
natural language systems In particular, a compar-
ison of the conversational usage of it with two types
of indexical pronouns indicates that certain uses of
it, referred to as local centering, resemble what Ka-
plan [Kap89] refers to as pure indexicals Several
functions of lhat are also identified and shown to
contrast with local centering with respect to their
preconditions and effects
T h i r d person, definite (3d) pronouns contrast
with indexical pronouns because the referents of
the former are arbitrary, and must be actively es-
tablished as part of the current universe of dis-
course in order for the intended referent to be
1 This p a p e r was written under the support of D A R P A
grant N000039-84-C-0165 a n d NSF grant IRT-84-51438 I
am grateful to K a t h y McKeown for her generous support
identified In contrast, the referents of index-
icals such as I and you (i.e., the speaker and addressee) are necessary c o m p o n e n t s of the dis-
course circumstances 2 Indexical pronouns can
be further classified into pure indexicals versus demonstratives, 8 depending on how the current dis-
course circumstances provide their referents T h e referent of a pure indexical is fully determined by the semantic rules and a context, which together pick out a unique referent for each use T h u s I refers to the person who utters it (assuming that I
is used to refer) A pure indexical cannot refer to alternative entities, nor can any other expression pick out the relevant e n t i t y via the same type of referring function P u r e indexicals do not add en- tities to a context, or change the a t t e n t i o n a l status
of their referents
In contrast, the referent of a demonstrative pro- noun is not completely determined by the context
plus the semantic rules An accompanying demon- stration is required, such as a physical or vocal gesture to something in the immediate discourse circumstances Further, demonstratives can refer
to anything in the c o n t e x t t h a t can be demon- strated In the cases of discourse deixis discussed by Webber [Web90], e.g., d e m o n s t r a t i v e pronouns are used to refer to discourse segments Webber notes
t h a t in these cases, the d e m o n s t r a t i o n consists in the intention to refer signalled b y the use of the demonstrative, plus the semantics of the contain- ing clause, plus a t t e n t i o n a l constraints on which discourse segments can be demonstrated 4 Thus, 3d pronouns, pure indexical pronouns, and demon- stratives all differ with respect to the set of con- textual elements t h a t are available referents, and the m a n n e r in which the referent is related to the referring expression Investigating their distinct discourse functions leads to extensions to the tri-
2The t e r m indexical includes devices whose meaning per- talns to the time, the place a n d the perceived environment of
a discourse context, e.g., tense, deictic adverbs (here, there)
a n d deictic pronouns (this, that) [Pei35]
3The view of indexicals p r e s e n t e d here is largely drawn from K a p l a n [Kap89]
4Webber [Weh90] argues t h a t only segments on the right frontier are available referents
6 3
Trang 2p a r t i t e discourse model of a t t e n t i o n a l state, inten-
tional s t r u c t u r e and segmental s t r u c t u r e proposed
by Grosz and Sidner [GS86] 5
T h e d a t a presented here come from a set of dia-
logic interviews, originally described in [Sch85] (cf
also [PasS9]) T h e methodology, fully described in
[Pas90], primarily involves the examination of lin-
guistic choices t h a t are in principle independent,
but which are found to co-vary significantly of-
ten Such co-variation is presumed to serve commu-
nicative functions t h a t discourse processing models
need to replicate and explain It should be remem-
bered t h a t the p a t t e r n s of co-variation •described
here represent pragmatically significant usage pat-
terns, rather t h a n obligatory ones
2 L o c a l C e n t e r
In previous work, I presented the results of an anal-
ysis of the distribution of occurrences of it and
d a t a from 4 interviews, involving 5 different speak-
ers ( g = 678) [Pas89] T h e two pronouns have
similar syntactic contexts of occurrence thus dif-
ferences in their distribution are pragmatic in na-
ture, and stem primarily from the semantic con-
trast of d e m o n s t r a t i v i t y with non-demonstrativity
Previously, I had noted t h a t the d a t a supported
the centering rule (CR) [GJW83] and the p r o p e r t y
sharing principle ( P S P ) [Kam86] A review of the
assumptions of the centering model, and of the con-
versational data, argues for an alternative view In
this section I reinterpret the results as establishing
a distinct a t t e n t i o n a l state, local center I explain
the two p r o p e r t y sharing p a t t e r n s of K a m e y a m a ' s
