Generating Cooperative System Responses 
in Information Retrieval Dialogues 
Markus Fischer 
Information Technolpgy Research 
Institute, University of Brighton (ITRI) 
markus.fischer~itrl.bt on.ac.uk 
Elisabeth Maier 
German Research Center for Artificial 
Intelligence GmbH (DFKI) 
maier@dfki.uni-sb.de 
Adelheit Stein 
Integrated Publication and Information 
Systems Institute (GMD-IPSI) 
st ein~darrastadt .grad.de 
Abstract 
This paper describes the Corinna system which integrates 
a theoretical approach to dialogue modeling with text 
generation techniques to conduct cooperative dialogues 
in natural language. It is shown how the dialogue model 
COR can be augmented by adding discourse relations as 
an additional level of description which is particularly 
valuable for the generation of dialogue acts. 
1 Introduction 
Text planning aid generation become more and more im- 
portant as components of intelligent multimodal user in- 
terfaces. It has been pi:oposed (see, e.g., Stein and Thiel, 
1993) to model interaction with an intelligent informa- 
tion system as a conversation between two participants. 
The conversation metaphor applies to natural language 
as well as to graphical and multimodal user interfaces, if 
an appropriate model is provided. Employing features of 
human-human communication into interface design (even 
if only to a small exterit), the system can respond to the 
user in a more natural way. Natural communication does 
not necessarily imply natural language, but undoubtedly, 
language plays an imp, ortant role in conversation. The- 
ories and tools developed in the areas of Natural Lan- 
guage Processing (NLP) and Natural Language Genera- 
tion (NLG) can help making models for conversations (or 
dialogues, as we will call it for the rest of the paper) more 
adequate and richer in Itheir descriptive power. 
The speech-act oriented dialogue model COR (COnver- 
sational Roles) has been developed at GMD-IPSI (Sit- 
ter and Stein, 1992). It covers the genre of informa- 
tion seeking human-computer interactions as opposed to 
human-human conversations and other (less restricted) 
genres such as spontaneous conversations, narratives, ar- 
guments, etc. Due to its strong focus on the pragmatics 
of a dialogue, leaving semantics aside, COlt is a domain 
independent conversation model. It is general in that it 
covers the basic illocutionary aspects and role expecta- 
tions of cooperative information-seeking exchanges. 
In the search for ways to enhance the COR model fur- 
ther, we found the Rhetorical Structure Theory (RST) 
to be a very good candidate. This theory - initially de- 
veloped for monologues only - has been formalized and 
integrated into many text generation and text planning 
systems, one of which has been developed at IPSI (of. 
Bateman et al., 1991). iSome empirical studies were car- 
ried out to find out possible points of integration (Fischer, 
1993). The results were very promising and hence used 
to specify a dialogue manager of a prototypical informa- 
tion system, called Corinna. This dialogue system uses 
automatically generated natural language as its major in- 
teraction modality. The incorporation of RST within the 
COR dialogue model served as an important parameter 
within the text generation process. 
The remainder of the paper is divided into two main 
parts: the theoretical part concludes with the descrip- 
tion of our own approach of integrating COR and RST 
(section 2.2), which is then elaborated and exemplified in 
the second part of the paper (section 3) presenting the 
basic features of the Corinna system. 
2 Theoretical Framework 
2.1 Related Work 
Research activities in several areas, such as NLG, dialogue 
modeling, information retrieval and multimedia interfaces 
played an important role in motivating our work. Two 
streams of research were particularly interesting in our 
context: on one hand the incorporation of dialogue mod- 
els in natural language systems, on the other hand the 
extension of RST and its application for (not exclusively 
natural-language-based) dialogue systems. The following 
gives some examples of work done in these two areas. 
As part of the Communal project (cf. Fawcett e~ al., 
1988), which includes generation as well as understand- 
ing of natural language, a dialogue model called SFM 
(Systemic Flowchart Model) was developed. It uses a 
discrimination network to describe situations and actions 
that can occur in a dialogue. Due to the fact that many 
different speech acts (based on Searle, 1969) and speech 
act sequences were to be considered, the network is quite 
complex. Attempts were also made to integrate this 
model with RST(cf. Fawcett and Davies (1992) and sec- 
tion 2.1.2). 
A system which is capable of performing dialogues with a 
user on the basis of speech, was proposed by Smith, Hipp 
and Biermann (1992). Its domain is the maintenance of 
electrical appliances, and the emphasis in this approach 
lies on (nested) communicative goals, and concepts such 
as intentional, attentional and linguistic structures (Grosz 
and Sidner, 1986). 
Another system for the treatment of spoken dialogues is 
reported in Bilange (1991). The approach, which has 
been developed in the framework of the SUNDIAL project, 
207 
7th International Generation Workshop • Kennebunkport, Maine • June 21-24, 1994 
is based on the assumption that dialogues can best be de- 
scribed by means of a multi-level approach. The author 
distinguishes four levels: a domain-specific transaction 
level, an exchange level, an intervention level modeling 
initiative, reaction and evaluation, and finally a level con- 
sisting of dialogue acts. The system has been developed 
for the domain of flight reservations. 
Plan-oriented approaches for dialogue modeling are de- 
scribed in Litman and Allen (1987) and Lambert and 
Carberry (1992). Both approaches distinguish domain 
or task, problem-solving and discourse levels. Knowledge 
from the various levels is employed to solve the task of 
plan recognition in dialogues. Connections do not only 
exist between the various levels but also between elements 
within one level. These links are modeled either as dis- 
course plans which follow the course of the interaction 
(e.g. CONTINUATION, CLARIFICATION and TOPIC-SHIFT, 
see Litman and Allen(1987)) or as discourse actions that 
link an utterance with" the context. A similar distinction 
of various levels of representation is made in O'Donnell 
h1990), except that the links between various elements of 
is discourse or exchange model are not made explicit. 
The EES project (Explainable Expert Systems) is the ba- 
sis for the theoretical and practical work of Moore and 
Paris (1989), Moore and Swartout (1990), and Carenini 
and Moore (1993). A central goal in EES was the cre- 
ation of a fiexibl~ explanation module for expert systems 
communicating in natural language, allowing the user to 
ask questions about explanations given by the system 
and generating appropriate system responses. Strate- 
gies incorporating parameters such as context and focus 
were used to disambiguate the user's utterances. To ad- 
dress the intentions behind utterances and communica- 
tive goals, concepts of RST (cf. Mann and Thompson, 
1987) and Speech Act Theory were exploited. The fo- 
cus, however, is on flexible explanation dialogues and, 
unlike COR-RST, not on modeling information-seeking 
dialogues as complex "negotiations" with flexible ways to 
withdraw and reject dialogue contributions. 
The Intelligent Documentation Advisory System (IDAS), 
developed by Reiter, Mellish and Levine (1992) repre- 
sents an attempt to use dynamically generated natural 
language in the framework of an information retrieval sys- 
tem using hypertext techniques. To obtain information, 
the user clicks on the object under consideration and then 
chooses one out of a list of request options displayed by 
the system. However, there is no dialogue model at all, 
the system only allows simple query - answer cycles. 
The concepts of speech acts in combination with RST are 
used in several multimedia presentation systems. Among 
the first systems following this approach were the WIP 
system developed at DFKI (see Andrd and Rist (1993)) 
and the system developed by Maybury (1991). However, 
neither the WIP project nor Maybury's system use high- 
level dialogue structures. As pointed out in Arens et 
al. (1993) global structures are necessary for establish- 
ing overall coherence in the context of multimedia inter- 
faces. COR-RST takes this into account. It focuses on 
Natural Language Generation, yet allows the extension 
to multimodal dialogue acts. 
In the following two sections the two theoretical ap- 
proaches (COR and RST) that were most influential for 
our work will be presented. Especially the COR model 
will be described in detail, because it is essential for un- 
derstanding our approach. 
2.1.1 The Conversational Roles Model (COR) 
In the field of information retrieval (IR) the interactive 
and communicative aspects of IR have only recently been 
emphasized (cf. for example, Belkin and Vickery, 1985; 
Belkin et al., 1993). There exist approaches to distin- 
guish various types of information retrieval strategies and 
tactics (cf. Bates, 1979), task hierarchies and global 
phases of the interaction. Itowever, no elaborate inter- 
action models are provided in this field (except simplistic 
iterative question-answer models). In the area of conver- 
sational analysis and discourse theory, on the other hand, 
we find various discourse and dialogue models which ad- 
dress local dialogue structures (e.g., Fawcett et al., 1988; 
Grosz and Sidner, 1986, 1990; Reichman, 1985). 
To be able to design a flexible dialogue system which can 
engage in cooperative information-seeking dialogues we " (° _ 4 
