THE GENERATION OF HIGH-LEVEL STRUCTURE 
FOR EXTENDED EXPLANATIONS 
David J. Mooney (mooney@dewey.udel.edu) 
Sandra Carberry (carberry@dewey.udel.edu) 
Kathleen F. McCoy (mccoy~dewey.udel.edu) 
Department of Computer and Information Sciences 
University of Delaware Newark, Delaware 19716 
ABSTRACT 2 
This paper analyzes the structural features of 
naturally-occurring extended explanations and ar- 
gues that current generation methodologies are in- 
adequate for determining high-level structure. It 
presents a computational model based on the hy- 
pothesis that high-level structure - composed of a 
unifying framework and its associated basic blocks - 
can be determined by bottom-up processes that at- 
tempt to satisfy speaker, listener, and compositional 
goals, after which top-down strategies can be used to 
organize the material about the selected framework. 
1 INTRODUCTION 
In this paper, we describe the structural character- 
istics of extended, planned 1 explanations involving 
complex physical devices and present a computa- 
tional model for generating such explanations. Our 
investigation suggests that the organizational strate- 
gies currently employed for structuring short expla- 
nations are inadequate for generating the high-level 
structure characteristic of that found in naturally- 
occurring extended explanations, which typically re- 
quire several pages of text. Our computational model 
is based on the hypothesis that text structure is not 
completely recursive as others have claimed (\[GS86\], 
\[Rei78\], \[Po186\], \[MT88\]), but rather that the high- 
level structure of extended explanations is deter-. 
mined by processes separate from those which or- 
ganize text at lower levels. 
Section 2.1 provides a brief overview of current 
models for structuring text, followed by a description 
of the basic block, the unit of discourse on which our 
model is based, in Section 2.2. Section 2.3 describes 
the characteristics of high-level structure of extended 
explanations, followed by a description of our strat- 
egy for generating this structure in Section 3; a com- 
plete description is contained in \[MCM89\]. 
1Emphasis is placed on planned to distinguish these ex- 
planations from discourse in which the material is developed 
mutually by the participants as the discourse progresses. 
276 
THE BASIC BLOCK MODEL OF 
EXTENDED DISCOURSE 
ill order to generate an extended explanation, a nat- 
ural language system must determine tile basic con- 
tent to be conveyed; the next step is to cohesively 
organize this material. As anyone who has had to 
organize large amounts of information into a coher- 
ent text can attest, there are many possible combi- 
nations of that material, some more cohesive than 
others. Frequently, deciding how to orgauize a large 
body of material is more difficult than determining 
what to include. Our research is concerned with the 
identification of a coherent unifying framework about 
which an extended explanation can be organized and 
the criteria for selecting from among several frame- 
works when more than one viable alternative exists. 
2.1 OTHER APPROACHES 
TEXT STRUCTURE 
TO 
2.1.1 COMPUTATIONAL 
APPROACHES 
A number of researchers (e.g., \[GSS6\], \[MT88\], 
\[Rei7S\], \[PolS6\]) have argued that discom-se is con> 
posed of hierarchically structured segments and that 
this structure is completely recursive in nature. 
Two general methodologies have been applied to 
the structuring of explanations: schemas (\[McK85\], 
\[McC85\], \[earS7\]) and rhetorical structure theory 
(RST) (\[HMS9\], \[HovS8\], \[MPS8\], \[MSSS\]). 
A schema is a discourse strategy that captures a 
typical pattern of discourse associated with a partic- 
ular discourse purpose, (e.g., providing an analogy 
or evidence). Schemas can be thought of as tem- 
plates composed of an ordered sequence of rhetorical 
predicates, which 
"characterize the predicating acts a speaker 
may use and delineate tile structural rela- 
tion between propositions in a text." ~' 
2From \[McK85\] page 9 
These predicates are intended to capture the struc- 
tural relations that hold between clauses in a text. 
The predicates are used recursively, capturing the 
structure of text at any level. 
RST, developed by Mann and Thompson 
(\[MT88\]), was originally a tool for the analysis of 
text. RST claims that, except for a small number 
of highly-stylized forms, all coherent texts have an 
RST ,decomposition. RST posits a small number of 
relations, comparable to McKeown's rhetorical pred- 
icates, that exist between segments of text. Because 
each relation has associated with it well-defined in- 
tended effects and conditions necessary for it to hold, 
RST lends itself well to a generation methodology 
based on a top-down, hierarchical planning formal- 
ism (\[Sac77\]). Thus, like MeKeown's rhetorical pred-- 
icates, RST claims to account for the structure of 
text at any level of the discourse hierarchy. 