P S P (subject and non-subject, [Kam86]) with re-
spect to local center, and discuss the similarity be-
tween local centers and pure indexicals Finally, I
discuss the relation • of local centering to intentional
and segmental structure
According t o the centering model, every utter-
ance has a backward-looking c e n t e r - - - t h e currently
most salient entity, b u t it need not be overtly men-
tioned in the c u r r e n t u t t e r a n c e [GJW83] T h e cen-
tering rule (CK) [GJW83], in combination with the
property-sharing principle ( P S P ) [Kam86], predict
certain preferred surface choices for maintaining
the backward-looking center (Cb) T h e C R says
t h a t when the same Cb is maintained in a new ut-
terance, it is likely to be expressed by ;a (3d) pro-
noun [GJW83] T h e P S P says t h a t when 3d pro-
nouns realize the Cb in adjacent utterances, they
5 T h e t e r m s e g m e n t a l s t r u c t u r e is u s e d in place of their
linguistic structure
FA and GR Lex Choice and Gr of N2
Pro,~on-SvB 43.1 21.9 21.9 19.1
NP,~o.-SUB 63.9 32.4 32.4 28.2
6 8 .0 4 I 0 0
OTH,on-SvB 48.4 24.6 24.6 21.4
1 3 3 1 5 3 1.3 6 3
Table 1: Effects of form and grammatical role of antecedents on pronoun choice, with observed fre- quency, expected frequency, and x-squares for each cell (individual cells are n u m b e r e d for convenient
• reference)
should b o t h be subjects (canonical center reten- tion) or b o t h not subjects (non-canonical center retention) [Kam86] Given t h a t the Cb can poten- tially be realized in non-preferred ways, t h a t the
Cb m a y change, or t h a t it m a y be unexpressed, Cb has m a n y possible surface realizations within a lo- cal discourse context of two s-adjacent utterances 6
T h e distinct effects of alternate realizations of Cb
on segmental s t r u c t u r e and intentional structure have not been explored Also, since the centering model focusses on 3d pronouns, no claims are made regarding the relation of indexical pronouns to the discourse model
T h e empirical results presented in [Pas89] showed t h a t two features of the utterances contain- ing a pronoun and its a n t e c e d e n t were extremely 6I u s e t h e s o m e w h a t a w k w a r d t e r m s-adjacen$ to connote adjacency w i t h respect to a c o n t a i n i n g s e g m e n t , a n impor-
t a n t a s p e c t of t h e Grosz-Sidner model; t h u s two s - a d j a c e n t
u t t e r a n c e s n e e d n o t be literally a d j a c e n t
Trang 3predictive of lexical choice between it and that: the
form of the a n t e c e d e n t (FA), and the g r a m m a t i -
cal role ( G R ) of b o t h expressions T h e best clas-
sifications were where FA had the three v a l u e s - -
p r o n o m i n a l a n t e c e d e n t ( P R O ) , full N P a n t e c e d e n t
(NP), and other ( O T H ) - - a n d where G R had the
two values -subject (SUB) and non-subject (non-
SUB) No other classifications of FA or G R were as
p r e d i c t i v e / It is crucial to note t h a t these classi-
fications were the m i n i m a l set t h a t still preserved
the distinctiveness of the distributions Seven other
surface features had previously been found to be
non-predictive [Sch85] s Table 1 shows a very
strong correlation (p 01%) 9 between the f o r m
and G R of the a n t e c e n d e n t (N1) and the lexical
choice and G R of the co-specifying pronoun (N2)
E x a c t l y 2 contexts selected for it, as shown b y the
combination of the high cell X2s, and the low val-
ues for expected frequency, which together indi-
cate t h a t the observed frequency was significantly
high These 2 contexts were where the a n t e c e d e n t
was P R O and where b o t h expressions maintained
the same G R value (cells 1, 7; PROGR, by itaR~)
O f these 2, t h e more significant context, and in-
deed the most significant context in the whole ta-
ble, was where the antecedent was PROGRsvv (cell
X 2 = 27.1) This is also the context t y p e where
the d e m o n s t r a t i v e was predicted not to occur (i.e.,
where the a n t e c e d e n t was PROscrBj; cells 2,4),
indicating a functional opposition between it and
that l° Most of the cases of the P R O antecedents
consisted of occurrences of it (65%), indicating t h a t
N1 and N2 often h a v e the s a m e form: it Previ-
ously u n r e p o r t e d d a t a bear on the likelihood t h a t
adjacent tokens of it will co-specify An analy-
sis of all adjacent u t t e r a n c e pairs where each con-
tained at least one token of referential it revealed
t h a t 30% were cases where b o t h were subjects, of
which 90% co-specified In contrast, it occurred
with opposing G R values 20% of the time, with
c o m p a r a t i v e l y fewer instances where b o t h tokens
co-specified (65%)
In sum, the d a t a show t h a t given ar~ occurrence
of it with an antecedent, the a n t e c e d e n t is likely