use the Conversational P~oles" model COR) developed 
by Sitter and Stein (1992). It has been used to design 
the interface of a multimedia information system, called 
MERIT (cf. Stein et al., 1992). The COR model was orig- 
inally influenced by the "Conversation for Action" (CfA) 
model (Winograd and Flores, 1986) which was applied to 
design computer-aided human-human interactions. 
By adopting basic concepts of speech act theory and ex- 
isting discourse models, and extending the CfA model for 
the situation of information-seeking human-computer in- 
teractions, the COR model shows the following features: 
• it depicts the interaction as a cooperative two-party 
"negotiation" where commitments (to supply infor- 
mation or meta-information) can be made, retracted 
or rejected; 
• it permits mixed-initiative dialogues and is flexible 
enough to describe all possible - even extremely com- 
plex - interaction patterns (this includes the tempo- 
rary role changes of information seeker/information 
provider, which frequently occur in highly vague task 
settings such as information-seeking); 
• it provides the means for an explicit representation 
of the dialogue history in an abstract form, i.e., dis- 
regarding the interaction mode (graphical, linguistic, 
mixed). 
According to COR the two participants (A and B) have 
dialogue goals and pursue specific conversational tactics 
to achieve these goals. The speaker's and addressee's 
mutual expectations about possible responses and about 
the subsequent course of the dialogue are essential. The 
model can be represented as a recursive state-transition 
network (see figure 1). The network defines the full 
potential of all possible interactions where successfully 
completed dialogue contributions/acts end in specialized 
states (circles). Transitions (arcs) represent the various 
types of dialogue acts: e.g., REQUEST, OFFER, and IN- 
FORM can exactly be mapped onto Searle's basic "illo- 
cutionary types": directives, commissives, and assertives 
(cf. Searle, 1979); the other generic acts belong to the 
same categories, but they are less significant bec.au.se they 
are merely responsive, e.g., PROMISE is a commlss~ve act, 
but it adopts the conditions of action expressed by the 
preceding REQUEST (cf. Sitter and Stein, 1992). The 
208 
7th International Generation Workshop • Kennebunkport, Maine * June 21-24, 1994 
Dialogue (A,B) \['~ 
est (B,A) 
\\ ~ ~'au~araw oner t~,~) / \\ be discontented \\ offer (B,Ap',,,.,~__ / ac.~,t (A,B) \ \ (A,B) 
~.~.~thdr'a~ w'ithdra., ~rejeet offer withdraw ~ withdraw. diree~A,B) directive ~ ~eornmissive 
(A,B) (13,A) N 
16 I171 L~J I._~.ll~9.J 
Figure 1: The DIALOGUE net of the COR model 
traversal of the graph stops in states which are marked by 
squares. State < 5 >, for instance, is reached when the in- 
formation seeker (A) h'as expressed contentment with the 
given information and quits the dialogue. States < 6 > to 
< 11 > are also terminal states, but here the information 
need could not be satisfied. Note that a dialogue which 
ends in one of these states can be well-formed, coopera- 
tive and complete (e.g,, B rejects a request of A, because 
the requested information could not be retrieved). 
The bold arcs leading from state < 1 > to state < 5 > de- 
note two "idealized" courses of the interaction which fol- 
low the basic role-expectations or role assignments. The 
two initiative acts (REQUEST, OFFER) typically establish 
new.conditions of action, whereas the subsequent acts are 
reactive and do not introduce new conditions (PROMISE, 
ACCEPT, INFORM, and BE-CONTENTED are all "expected" 
in that they are positive responses to the preceding acts). 
Really encountered information-seeking dialogues, how- 
ever, often do not follow such a simple, linear conversa- 
tional development. Directive acts (REQUEST, ACCEPT) 
can be rejected by the addressee, commissives (OFFER, 
PROMISE) are often withdrawn. Both rejections and with- 
drawals (called "alternative" acts or responses) can either 
lead back to state < 1 >, where the dialogue is entered 
again (begin of a new: dialogue cycle), or to a terminal 
state (definite REJECT, WITHDRAW). 
Another way of departing from the linear course of in- 
teraction is even mor e important. So far we have only 
considered "atomic" dialogue acts which are not further 
decomposed. But consider the following situations which 
often occur in information-seeking interactions: if the 
meaning of an utterance (atomic act) has not been under- 
stood by the addressee, or, if she needs additional infor- 
mation to be able to proceed, an embedded clarification 
dialogue might be necessary. In order to resolve this prob- 
lem, the transitions in. figure 1 must not be interpreted 
as atomic acts but as Structured dialogue contributions. 
The extended COR model, therefore, defines basically 
two types of subnetworks: figure 2 displays the net of 
an INFORM contribution; figure 3 shows a representative 
net for all other types of contributions (here: REQUEST 
as an example). Thus, recursion is taken into account. 
In the figures 2 and 3 "A: request" and "A: inform" 
denote atomic acts, whereas a suffix notation indicates 
structured contributions, for example: ASSERT(A,B) or 
DIALOGUE(B,A, solicit context information). The traver- 
sal of the INFORM net is quite simple: A's INFORM act 
can be followed by a subdialogue initiated by B (e.g., by 
a REQUEST such as a clarifying question), or, if B does 
not need additional information, state < c > is reached 
immediately (jump). Thus, A's ("nuclear") INFORM act 
might be sufficient, whereas the ("satellitic") subdialogue 
is optional and depends on B's decision. 
Inform/Assert (A,B) jump 
dialogue (B,A, solicit 
context information) 
Figure 2: The INFORM net 
Request (A,B) dialogue (B,A, solicit context in formation) 
jump dialogue (B,A, identify request) 
Figure 3: The REQUEST net 
The transition net given in figure 3 is more complex but 
209 
7th International Generation Workshop • Kennebunkport, Maine • June 21-24, 1994 
follows the same principles. The subdialogues are also 
optional; A has two possibilities to start in state < a >: 
• A may start with a nuclear act (here, a REQUEST for 
information, such as a query to the database)and has 
then the opportunity to supply additional context 
information (e.g., an assertion to explain, specify or 
illustrate the request). If A does not give this context 
information voluntarily, B may enter a subdialogue 
to solicit the required context information. Thus, the 
ASSERT transition and the DIALOGUE leading from 
state < b > to < c > have a similar function, both 
being optional, i.e., satellitic. 
• A may start with an ASSERT (entering the subnet 
type displayed in figure 2) to give context informa- 
tion concerning her request; she may add the explicit 
REQUEST immediately afterwards, or may skip the 
explicit utterance (jump), in case she believes that 
B is able to infer the intended request from the con- 
text. If B is not able to identify the request, B has 
the option to initiate a clarification DIALOGUE. 
The COR model focuses on the illocutionary aspects of 
the conversation and abstracts away from the specific 
propositional content of dialogue contributions. However, 
it has been recognized that COR in its first version was 
only a partial model which had to be further enhanced by 
addressing rhetorical and semantic aspects (cf. Maier and 
Sitter, 1992; Stein and Maier, 1993). This was recently 
verified by Fischer (1993) who used the COR model to 
analyze a corpus of real dialogues between humans (infor- 
mation seekers communicating with information brokers 
to prepare a database search). 
2.1.2 Rhetorical Structure Theory (I:tST) 
Among the theories for modeling discourse, the RST - 
Rhetorical Structure Theory (cf. Mann and Thompson, 
1987) is the theory most exploited for natural language 
processing, particularly for natural language generation. 
RST is a theory which describes the structure of written 
monologues. One of the most basic assumptions of RST 
is that coherence can be modeled by means of named 
relations which hold between adjacent text units. Such 
relations can be used to structure texts by iteratively ap- 
plying relations thereby composing complex text units 
out of smaller ones. 
Another assumption of RST is that, in general, a rela- 
tion imposes an asymmetric structure on the connected 
text units. For a given pair of related text units, the 
so-called nucleus corresponds to the unit which contains 
highly relevant information, while the satellile carries less 
significant information; the satellite can be either substi- 
tuted or left out without significantly changing the overall 
meaning of the discourse. 
Since RST relations have been specified in a semi-formal 
way, this theory was a good candidate for a computational 
specification of coherence and later for an implementation 
of text planning and text generation systems. 
Recent attempts have been made to use this theory 
also for modeling dialogues, in particular for modeling 
both the connections within and between various dialogue 
contributions which is in contrast to approaches which 
only use RST to model links within a dialogue contri- 
bution like, e.g., Moore and Paris (1993). Nearly all ap- 
proaches (see, for example, Fawcett and Davies, 1992, and 
Daradoumis, 1993) are in the area of human-computer 
interaction, where a generation component is responsible 
for the automatic production of system utterances. An- 
other approach, which was developed for the domain of 
information-seeking dialogues, is reported in Maier and 
Sitter (1992). The authors showed that in such dialogues 
a specific subset of relations, the so-called interpersonal 