rCf [Sch85] [Sch84] for h o w it was d e t e r m i n e d t h a t t h e s e
were t h e m a x i m a l l y p r e d i c t i v e classifications
sViz., s p e a k e r a l t e r n a t i o n , clause t y p e ( m a i n o r s u b o r -
dinate), parallelism, a n d various m e a s u r e s of d i s t a n c e be-
t w e e n p r o n o u n a n d a n t e c e d e n t (e.g., intervening u t t e r a n c e s ,
i n t e r v e n i n g referents, s y n t a c t i c d e p t h ) N o t e also t h a t n o
significant variation w i t h r e s p e c t t o FA a n d ( l i t was f o u n d
across individual s p e a k e r s o r conversations
9 N o t e t h a t a p r o b a b i l i t y o f 5 ~ o r less is generally t a k e n
to b e higtfly significant
10 T h e r e m a i n i n g 4 o f t h e 8 P R O a n t e c e d e n t c o n t e x t s were
non-predictive
to be it, the G R of b o t h expressions is likely to be SUB, and in either case (SUB or non-SUB), t h e y will have the s a m e G R value T h e opposing G R
p a t t e r n is not predictive (where G R of N1 is not the s a m e as G R of N2) Nor is it predicted to oc- cur with an a n t e c e d e n t NP, and is predicted not
to occur with an a n t e c e d e n t O T H T h e 2 contexts
singled out here indicate t h a t it is a likely f o r m for
re-referring to a known, given e n t i t y - - b e c a u s e the
a n t e c e d e n t is P R O Conversely, successive occur-
rences of it in Ui and Ui+I generally co-specify if
they have the s a m e G R T h e e n t i t y referred to by
it in these two p a t t e r n s is called a local center T h e
following local center e s t a b l i s h m e n t (LCE) rule en- capsulates how a local center is a n t i c i p a t e d and maintained, b o t h for discourse u n d e r s t a n d i n g (.4) and generation (B)
A: R e c o g n i z i n g a L o c a l C e n t e r : Two s- adjacent utterances U1 and U2 establish en- tity £ as a local center only if U1 contains a 3ds pronoun N1 referring to g, U2 contains
a co-specifying 3ds pronoun N2, and N1 and N2 are both subjects or both non-subjects
In the canonical case, both are subjects
B: G e n e r a t i n g a L o c a l C e n t e r : To estab- lish g as a local center in a pair of s-adjacent utterances U1 and U2, use an expression of type N to refer to g in both utterances where each token, N1 and N2, is a 3ds pronoun, and each is the subject of its clause or each is not the subject of its clause In the canonical
case, both should be subjects (Precondition:
To establish an entity ,~ as a local center, C must be in the current focus space, and it must be possible to refer to it with a 3ds pro-
n o u n ) Recall f r o m §1 t h a t the process of interpreting
a pure index requires no search or inference, but depends only on how the discourse circumstances are c u r r e n t l y construed T h e s e m a n t i c value of a
pure index is a c o n t e x t u a l a t t r i b u t e - - e g , c u r r e n t
s p e a k e r - - t h a t m u s t have a particular referential value whenever an u t t e r a n c e occurs In m a n y ways,
a pronoun fulfilling the LC function is like a pure in- dex Discourse initially, there is no LC, because the
L C E rule d e p e n d s minimally on an u t t e r a n c e pair
B u t for a n y u t t e r a n c e pair where the L C E rule has applied, there will be a discourse e n t i t y - - a com-
p o n e n t of the speech s i t u a t i o n - - t h a t is by default indexed to the use of subsequent referring expres- sions with the right l e x i c o - g r a m m a t i c a l properties
An LC conforms to the characteristics of a pure in- dexical in t h a t it becomes established as a transient
a t t r i b u t e of the speech s i t u a t i o n analogous to the
essential a t t r i b u t e c u r r e n t speaker T h e relation
of the referent to the referring expression is one-
6 5
Trang 4to-one; no o t h e r referents are candidate LCs, and
no other form can access the LC T h e processing
mechanism for interpreting subsequent expressions
conforming to the L C E rule is highly constrained
It is analogous to, although not identical with, that
for pure indexicals T h e difference is t h a t the lo-
cal center is not lexicalized, but rather, must be
established and maintained by certain conventions
of usage C P s can choose not to establish a LC, or
can choose not to maintain it 11
K a m e y a m a [Kam86] proposed canonical and
non-canonical p r o p e r t y sharing patterns, but did
not discuss what governs the choice between them
Here it is suggested t h a t the non-canonical LC pat-
tern, illustrated in 1), results from the interac-
tion of two distinct pragmatic effects In the non-
canonical LC contexts, where the LC was realized
by non-SUB, the grammatical subjects were most
often 1st or 2nd person pronouns 12 This d a t a con-
forms to an empirically supported proposal made
by Givon and others [Giv76] [Li76] t h a t preferred
subjects are animate r a t h e r t h a n inanimate, defi-
nite r a t h e r t h a n indefinite, p r o n o u n s rather t h a n
full NPs, and 1st or 2nd person r a t h e r t h a n 3rd
person, due to the facts t h a t in English, gram-
matical subjects often express discourse topics (cf
also IF J84]), t h a t people prefer to talk a b o u t them-
selves and other people, and t h a t discourse topics
are given T h e interviews examined here were in-
tentionaUy biased towards the discussion of non-
animate entities, is B u t Givon's subject hierarchy
predicts that, given a non-animate and an animate
entity in a single utterance, the latter will more of-
ten occur as the subject Since every matrix-clause
utterance c a n have only one subject, there is po-
tential competition for the subject role T h e d a t a
show t h a t when SUB, reserved by the L C E rule for
establishing a local center, is pre-empted by a Ist or
2nd person pronoun, it is still possible for LC to be
realized by a l t e r n a t e means, namely by sharing of
non-SUB T h u s the sharing of the G R value across
utterances is a defining characteristic, as noted by
K a m e y a m a [Kam86] T h e non-canonical LC con-
11 T h a t C P s o f t e n d o m a i n t a i n a n L C is b o r n e o u t b y d a t a
p e r t a i n i n g t o cohesive chains, a s u c c e s s i o n o f u t t e r a n c e p a i r s
in w h i c h every u t t e r a n c e c o n t a i n s a co-speclfying p r o n o u n
token T h e r e were 101 cohesive c h a i n s in t h e i n t e r v i e w d a t a ,
r a n g i n g in l e n g t h f r o m 2 t o 13 successive u t t e r a n c e s , con-
t a l n i n g 506 p r o n o u n s , t h e m a j o r i t y o f w h i c h involved LC
c o n t e x t s ; cf [Pas90]
12The t w o n e x t m o s t likely possibilities were a n sallmate
full NP, o r a n o n - r e f e r e n t i a l p l e o n a s t i c e l e m e n t , e.g., existen-
tial there A f t e r t h a t , t h e r e was a very small h e t e r o g e n e o u s
category Note: s u b j e c t a l w a y s refers t o a surface g r a m m a t -
ical function
13E.g., college courses, degree r e q u i r e m e n t s , c a r e e r op-
tions, resttm~s, a n d s o o n
figuration results from an interaction between two separate organizing forces: the LC status of the 3d pronoun referent, and the a t t e n t i o n a l prominence
of the speaker and hearer
(1)
S l a : Slb:
S1¢:
Sxd:
S l e : S2a:
S2b:
$3 :
I don't have t h e m e n t a l c a p a c i t y
to handle uh what I would like to teach which'd be philosophy
or history at U of C
uh with t h a t level students u m
m a y b e with time and experience I'll gain it
b u t I d o n ' t have it now
In example 1), the u t t e r a n c e pair $2 and $3 share a 1st person subject and a non-canonical lo- cal center 14 In this case, the centering model can- not provide a principled answer to the question of whether the s p e a k e r - - t h e grammatical s u b j e c t - -
or the speaker's 'mental c a p a c i t y ' - - r e f e r r e d to by successive 3d pronoun direct o b j e c t s - - i s the cur- rent center In the model proposed here, $2 and S~ establish ' m e n t a l capacity' as a local center, an at- tentional status for regulating the generation and production of 3ds pronouns, and the question of which e n t i t y is more salient does not arise B u t lo- cal centering does seem to have a secondary func- tion pertaining to the linkage between utterances
at the level of intentional and segmental structure
In addition to sharing a default referent, clauses containing LC pronouns are often semantically alike in other ways In an initial a t t e m p t to test for this similarity, u t t e r a n c e pairs with P R O an- tecedents were classified into those t h a t did and did not conform to the LCE rule No other con- texts were examined because contexts with O T H and N P antecedents were presumed to be even less like LC contexts These u t t e r a n c e pairs were sorted into cases where the lexical root of the m a t r i x verb
in b o t h clauses was identical (i.e., the verb of which the pronoun was an argument), but where the ut- terances were not v e r b a t i m repetitions 15 T h e re- sults were t h a t 30% of the LC contexts had the same verb, b u t only 11% of contexts differing from
LC in t h a t N2 was that instead of it None of the contexts with opposing GI~ values for the two pronouns had the same verb, which is not surpris- ing given t h a t for most verbs, each a r g u m e n t po- sition has a very distinct semantic role In sum,
by maintaining an LC and the same lexical verb, 14In interview e x c e r p t s , S is t h e s t u d e n t a n d C t h e coun- selor F e e d b a c k cues f r o m t h e a d d r e s s e e i n d i c a t i n g contin-
u e d a t t e n t i o n (e.g., uhhuh) h a v e b e e n o m i t t e d 15In c o p u l a r clauses, t h e b e - c o m p l e m e n t s were c o m p a r e d