relations (see Maier and Hovy (1991)), are used. This 
classification of relations is based on three types of mean- 
ing as distinguished in Halliday (1985): ideational, in- 
terpersonal and textual meaning. Ideational meaning is 
the representation of experience of the world. Interper- 
sonal meaning refers to what the speaker or writer does 
in order to address the goals of the recipient. Textual 
meaning, finally, relates pieces of discourse to the context 
and indicates how the discourse structure has to be inter- 
preted. Interpersonal relations, therefore, share the be- 
havior that they mainly address features of the discourse 
participants. Among these relations we find, for instance, 
JUSTIFICATION, where the satellite provides reasons why 
the speaker or the listener should carry out actions spec- 
ified in the nucleus, or EVALUATION, where the satellite 
presents a subjective account of the information given in 
the nucleus. 
2.2 Integration of COR and RST 
2.2.1 Corpus Analysis 
To find out how COR and RST can be integrated for 
modeling information-seeking dialogues, a corpus of dia- 
logue transcripts - obtained from Prof. Saracevic, Rut- 
gets University, New Jersey - was carefully analyzed. The 
transcribed dialogues were conversations between a per- 
son seeking information and an information broker spe- 
cialized in database search. 
The transcripts contained oral communication with fre- 
quent syntactical mistakes, incomplete and halfway refor- 
mulated sentences. Our approach was not to try to model 
these attributes of dialogues. Instead, the utterances were 
adapted to match written, error-free text. 
First, a COR analysis was carried out, resulting in a seg- 
mentation of the transcribed dialogues into acts, contri- 
butions and whole dialogue cycles. Then these dialogue 
elements were assigned an illocutionary point and a nu- 
cleus or satellite status. 
Taking these results, an RST analysis was performed. 
The existing segmentation into acts, contributions and 
dialogue cycles was used to create the text spans that 
make up the constituents - nuclei and satellites - of the 
RST analysis. In our analyses no major problems were 
encountered by applying RST to dialogues, even though 
RST was developed for monologues only. 
2.2.2 Relations in the basic COR model 
There is a relatively small number of typical lIST re- 
lations that connect pairs of dialogue acts of the basic 
COR model which does not incorporate the recursive 
structure for subdialogues. The most important ones are 
SOLUTIONHOOD, EVALUATION, EVALUATION* and BACK- 
GROUND. Note that EVALUATION* is a newly defined re- 
lation that inherits aspects of EVALUATION; it will be de- 
scribed later in this section. 
210 
7th International Generation Workshop • Kennebunkport, Maine • June 21-24, 1994 
How these relations apply to illocutionary acts as defined 
in the COlt model is shown in a sample dialogue and its 
RST analysis (figures 4 and 5). As can be seen there, in 
the first phase of the dialogue an act of type REQUEST (for 
information) gets acknowledged positively by a PROMISE 
(to search for information and to present what has been 
found). The nucleus of the EVALUATION relation holding 
between the two acts is the REQUEST. 
User 
System 
System 
User 
(REQUEST) How many EC-funded projects dealing 
with Recombination are there? 
(PROMISE) Let's see. (INFORM) 
There axe 15 projects. 
(BE-CONTENTED) Okay. 
Figure 4: A sample dialogue 
evalua~on ~i:i evaluation informal *valua~on inform" \[ 
\[ request promise ~ request promise ~! rcqu~t promise 
Figure 5: Development of a dialogue as a RST diagram 
In the second stage, the requested information is given. 
REQUEST and PROMISE become satellite of a new relation 
which has an INFORM act as its nucleus. The suitable 
relation here is SOLUTIONHOOD, since the INFORM carries 
the answer to the REQUEST. Finally, the appropriate- 
ness of the provided data is confirmed by an act of type 
BE-CONTENTED. The recursively determined nucleus of 
the whole dialogue turn is the INFORM act. This is in 
line with what can be expected for information retrieval 
dialogues: the presentation of information to be looked 
for is the most central part of the whole dialogue. Ini- 
tiative acts (REQUEST, OFFER) are also important, but 
they merely open the Structure span which is completed 
when the INFORM act i S given, or when it is REJECTed or 
WITHDRAWn. 
Several dialogue cycles which appear within one level of 
dialogue are usually connected by the BACKGROUND re- 
lation (not shown here, see Fischer (1993) for examples). 
2.2.3 Expectation and Nuclearity 
In the COR model, roles - or rather expectations of spe- 
cific role behavior - are essential. Some acts are expected, 
while others are not; the latter ones are called alternative. 
The following describes \]how the expectation of certain di- 
alogue acts influence their status as a constituent within 
an RST relation, i.e., whether they are considered the 
nucleus or the satellite of the relation. 
On the dialogue level (see figure 1), the acts REQUEST, 
OFFER, and INFORM are most important. They actually 
contribute to the progression of the dialogue insofar as 
they actively model the negotiation of information. Other 
acts are merely evaluations of these three acts, they can 
either be positive or negative. The positive ones include 
ACCEPT, PROMISE, an d BE-CONTENTED. The negative 
ones are WITHDRAW, REJECT ' and BE-DISCONTENTED. 
The acts ACCEPT, PROMISE, and BE-CONTENTED are all 
expected ones. In our corpus, between any of these acts 
and their respective preceding acts the (positive) EVAL- 
UATION relation holds. Since the EVALUATION relation 
defines the constituent that contains the evaluating ex- 
pression as the satellite of the relation and the evaluated 
expression as the nucleus, this behavior matches features 
of the COR model: The evaluated dialogue acts are RE- 
QUEST, OFFER, Or INFORM which make significant propo- 
sitional contributions to the dialogue. For example, in the 
REQUEST-PROMISE pair of dialogue acts, REQUEST is the 
nucleus and PROMISE the satellite - the latter one being 
the expected positive acknowledgement. 
Evaluations that are not expected are WITHDRAW, RE- 
JECT, and BE-DISCONTENTED. They give the dialogue 
an alternative turn. This means that the evaluation is 
of high relevance and overrides the importance of previ- 
ous acts. In order to model this fact by means of RST, 
an alternative evaluation relation has to be introduced, 
which defines the evaluating expression to be the nucleus 
of the relation, in contrast to its definition as given above. 
We called this relation EVALUATION* because it resem- 
bles EVALUATION, except that it swaps the roles of the 
involved constituents. 
2.2.4 Context relations for the extended COIl, 
model 
The extended COR model contains complex dialogue con- 
tributions with a recursive structure (see figures 2 and 3). 
These complex acts have two constituents. One of these 
constituents expresses the illocutionary point of the whole 
contribution. Therefore, it takes on the role of the nu- 
cleus. Examples are: A: INFORM, A: REQUEST, A: OF- 
FER. The other constituent is either an ASSERT contribu- 
tion (atomic or complex) or a DIALOGUE to negotiate the 
contextual information (see also section 2.1.1). 
The fact that ASSERT and DIALOGUE serve the same pur- 
pose in COlt - namely to provide context information 
- also had to be modeled adequately in terms of ltST. 
This was achieved by simply assigning the INFORM act 
to be the nucleus of the whole DIALOGUE. Both ASSERT 
and INFORM may contain the same proposition (contex- 
tual information). The only difference is the way to get 
to this state: either A gives the information voluntarily 
ASSERT) or B initiates a sub-dialogue to ask for the in- 
rmation (DIALOGUE). 
Concerning the question about the types of relations typ- 
ically holding between the nuclear act - carrying the illo- 
cutionary point of the compound contribution (REQUEST, 
for example) - and the accompanying satellitic (assertive) 
act, our analyses resulted in finding two distinct types: 
(1) The first type is for additional information that is 
needed in order to answer a question or to understand 
a certain statement. In our genre of dialogues, this in- 
formation is obtained by applying Information Retrieval Tactics. (2) 
The second type is for supplementary infor- 
mation that explains the underlying reasoning behind an 
act made by a dialogue participant. We call this infor- 
mation type meta-information, The two relation types 
outlined above are now described in more detail. 
Context relation type 1: Information Retrieval 
Tactics In information-seeking dialogues, it is very un- 
common that questions can be answered immediately. In 
most cases additional information is necessary. One way 
211 
7th International Generation Workshop • Kennebunkport, Maine • June 21-24, 1994 
to do this is to use Information Retrieval Tactics. Their 
purpose is the procurement of data that is needed to suc- 
cessfully answer a request. In the transcribed dialogues, 
the tactics were employed by the information broker. An 
intelligent information retrieval system should also be 
able to handle at least some of these tactics. Tactics in- 
elude replacement or addition of search terms, pursuing 
or neglecting search paths; for a detailed collection see 
Bates (1979). 
The short dialogue shown in figure 6 exemplifies the use of 
the tactic SUPERTERM, which is used to replace a specific 
term by a more general one. The system cannot give an 
appropriate answer to the original request for all known 
EC-funded projects involving "NL Generation" because 
that search item is not contained in the database. Inor- 