i n s t e a d o f t h e verb; ellipsis w a s c o u n t e d as identity
Trang 5the speaker continues to predicate the same type
of information a b o u t the same entity This pre-
sumably serves as a cue t h a t the speaker main-
tains a common Discourse Segment Purpose (DSP,
[GS86]) t h r o u g h o u t b o t h u t t e r a n c e s - - t o convey in-
formation a b o u t the local center with respect to
the state of affairs conveyed by the verb Insofar as
local centering pertains to segment continuation,
or to relating a new utterance to the DSP of a
preceding utterance, a discourse plan to continue
the current DSP need not refer directly to the sur-
face grammatical choices which reflect t h a t plan,
but only to the current status of LC If there is a
current LC, then maintaining it would reflect the
speaker's intention for the next utterance to con-
tinue the same DSP as the prior utterance
T h e d a t a assembled here indicate t h a t local cen-
tering not only constrains the interpretation of cer-
tain pronouns, but also conveys the inter-utterance
relevance of locally centered entities in a larger dis-
course segment, or in the discourse as a whole
However, most of the entities referred to in the con-
texts represented in Table I are not LCs Logically,
t h a t means they can fall into several classes: e.g.,
entities that are former or potential LCs because
they are in the universe of discourse and are rele-
vant to a former or future DSP; entities t h a t are
in the universe of discourse but are not LCs be-
cause they are peripheral to the current DSP; and
finally, entities t h a t are not yet in the universe of
discourse T h e next section will illustrate how the
demonstrative picks out entities in the latter two
c l a s s e s
3 N e w E n t i t i e s , A n t i - c e n t e r s ,
a n d N o n E n t i t i e s
T h e results presented in the preceding section in:
dicate t h a t referential it has different discourse ef-
fects, depending on its grammatical role, and on
various properties of its antecedent, which in turn
depend on the status of the referent in the discourse
context Just as local centering is only one dis-
course referring function that it participates in, it
will be seen t h a t there are several referring func-
tions the demonstrative participates in, each with
distinct preconditions and effects Although pro-
nouns are often t h o u g h t of as identifying topical
entities, that is not necessarily the case English
has a relatively impoverished inventory of pronouns
in comparison to the B a n t u language Chich~wa,
which has two sets of definite pronouns, one of
which is morphologically incorporated into the verb
stem, and the other of which consists of indepen-
NP Antecedent IT THAT
Probability " ] , 0001 Table 2: Givenness and Lexical Choice of P r o n o u n
dent morphemes IBM87] is In their analysis of Chich~wa, Bresnan and M c h o m b o argue t h a t of the two n o n - a r g u m e n t grammatical roles in LFG, WOP(ic) and f O C ( u s ) , the independent pronouns can only fill the F O C role, not T O P [BM87] In their framework, no expression can simultaneously
be T O P and FOC x7 This is reminiscent of the
pragmatic contrast in English between it and that
in focus-marking constructions, as illustrated in 2a-
b) below That is acceptable, while it is not, as a
syntactically focussed element:
(2)
a T h a t / * I t I b o u g h t for m y mother, but I could get a n o t h e r one for y o u
b P e p p e r is okay, b u t d o n ' t add more curry It's ? t h a t / * i t t h a t makes me sneeze These examples are compatible with the conver- sational d a t a in the following way If T O P and
F O C are truly contrastive grammatical functions,
the above examples show t h a t that is more accept- able as FOC We have seen t h a t it is less likely
when the antecedent is N P or O T H t h a n when it
is P R O , that it occurs often as SUB, and often
with SUB antecedents T h u s it, w h e t h e r fullfill-
ing LC or not, correlates with other properties of discourse topics An e n t i t y t h a t has been referred
to by an antecedent pronoun has already been lo- cated in the universe of discourse, and already has the informational status of given prior to the oc- currence of the pronoun itself, and thus is a likely
topic We have also seen that that is unlikely with
P R O s v B antecedents, which would correlate with
a presumed likelihood for that to not express TOP
But further evidence regarding the informational
and attentional status of the likely referents of that
reinforces the presumed T O P / F O C contrast
T h e first case we'll look at involves NP an- tecedents Table 2 shows the distribution of an- tecedent NPs, classifed according to whether they
were given or not, by lexical choice of it or that
A referent was classified as given if it had been 16In a d d i t i o n , t h e r e is a s e p a r a t e s e t o f d e m o n s t r a t i v e
p r o n o u n s 17More specifically, n o t a t t h e s a m e level of L F G f u n c -
t i o n a l c l a u s e s t r u c t u r e
67
Trang 6mentioned previously, if it was closely associated
with a previously mentioned e n t i t y (e.g., social
worker and the social work profession), or if it was
a commonly known individual e n t i t y whose iden-
t i t y would would be known to either speaker (e.g.,
places such as New York City) T h e very low prob-
ability for the X 2 of Table 2 (p = 01%) indicates
t h a t the tendency for that to occur with new an-
tecedents and for it to occur with given antecedents
is extremely significant F u r t h e r classifying the lo-
cal u t t e r a n c e contexts by G R in various ways did
not reveal any f u r t h e r significant distinctions This
result, while not counter-intuitive, is not one t h a t
would be obvious w i t h o u t looking at frequency dis-
tributions in actual on-line discourse, since it can
easily and naturally be used to co-specify with a
new antecedent, or that with a given antecedent
Some examples from the interviews are shown in
3-4) with the relevant pronoun token and its NP
a n t e c e d e n t in boldface T h e y have been particu-
larly selected to show t h a t the occurring pronoun
can be felicitously replaced with the opposite choice
(shown in parentheses)
(3)
Cla:
Clb:
Clc:
C~ :
Ca :
(4)
it is the service t h a t you give to other
people be it as a d o c t o r or a social
worker a psychiatrist or a lawyer
you have a c e r t a i n e x p e r t i s e
and people use t h a t (it)
C1 "
C2a:
Cab:
Ca :
C4 :
I know we've had information a b o u t it
and uh if not you can a-
just write directly to B r y n Mawr
and ask t h e m a b o u t t h e p~ogrRm
and see if t h e y still have it ( t h a t )
One way to i n t e r p r e t these results is t h a t a single
reference to a new e n t i t y is insufficient to establish
the e n t i t y as p a r t of the universe of discourse, given
the processing demands of actual on-line discourse
In the cases where an e n t i t y is already given, b u t
is referred to by a full N P r a t h e r than a pronoun
(for whatever reason), the e n t i t y can be successfully
reinvoked in the immediately following utterance
by a 3ds pronoun If the e n t i t y is new, a single prior
mention is not in general sufficient, with respect to
these data, to predispose the use of a 3ds pronoun
to reinvoke it Instead, the demonstrative functions
to incorporate these new entities into the context
T h e d e m o n s t r a t i v e has a n o t h e r singular func-
tion with N P antecedents Table 1 singles out
2 significant contexts where there was a full NP
N P s u B / T H AT sub 3:t 9
Table 3: Subsequent Discourse Relevance
antecedent (cells 13, 16) If the a n t e c e d e n t was
an NPnonstrBs, there was an increased likelihood for thatnonstrns and a decreased likelihood for
itst~B.r Because itsunJ is the canonical indi- cator of LC, and because LCs are presumed to have discourse relevance (i.e., play a central role
in the current DSP), I hypothesized t h a t the link- age between an a n t e c e d e n t NPnonSUBJ and a co- specifying thatnonSUBJ served to mark the referent
as being unlike a local center by being peripheral
to the current DSP This was tested by examin- ing how often an e n t i t y mentioned in the NP con- texts was mentioned later in the discourse Table
3 depicts the contexts in which an a n t e c e d e n t NP was followed by it or that, where G R for each was SUB or non-SUB, or where the G R values differed (X) These 6 contexts were coded for whether the referent was referred to again within the 10 utter- ances following the u t t e r a n c e containing the pro- noun If so, the e n t i t y was coded as relevant; else
it was non-relevant T h e low probability of 1.6% indicates a significant correlation T h e 2 cells con- tributing the most to the overall significance were for the NPno~-StrB/THATno,,-strB context, with non-relevant entities occurring significantly often, and relevant entities occurring significantly rarely This evidence supports the view t h a t the features
of this context function to re-invoke entities while simultaneously signalling their peripheral status
T h e final referring function discussed here is where the d e m o n s t r a t i v e has an O T H antecedent
W h e n N1 is OTHnonSUB (contexts 21-24), itSUB
is unlikely (context 21), and both cases of thatsuB
(context 22) and thatno,,-SVB (context 24) are sig- nificantly frequent I will argue t h a t these O T H contexts exemplify i n t r a - t e x t u a l deixis, which is analogous to the cases of discourse deixis stud- ied by W e b b e r [Web90] I refer to these cases