der to get some relevant entities, it tries to extend the 
search, by replacing the original search term by a supert- 
erm. Therefore, it initiates a subdialogue asking the user 
to provide another term, and the user chooses "Natural 
Language". 
User (REQUEST) Show me all EC-funded projects dealing 
with NL Generation. 
Subdialogue to solicit contextual information, to 
System (REQUEST) Cart you tell me a more general 
term than NL Generation? 
Subdialogue to solicit contextual information 
User - (REQUEST) Why? 
System (INFORM) I ask you because NL 
Generation is unknown. 
User (BE-CONTENTED) I see. 
User (INFORM) How about Natural Language. 
System (BE-CONTENTED) OK. 
System (INFORM) Here is what I found. 
(System presents table o\] results.) 
System (OFFER) Would you like to see projects of a 
particular company? 
User (ACCEPT) Yes .... 
Figure 6: A dialogue example 
Information retrieval tactics can be integrated into RST 
without significant problems. They are subsumed by 
existing RST relations, such as GENERAL-SPECIFIC and 
ABSTRACT-INSTANCE. We have observed that most com- 
mon tactics used in Information Retrieval are in fact 
forms of (the very broad) ELABORATION relation. A more 
exhaustive analysis regarding the mapping of tactics to 
RST relations is required for a complete picture. 
Context relation type 2: Meta-Information 
Apart from the contextual information concerned with 
the contents of the database, there is a second type of 
context which deals with underlying reasoning behind ut- 
terances. Making these explicit means to make this un- 
derlying reasoning transparent to the dialogue partner. 
An information retrieval system can enhance confidence 
of the information seeker by giving this meta-information. 
There are several RST relations available for connecting 
assertive acts with meta-information to the nuclear act. 
Examples are CAUSE, PURPOSE and INSTRUMENT. 
The RST analysis of the sample dialogue given above is 
shown in figure 7. It shows examples for tactics (SU- 
PERTERM) and meta-information (CAUSE). Note that 
the context relations are connecting dialogue acts (RE- 
QUESTs in this example) with complete (sub-) dialogues, 
to be precise: the respective INFORM act within these sub- 
dialogues. Indeed, this consequently means that at some 
stage of the dialogue, no real relation holds between two 
adjacent dialogue acts, e.g., two consecutive REQUESTs. 
For the text generation processes, however, this does not 
involve a drawback. The COR model itself transposes the 
RST structure given for the whole sub-dialogues to single 
dialogue acts. 
3 The Corinna System 
In order to create an information system based on the 
integrated model of COR and RST, several aspects had 
to be considered which go beyond the previously outlined 
theoretical framework. 
The models COR and RST are of descriptive nature; they 
are not sufficient to actually construct a dialogue between 
user and system. To do this, a dialogue manager also 
needs further semantic information which constitutes the 
propositional content of the dialogue act to be generated. 
This semantic knowledge depends on and is provided by 
the genre (information seeking), the application area (in- 
formation retrieval) and the actual domain. 
As domain we chose the content of the CORDIS 
databases, which represent information about EC-funded 
research projects, participating organizations and strate- 
gic funding programs. The databases are offered online 
by ECHO, a database host of the Commission of the Eu- 
ropean Community. They have been transformed to and 
integrated into a relational database system which is pub- 
licly available. For the implementation of Corinna a way 
of incorporating this domain knowledge into the dialogue 
management and the generation of the dialogue acts had 
to be found. 
Natural language was decided to be the main mode of 
interaction in the Corinna system, because it achieves a 
high level of expressive power. The user is presented a 
menu of possible dialogue acts (according to the current 
dialogue state) among which she may choose the most 
appropriate one. The PENMAN generation system was 
used to produce both the system's and the user's dia- 
logue contributions. An additional modality of deictic 
gestures was introduced to allow convenient interaction 
with the system when the type of user contribution can- 
not be generated, for example when the basic search term 
and method is selected from a visual presentation. 
3.1 Discourse and Knowledge Structures 
3.1.1 Concepts and relations 
COR focuses on the dialogue pragmatics and omits se- 
mantic aspects in order to be domain-independent. As 
described above, RST can be used to introduce an ad- 
ditional level of coherence description between dialogue 
contributions. Additionally, some of the RST relations 
can be specialized in such a way that they represent rela- 
tions between concepts that constitute the propositional 
content of dialogue acts. 
The propositional level was modeled and implemented in 
a completely object-oriented way. Concepts of all lev- 
els of knowledge (dependent as well as independent from 
the domain) were classified as classes, instances and at- 
tributes. The object-oriented approach makes use of the 
212 
7th International Generation Workshop * Kennebunkport, Maine * June 21-24, 1994 
background 
f 
solu~onhood 
S: Here S: Would you like U: Yes,., 
is what to see projects of a I found, particular comF.ny? 
offer accept U: Show me all EC-fundcd proj- 
ects dealing with 
NL Generation. 
request 
Ir 
S: Can you roll me a 
rnom genera! term 
than NL Gmcration? 
request 
superterm 
evaluation inform 
solutionhood ~1 S: OK. ~'~l be contented 
cause U: How about 
~-~ Natural Language inform 
Lr .. 
U: I see. ~1 becontented 
U: Why? S: Because NL Gee- eration is unknown. 
request inform 
Figure 7: RST analysis 
notion of "relations". In the Corinna system we focused 
on the representation of relations on the meta-lev'el of the 
object-oriented paradigm. Examples include SUBCLkSS, 
CLASS-INSTANCE, and!INSTANCE-ATTRIBUTE. 
The ideational branch of Maier and Hovy's (1991) meta- 
functionally motivated taxonomy of RST relations made 
it particularly easy to augment the hierarchy, achieving 
a powerful model that uses relations as a unified way to 
structure knowledge. 
3.1.2 Developing the dialogue 
Within the process of information seeking, the situations 
of the search - specifically failure or partial success - direct 
the dialogue in certain:ways. Three main decision criteria 
have been incorporate~d in the Corinna system: 1) the 
requested data must be available in the database, 2) the 
data must be presentable in such a way that it is of use 
for the information seeker, and 3) the data has to be 
relevant with respect to the expressed information need. 
Depending on these factors, a follow-up dialogue act can 
be selected. For example, an OFFER is only possible when 
availability and presentability of the data can be granted. 
If the data is relevant, a follow-up ACCEPT is possible. If, 
on the other hand, the information need is not matched 
by the offered information, the OFFER will have to be 
answered (i.e., evaluated) by a REJECT-OFFER. 
Currently, the system is mainly able to judge availability 
and presentability of data. It is able to reason about 
the user's information need only in a very limited way. 
However, this does not lead to a significant drawback, 
since the user is always given the choice between several 
alternatives, as for example between an ACCEPT and a 
REJECT-OFFER. 
Taking into account what has been said above about the 
integration of COR and RST, the process for the gen- 
eration of new dialogue contributions also includes the 
of the sample dialogue 
treatment of RST relations. Given the propositional con- 
tent of a dialogue act, the determination of a follow-up 
act is influenced by the RST relation which holds be- 
tween the two acts. For example, if a relation of the 
type rneta-inforrna~ion (see section 2.2.4) has been cho- 
sen where contextual information has to be supplied, an 
appropriate proposition expressing the state or direction 
of the database search has to be composed. By this way 
sequences of coherent propositions for dialogue acts are 
produced. 
3.1.3 Access to the database 
Typically, an information retrieval system has its main 
challenge in the treatment of very large amounts of data. 
The CORDIS databases contain information about some 
14,000 projects and 10,000 organizations, persons and 
other entities. 
The object-oriented structure of Corinna enabled us 
to implement retrieval methods for several types of 
data storage - internally stored discourse knowledge and 
database contents represented externally - still having the 
benefit of a uniform programming interface. Thus, the 
dialogue manager is transparently accessing two distinct 
types of data, simplifying the design and implementation 
of the system significantly. 
3.2 Realizing Dialogue Acts 
So far the focus was on the underlying models of a dia- 
logue. This is motivated by the fact that the described 
mechanisms for performing a dialogue based on COR and 
RST apply to any modality, not necessarily natural lan- 
guage. For the Corinna prototype we chose the linguistic 
form as interaction mode because it incorporates a high 
level of expressive power needed for Information Retrieval 
tasks. Also, it has significant advantages over other in- 
teraction modes, as far as the clarification of misunder- 
213 
' 7th International Generation Workshop * Kennebunkport, Maine • June 21-24, 1994 
standing and dialogue failure are concerned. Information 
types required in such dialogue states cannot be presented 
raphically - language is the preferable interaction mode 