as intra-textual deixis because the deictic refer- ence involves referents related to grammatical con- stituents r a t h e r t h a n to discourse segments
Trang 7In previous work, I pointed out t h a t the criti-
cal feature of the a n t e c e d e n t t y p e which favors the
lexical choice of that is syntactic, namely the dis-
tinction between N P s with lexical noun heads and
other types of constituents [Sch84] C o n t e x t s where
N1 is an NP whose head is a derived nominaliza-
tion (such as the careful choice of one's words)pat-
tern like those where the head is a lexical noun Is
Gerundives fall into the O T H class Unlike NPs,
the O T H antecedents c a n n o t be m a r k e d for def-
initeness: *a/*the carefully choosing one's words
versus a/the careful choice of words Definiteness
is one of the means for indicating whether a refer-
ent is presupposed to be p a r t of the current context
T h u s a possible difference between the interpreta-
tion of the two t y p e s of phrases carefully choosing
one's words and a careful choice of words would
have to do with w h e t h e r there is a discourse e n t i t y
in the context as a consequence of the occurrence
of t h e p h r a s e itself
(5)
V i a :
C:~a :
C2b:
C s a :
Csb:
C4 :
there are some books t h a t we
have t h a t talk a b o u t interviewing
u m o n e ' s called S w e a t y P a l m s
which I think is a great title (laugh)
u m b u t i t talks very interestingly
a b o u t h o w t o g o a b o u t i n t e r v i e w i n g
and t h a t ' s t h a t ' s going to be i m p o r t a n t
A n o t h e r feature of O T H a n t e c e d e n t s pertains to
their ability to evoke specific entities into the uni-
verse of discourse C o m p a r e the two pronouns in
example 5) T h e token of it in C~a unambiguously
refers to the one book called Sweaty Palms T h e
referent of that in C4 is much harder to pin down
Does it correspond to interviewing, or to how to
go about interviewing? This example illustrates an
inherent vagueness in the processing of finding a
textual referent for a d e m o n s t r a t i v e which I will
now describe in more detail
W e b b e r [Web90] notes t h a t deictic reference is
inherently ambiguous, although I prefer the t e r m
vague, in t h a t vagueness connotes an underspeci-
fled i n t e r p r e t a t i o n t h a t can be given a n u m b e r of
more specific readings W e b b e r argues persuasively
t h a t deictic reference to a discourse segment is re-
stricted to references to open segments on the right
frontier, but t h a t there is still an ambiguity as to
which segment might be referred to, due to the
recursive n a t u r e of discourse segmentation Since
an open s e g m e n t on the right frontier m a y contain
lSMixed nominals, such as the careful choosing of one's
words, occurred too rarely to have a discriminating effect on
contexts favoring it or that
within it an e m b e d d e d open s e g m e n t t h a t is also on the right frontier, a token of the d e m o n s t r a t i v e t h a t refers to a discourse s e g m e n t can be a m b i g u o u s be- tween a more inclusive s e g m e n t and a less inclusive one [Web90] T h e vagueness m a y b e eliminated if the context in which the deictic expression occurs clearly selects one of the possible readings This
p h e n o m e n o n p e r t a i n i n g to deictic reference to seg-
m e n t s is replicated in the cases where that has an
O T H antecedent, thus in C4 of 5), the antecedent of the d e m o n s t r a t i v e p r o n o u n could be interviewing,
or the more inclusive expression go about interview- ing, or the more inclusive one yet how to go about interviewing I will now argue t h a t such expres-
sions do not in and of themselves introduce entities into the universe of discourse
(6)
UI:
V2:
Us:
I noticed t h a t Carol insisted on sewing h e r dressesk f r o m non-synthetic fabric That's an e x a m p l e o f how observant I am
And t h e y k always t u r n out beautifully
(7)
UI:
V 2 :
Us:
I noticed t h a t C a r o l i insisted on
sewing her dresses f r o m non-synthetic fabric
T h a t ' s an e x a m p l e of how o b s e r v a n t I am
* T h a t ' s because s h e i ' s allergic to synthetics
(8)
UI:
V 2 :
U3:
I noticed t h a t C a r o l / insisted on sewing her dresses f r o m non-synthetic fabric
S h e / s h o u l d t r y the new rayon challis
* T h a t ' s because she's allergic to synthetics
T h e examples in 6)-8) show t h a t entities intro- duced b y referential N P s in U1 are still available for p r o n o m i n a l reference in Us, after an intervening U2 Ux introduces the referring expressions Carol
and her dresses E x a m p l e 6) shows t h a t the refer-
ent of her dresses is still available in U 3 even though
it is not mentioned in U2 Instead, Us contains
a pronoun t h a t refers to the fact t h a t is asserted