f. also Dilley et al., 1992). 
The realization of dialogue contributions is divided into 
two major steps. First, the internal representation of a di- 
alogue act is transferred into a high-level logical form that 
is capable of dealing with language-oriented attributes of 
an utterance, out of which English text is then automat- 
ically generated in the second step. For this task we 
used the PENMAN generator (PENMAN (1989))with 
the Sentence Plan Language SPL (Kasper and Whitney 
(1989)) as intermediate representation. 
3.2.1 Sources of Data for Realizing Acts 
We want to focus here on three main parameters influ- 
encing the process of realizing a dialogue act: illocntion, 
proposition and RST relation. 
It has been pointed out earlier that the propositional con- 
tent and illocutionary point of dialogue acts are managed 
independently. During the design of Corinna, it became 
evident that also the realization process can take place 
quite independently, as far as the creation of the SPL plan 
is concerned. The flexibility needed for natural language 
utterances is achieved by combining various intermedi- 
ate logical repreaentations of realizations for propositions 
and illocutions. Several typical linguistic attributes for 
the illocutionary acts in the COlt model can be assumed 
in order to allow management in the context of a lim- 
ited computer program. For example, many forms of rtE- 
QUEST incorporate the use of imperative forms like "tell 
me", "show me", etc. 
In cases where contextual information is to be given for an 
assertive dialogue act the ltST relation can make signif- 
icant contributions towards a coherent realization of the 
two utterances involved. For example, relations like PUrr- 
POSE or INSTRUMENT can be signaled by markers such as 
"in order to", "for", etc. 
Taking together 1) the fragmentary realization plan of the 
proposition on a purely assertive level, 2) the generated 
fragments for an illocutionary act of COlt, and 3) the 
RST relation that connects this act to the previous act, a 
suitable SPL statement for PENMAN can be composed. 
The final result is an English sentence, which is passed to 
the user interface. 
3.2.2 An example 
The following simplified example illustrates the various 
steps !n the planning of a contribution. We assume the 
scenario as given in figure 6 where the system has just 
asked for a SUPERTERM of "NL Generation". In this ex- 
ample the user wants to know the reason of the request 
and the system gives the corresponding answer. The SPL 
statement for this utterance is given in figure 8. Three 
distinct types of information were used to compute this 
sentence plan: (1) The RST relation CAUSE was selected, 
since the user asked the question "Why? . (2) The do- 
main of this relation is the verbal action that was per- 
formed by the system previously in the dialogue: The 
REQUEST for the more general term. (3) The motivation 
for making the utterance is the situation in the database 
- the search term is unknown. This fact is represented 
by a propositional item describing the state of the search. 
Note that the system's response is a dialogue act of type 
INFORM. As it has been stated earlier, the acts of type 
INFORM become the nucleus of the whole dialogue, which 
means that the (sub)dialogue takes on the role of an as- 
sertion after its completion. In case of a subdialogue, this 
assertion then carries the accompanying contextual infor- 
mation for another act one dialogue level up - in our case 
this is the REQUEST of the system, asking for a replace- 
ment term, and the contextual information is the CAUSE 
of that question. The RST relation plays then an impor- 
tant role in creating the coherence between the original 
question of the system and the contextual information, 
which is the reason, why it asked this question. 
(connecting-relation / rst-volitional-cause 
:domain ;; The system has asked for a 
;; superterm of "NL Generation" 
(act / addressee-oriented-verbal-process 
:lex ask 
:sayer speaker 
:addressee hearer) 
:range ;; The system informs the user 
;; about the reason why it asked. 
(proposition / property-ascription 
:tense present 
:domain (concept / object 
:name NL-Generation) 
:range (q / quality 
:lex unknoen))) 
Figure 8: Sentence plan for "I ask you, because NL Gen- 
eration is unknown" 
3.3 Interaction with the system 
It has been pointed out previously that we aimed at hav- 
ing a mainly language-driven interface. Additionally to 
the generation of the system's utterances we decided to 
also generate the dialogue contributions of the user. That 
is, there is no Natural Language Understanding module 
in Corinna; instead, in each dialogue situation where it is 
the turn of the user, Corinna generates several suggested 
utterances, from which the user can then choose the one 
that suits her best in the context of the current dialogue 
situation. We believe that this does not pose a severe 
limitation, the dialogue model based on COlt-RST, as 
well as the semantic knowledge about the search in the 
database allow predictions about how the dialogue might 
be continued. 
Figure 9 shows a sample interaction. The screen is di- 
vided into two halves. In the upper one, the instantiated 
dialogue history is displayed and updated after each con- 
tribution. The lower half of the interface is reserved for 
presentation of the database queries and their results. In 
the dialogue state as represented by the snapshot, the sys- 
tem has just provided the table of results. The user then 
has the choice to select one out of three possible acts: BE- 
CONTENTED ("OK"), CONTINUE ("This is not sufficient") 
and BE-DISCONTENTED ("Let's stop this dialogue"). 
Note that, depending on the situation of the dialogue, a 
participant - system or user - may utter two acts in a 
row, as opposed to simple system-user alternations. For 
example, after the user is contented with the answer of 
the system about why it asked for a term replacement, 
214 
7th International Generation Workshop • Kennebunkport, Maine • June 21-24, 1994 
CORINI#A - Cooperative Retrieval interface for Natural Language Acts 
User 
System 
User 
System 
User 
User 
System 
System 
Show me all projects with a title like NL GENERATION. 
Can you tell me a term that is more general than NL GENERATION? 
Why7 
I am asking you, because NL GENERATION Is unknown. 
Okay 
NATURAL LANGUAGE 
NATURAL LANGUAGE Is okay. 
The table contains all relevant entities. t User's choice 
OK 
This Is not sufficient. 
I do not llke ihat. Let's stop this dialogue. 
ACRONYM PROGRAM- ACRONY M REFERENCE 
ACQUILEX ESPRIT 2 3030 
8TART-DATE END-DATE 
Jun 1 1989 Mar 91 19922 
12:00:00:000AM 1 2:00:00:O00AM 
TITLE 
ACQUISITION OF 
LEXI CAL 
KNOWLEDGE FOR 
NATURAL LANGUAGE 
PROCESSI NG 
SYSTEMS 
ACORD ESPRIT 1 393 CONSTRUCTION Jan 1 198,5 Dec ~1 1989 
AND 12:00:O0:000AM 12:00:00:O00AM 
INTER~OGATION OF 
KNOWLEDGE BASES 
USING NATURAL 
LANGUAGE TEXT 
AND GRAPHI C~ 
DANDELION ESPRIT 3 DANDELION DISCOURSE Sep 1 1992 Aug 31 1995 
FUNCTIONS AND 12:00:00:000AM 12:00:00:O00AM 
DI SCOURSE 
REPRESENTATION: , Alkt CL~n"DUDI¢'~'AI I ~/ 
91~e,.1.oque Driver i 
IVed 9 Feb 6:14:43J F~.SC,I~Ler CL PE.~Y, hN: User Zn\]~.~.t 
Figure 9: A sample interaction with Corinna 
("Okay"), she then makes a contribution to the dialogue 
one level up - levels are represented by indentation of con- 
tributions - which was a term the system had asked for 
("Natural Language"). A similar phenomenon can be ob- 
served in the last two utterances of the system. To decide 
whether the user is allbwed to utter two acts in a row, the 
system applies a set of (relatively simple) heuristics; they 
are giving the user preference in guiding the dialogue. 
4 Summary 
The Corinna system iis the result of an integration of 
two models: The CQR model and RST. Compared to 
the single theories the integrated model has a number 
of significant advantages: (1) The scope of RST could 
be extended from monologues to (information seeking) 
dialogues - without changing major assumptions of the 
original theory; (2) RST adds an additional level of coher- 
ence to the existing COR model. While COP,. focuses on 
illocutions, dialogue acts and their sequential order, PaST 
provides means to describe the semantics of the links be- 
tween the single acts; (3) The description of dialogues in 
terms of COR-RST gives rise to a systematic approach to 
dynamically generate dialogue contributions for either of 
the participants. Corinna, as a prototypical implementa- 
tion of the combined theories, uses PENMAN to generate 
natural language utterances. It accesses a database typi- 
cal for the information retrieval task (CORDIS). 
Of course there are many possibilities for improvement. 
First, the flexibility and power of the generated utter- 
ances can be further increased. The plans for the dia- 
logue contributions could be further parameterized, also 
the knowledge base covering the current situation of the 
search in the database needs additional concepts. On the 
side of Information Retrieval, more tactics could be im- 
plemented making the search more efficient. 
We think that the outlined theoretical framework is pow- 
erful enough to go beyond the state of the system as it 
is implemented so far, and we will continue this research 
especially in the context of concrete application areas. 
Acknowledgements 
We would like to thank Stefan Sitter for support and in- 
spiring discussions. We also gratefully acknowledge the 
comments of three anonymous reviewers who made many 
helpful suggestions. Part of the work of the second au- 
thor was funded by the German Ministry for Research 
and Technology (BMFT) under contract 01 IV 101 k/1 (V RBMOBI0. 
215 
7th International Generation Workshop • Kennebunkport, Maine • June 21-24, 1994 