b y the whole u t t e r a n c e U1 In contrast, the refer- ent of the non-nominal sentence constituent Carol insisted on sewing her dresses from non-synthetic fabric is not available after an intervening sen-
tence t h a t contains a deictic reference to a differ- ent non-nominal constituent, as in 7), or after an intervening sentence t h a t contains a reference to a discourse e n t i t y m e n t i o n e d in U1, as in 8)
T h e preceding examples show t h a t O T H con- stituents do not introduce entities into the dis- course context W i t h such antecedents, the d e m o n -
69
Trang 8s t r a t i v e does not access a pre-existing discourse en-
tity, b u t rather, plays a role in a referring function
b y virtue of which a new discourse e n t i t y is added
to the context T h e occurrence of the d e m o n s t r a -
tive triggers a referring function t h a t is constrained
b y the semantics of the d e m o n s t r a t i v e p r o n o u n
and its local s e m a n t i c context, t h e antecedent, and
other contextual considerations T h e result of ap-
plying an a p p r o p r i a t e referring function is to in-
c r e m e n t the c o n t e x t with the new discourse e n t i t y
t h a t is found to be the referent of the d e m o n s t r a -
tive pronoun
T h i s investigation has shown t h a t a p r o n o u n
does not achieve discourse reference in and of it-
self In c o m b i n a t i o n with various linguistic prop-
erties of the prior utterance, and depending on the
s t a t u s of the referent in the context, a p r o n o u n m a y
have distinct referring functions Although this in-
vestigation has focussed primarily on n o n - a n i m a t e
pronouns, future research is expected to show t h a t
elements of t h e c o n t r a s t b e t w e e n it and that oc-
cur with a n i m a t e 3d p r o n o u n s (e.g., he, she) since
these p r o n o u n s h a v e b o t h d e m o n s t r a t i v e and non-
d e m o n s t r a t i v e uses
R e f e r e n c e s
IBM87]
[rJ84]
[Giv76]
[GJW83]
[Gro77]
[GS86]
[Isa7S]
Joan Bresnan and Sam A Mchombo Topic,
pronoun and agreement in Chiche@a In
M Iida, S Wechsler, and D Zec, editors,
Working Papers in Grammatical Theory and
Discourse Structure, pages 1-60 CSLI, 1987
William A Foley and Robert D Van Valin
Jr Functioned Syntoz and Universal Gram-
mar Cambridge University Press, Cam-
bridge, 1984
Talmy Givon Topic, pronoun, and gram-
matical agreement In Charles N Li, editor,
Subject and Topic, pages 149-188 Academic
Press, New York, 1976
Barbara J Grosz, A K Joshi, and S Wein-
stein Providing a unified account of definite
noun phrases in discourse In Proceedings of
the $lst ACL, pages 44-50, 1983
Barbara Grosz The Representation and Use
o] Focus in Dialogue Understanding PhD
thesis, University of California, Berkeley,
1977
tention, intentions and the structure of dis-
course Computational Linguistica, 12:175-
204, 1986
S Isard Changing the context In
E.L.Keenan, editor, Formal Semantics o]
Natural Language, pages 287-296 Cam-
bridge U Press, Cambridge, 1975
[Kam86]
[Kap89]
[Li76]
[Pas89]
[Pasg0]
[Pei35]
[Pri81]
[Sch84]
[Sch85]
[Sid83]
[Webg0]
Megumi Kameyama A property-sharing constraint in centering In Proceedings of the 24th Annual Meeting of the A CL, pages 200-
206, 1986
David Kaplan Demonstratives In J Almog,
J Perry, and H Wettstein, editors, Themes from Kaplan, pages 481-566 Oxford Univer-
sity Press, New York, 1989
Charles N Li Subject and Topic Academic
Press, New York, 1976
Rebecca J Passonneau Getting at discourse referents In Proceedings of the 27th Annual Meeting of the ACL, pages 51-59, 1989
Rebecca J Passonneau Getting and keep- ing the center of attention In R Weischedel and M Bates, editors, Challenges in Natu- ral Language Processing Cambridge Univer-
sity Press, 1990 To appear; also available as Tech Report GUCS-060-90, Dept of Com- puter Science, Columbia University
Charles S Peirce In C Hartshorne and
P Weiss, editors, Collected Papers o] Charles Sanders Peirce Harvard University Press,
Cambridge, MA, 1931-35
Ellen Prince Towards a taxonomy of given- new information In P Cole, editor, Radical Pragniatics, pages 223-55 Academic Press,
New York, 1981
Rebecca J (Passonneau) Sehiffman The two nominal anaphors it and that In Proceedings
of the 20th Regional Meeting of the Chicago Linguistic Society, pages 322-357, 1984
Rebecca J (Passonneau) Schiffman Dis- course Constraints on it and that: A Study o] Language Use in Career-Counseling Inter- views PhD thesis, University of Chicago,
1985
Candace L Sidner Focusing in the compre- hension of definite anaphora In M Brady and R C Berwick, editors, Computational Models of Discourse, pages 267-330 The
M I T Press, Cambridge, Massachusetts, 1983 Bonnie L Webber Structure and osten- sion in the interpretation of discourse deixis Technical Report MS-CIS-90-58, LINC LAB
183, University of Pennsylvania Computer and Information Science Department, 1990
To appear in Language and Cognitive Pro- cesses