References 
\[Andr~ and PAst, 1993\] Elisabeth Andr~ and Thomas PAst. The 
design of illustrated documentation as a plvaming pask. In Mark 
T. Maybury (ed.), Intelligent Multimedia Interfaces. pages 94- 
116, AAAI Press/MIT Press, Menlo Park, California, 1993 

\[Areus et al., 1993\] Yigal Areas, F_xluard Hovy and Susanne van 
Mulken A tree traversing prototype that allocates presentation 
media. ISI, University of Southern California, 1993 

\[Bateman et aL, 1991\] John Bateman, Elisabeth Maler, Elke Teich 
and Leo Wanner. Towards an architecture for situated text gen- 
eration. In Proc. of the International Conference on Current 
Issues in Computational Linguistics, Penang, Malaysia, 1991 

\[Bates, 1979\] Marcia Bates. Information search tactics. Journal of 
the American Society of Information Science, 30:205-214, 1979 

\[Belkin et al., 1993\] Nicholas Belkin, Colleen Cool, Adell'teit Stein, 
• and Ulrich Thiel. Scripts for information-seeking strategies. In 
Case-Based Reasoning and Information Retrieval. Papers from 
the AAAI Spring Symposium Series. AAAI Technical Report 
SS-93-07, pages 8-17, 1993 

\[Belkin and Vickery, 1985\] Nicholas Belkln, and A. Vickery. Inter- 
action in information systems. The British Library, London, 1985 

\[Bilange, 1991\] Eric Bilange. A task independent oral dialogue 
model. In Proceedings of the Fifth Conference of the Euro- 
pean Chapter of the Association for Computational Linguistics 
(EA CL-91). pages 83-88, Berlin, Germany, April 1991 

\[Carenini and Moore, 1993\] Giuseppe Carenini and Joharma 
Moore. Generating explanations in context. In Proc. of the 1993 
International Workshop on Intelligent User Interfaces. pages 
175-192, ACM Press, New York, 1993 

\[Daradoumis, 1993~ Thanasis Daraxtoumis. Building a dynamic 
RST-based grammar of coherent interactive dialogues. In 
Michael Zock, Giovanni Adorni, and Giacomo Ferrari, editors, 
Proceedings of the Fourth European Workshop on Natural Lan. 
guage Generation. pages 91-101, Pisa, Italy, April 1993 

\[Dilley et al., 1992\] Stephan Dilley, John Bateman, Ulrich Thiel, 
Anne Titian Integrating natural language components into 
graphical discourse. In Proc. of the 3rd Conference on Applied Natural Language Processing. pages 72-79, ACL, Trento, Italy, 
1992 

\[Fawcett et al., 1988\] Robin P. Fawcett, Anita van der Mije, and 
Carla van Wissen. Towards a systemic flowchart model for dis- 
course structure. In Robin P. Fawcett and David J. Young, 
editors, New Developments in Systemic Linguistics; Volume 2: 
Theory and Application. pages 116-143. Pinter Publishers, Lon- 
don, 1988 

\[Fawcett and Davies, 1992\] Robin P. Fawcett and Bethan Davies. 
Monologue as a Turn in Interactive Discourse: Towards an Inte- 
gration of Exchange Structure and Rhetorical Structure Theory. 
In Proceedings of the 61h International Workshop on Natural 
Language Generation. Berllng, Springer, 1992 

\[Fischer, 1993\] Markus A. Fischer. Weiterentwicklung und Ira. 
plementierung eines Dialogsystems filr kooperative Information- 
ssysteme. GMD-Studien , No 228, Sankt Augustin: GMD, 1994 
(in German) 

\[Grosz and Sidner, 1986\] Barbara J. Grosz and Conduce L. Sidner. 
Attention, intentions and the structure of discourse. Journal of 
Computational Linguistics, 12(3):175-204, 1986 

\[Grosz and Sidner, 1990\] Barbara J. Grosz and Conduce L. Sidner. 
Plans for discourse. In Philip R. Cohen, Jerry Morgan, and 
Martha E. Pollack, editors, Intentions in Communication. pages 
417-444. MIT Press, Cambridge, MA, 1990 

\[Halliday, 1985\] M.A.K. Halliday. An introduction to functional grammar. Edward Arnold, 1985 

\[Kasper and Whitney, 1989\] Robert Kasper and Richard Whitney. 
SPL: A sentence plan language for text generation. Technical 
report, USC/ISI, 1989 

\[Lambert and Carberry, 1992\] Lynn Lambert and Sandra Car- 
berry. Using linguistic, world, and contextual knowledge in a 
plan recognition model of dialogue. In Proceedings of COLING- 
9~, Nantes, France. pages 310-316, 1992 

\[Litman and Allen, 1987\] Diane J. Litman and James F. Allen. A 
plan recognition model for subdialogues in conversations. Cog- 
nitive Science, 11:163-200, 1987 

\[Mater and Hovy, 1991\] Elisabeth Mater and Eduard Hovy. A metaftmctionally motivated taxonomy for discourse structure re- 
lations. In Proc. of the 3rd European Workshop on Natural Lan- 
guage Generation. Judenstein, Austria, March 1991 

\[Mater and Sitter, 1992\] Elisaheth A. Mulet and Stefan Sitter. An 
extension of rhetorical structure theory for the treatment of re- 
trieval dialogues. In Proc. of the l~th Annual Conference of 
The Cognitive Science Society. pages 968-973, Lawrence Erl- baum Associates, Hilisdale, N J, 1992 

\[Mann and Thompson, 1987\] William C. Mann and Semdra A. 
Thompson. Rhetorical Structure Theory: A theory of text orga- 
nization. In Livia Polanyi, editor, The Structure of Discourse. 
pages 85-96. Ablex Publishing Company, Norwood, N J, 1987 

\[Maybury, 1991\] Mark T. Maybury. Planning Multimedia Expla- nations Using Communicative Acts. In Proc. of the 9th National 
Conference on Artificial Intelligence (AAAr93). pages 61-66, 
AAAI Press/MIT Press, Anaheim, California, 1991 

\[Moore and Paris, 1989\] Johanna Moore and C6cile Paris. Plan- 
ning text for advisory dialogues. In Proc. of the ~Tth Annual 
Meeting of the Association of Computational Linguistics. pages 26--29, Vancouver, B.C., Canada, June 26-29, 1989 

\[Moore and Paris, 1993\] Johanna Moore and C~eile Paris. Plan- 
ning text for advisory dialogues. Capturing intentional and 
rhetoricalinformation. In Computational Linguistics, 19(4):651- 694, 1993 

\[Moore and Swartout, 1990\] Johanna Moore 
and William Swartout. Pointing: A way towards explanation 
dialogue. In Proc. of the 8th International Joint Conference on 
Artificial Intelligence (AAAr90), Vol. 1. pages 457-464, AAAI 
Press/MIT Press, Menlo Park, California, 1990 

\[O'Donnell, 1990\] Michael O'Donnell. A dynamic model of ex- 
change. Word, 41(3), December 1990 

\[PENMAN, 1989\] The PENMAN User Guide. Information Sci- 
ences Institute, Marina del Ray, California, 1989 

\[Reichman, 1985\] Rachel Reichman. Getting computers to talk like 
you and me. MIT Press, Cambridge, MA, 1985 

\[Ratter, Mellish and Levine, 1992\] Ehud Ratter, Chris Mellish and 
J. Levine. Automatic generation of on-line documentation in the 
IDAS project. In Proc. of the 3rd Conference on Applied Natural 
Language Processing. Trento, Italy, 1992 

\[Searle, 1969\] John R. Scoria. Speech acts. Cambridge University 
Press, UK, 1969 

\[Searle, 1979\] John R. Seaxle. A taxonomy of illocutionewy acts. In 
Expression and Meaning. Studies in the Theory of Speech Acts. 
Cambridge University Press, Cambridge, GB, 1979 

\[Sitter and Stein, 1992\] Stefan Sitter and Adelheit Stein. Modeling 
the illocutionary aspects of in.formation-seeking dialogues. Infor- 
mation Processing and Management, 28(2):165-180, 1992 

\[Smith, Hipp and Biermann, 1992\] R.W. Smith, D.R. Hipp, and 
A.W. Biermann. A dialogue control algorithm and its perfor- 
mance. In Proc. of the 3rd Conference on Applied Natural Lan- 
guage Processing Trento, Italy, 1992 

\[Stein and Maier, 1993\] Adelheit Stein, and Elisabeth Mater. Modeling and guiding cooperative multimodal dialogues. In 
Human-Computer Collaboration; Reconciling Theory, Synthe- 
sizing Practice. Papers from the AAAI Fall Symposium Series. 
AAAI Technical Report FS-93-05, pages 107-112, 1993 

\[Stein and Thiel, 1993\] Adelheit Stein and Ulrich Thiel. A conver- 
sational model of multimodal interaction in information systems. 
In Proceedings of the llth National Conference on Artificial In- 
telligence (AAAI'93). pages 283-288. AAAI Press/MIT Press, 
Menlo Park, CA, 1993 

\[Stein et al., 1992\] Adelheit Stein, Ulrich Thiel, and Anne Tiflen. 
Knowledge based control of visual dialogues. In Proceedings of 
the Ist International Workshop on Advanced Visual Interfaces 
(AVI-9$). pages 138-155. World Scientific Press, Singapore, 1992 

\[Winograd and Flores, 1986\] Terry Winograd and Fernando Flo- 
res. Understanding computers and cognition. Ablex Publishing 
Company, Norwood, N J, 1986 