While these methods have proven to be effective 
for organizing short pieces of text, we maintain that 
they are inadequate for generating the characteristic 
structure of extended explanations at the level of the 
primary segments, which occupy the first level of the 
discourse hierarchy. We contend that the character- 
istics exhibited by the primary segments of extended 
explanations, to be described in the next section, 
cannot be captured by recursive processes. Rather, 
we maintain that high-level structure must be gener-- 
ated 'by a separate, bottom-up process, after which 
recursive organizational strategies can be applied at 
lower levels. 
2.1.2 RHETORIC 
Rhetoric, the formal study of the art of good writ- 
ing, provides general strategies for organizing text 
at a high level that are absent from the computa- 
tional models. Analysis - "the method of explana- 
tion whereby a subject is divided into its separate 
component parts ''3 - is possibly the most instru- 
mental of these strategies. There are no hard-and- 
fast rules for determining what constitutes an ap- 
propriate analysis of a subject. As \[WA60\] observes, 
a subject may be classified in as many ways as it 
has characteristics/parts/stages/etc. However, there 
are three criteria which experts (\[WA60\], \[Are75\], 
\[Tho57\], \[Dan67\], \[KP66\]) mutually consider essen- 
tial for a satisfactory organizational strategy: 
1. The 
tent 
(e.g. 
scheme should be logical; a single, consis- 
criterion should be used for tile analysis 
time, steps in a process). 
2. The scheme should exhaust all of the possibil- 
ities; everything to be conveyed should be en- 
compassed by the scheme. 
3. The resultant categories should be mutually ex- 
clusive; nothing should belong to more than one. 
3\[Are75\] page 107 
While the type of explanation with which this pa- 
per is concerned exhibits a high-level organization 
reflective of these criteria, the criteria by themselves 
do not provide the specificity necessary for computa- 
tional generation. These guidelines include no sug- 
gestions for dealing with situations in which no logi- 
cal, all-inclusive framework Call bc identified, nor do 
they offer suggestions for selecting among several or- 
ganizational schemes which meet the prescribed cri- 
teria equally well. Furthermore, the guidelines are 
not sufficient in-and-of themselves to account for all 
of the observed phenomena discussed in the following 
sections. 
2.2 BASIC BLOCKS 
Our model is based on a discourse unit which we 
have termed a basic block. A basic block consists of 
two elements: 
1. an organizational focus, such as a person or lo- 
cation, and 
2. a set of concepts related to that focus. 
The focus is what makes a cohesive unit of the ma- 
terial in the block; it is the thread cominon to all of 
this material, whether directly or indirectly. 
A basic block will be realized as a primary segment 
of text which occupies the first level of the discourse 
hierarchy. In a coherent discourse, the loci on which 
the basic blocks are based are themselves related, 
each representing a different aspect of some unifying 
framework. These points are demonstrated by the 
testimony fi'om which the basic block in Figure 1 
was extracted 4. 
This block references a particular time frame: zero 
to thirty seconds of the accident at Three Mile 
Island 5. The remaining blocks of that testimony are 
similarly constructed around time frames, e.g., one to 
six minutes, six minutes to one hour, etc. Observed 
frameworks demonstrate a gamut of types: proper- 
ties of the concepts (location, time), planning strate- 
gies in which events are involved (medical diagnosis), 
and characteristics that are not only inherent in the 
material but also due in part to the speaker's per- 
ception of them (significant factors). There appears 
to be no limit to what can constitute an acceptable 
framework, only that it is derived from the material 
itself and not from an independent device solely con- 
cerned with text structure. What may be a potential 
framework for one set of material may be totally in- 
adequate for another. Note that these features are 
reflective of the guidelines suggested by analysis. 
In addition to forming a cohesive unit, basic block 
structure is explicitly distinguished in the following 
two ways. First, it is often explicitly marked. In 
4 Space limitations prevent inclusion of the complete text. 
5Three Mile Island is a nuclear power plant located in the 
state of Pennsylvmfia in the United States. It suffered a near- 
meltdown in 1979. 
277 
Now, what happened at Three Mile Island was that 
a feedwater transient was experienced, and by that 
I mean, simply, we lost feedwater to the plant mo- 
mentarily. 
Now, wittl loss of feedwater to the steam generator, 
the plant will experience a momentary pressuriza- 
tion above its normal pressure. This pressurization 
is released by a relief valve located at the top of the 
pressurizer. It simply opens and lets out a little bit 
of stean't to take care of the excess pressure that is - 
interruption- 
Then at 15 seconds into the event-keeping in mind 
that the valves opened maybe 5 seconds into the 
event-at 15 seconds the pressure started coming 
down because the valve had opened and cut off the 
pressure. 
The valve should have reclosed when it got back 
down to about 2,250 pounds; it did not reclose. The 
pressure proceeded to come on down. At about 30 
seconds into the event, this water here started disap- 
pearing, of course, because now you are continuing 
to remove a very large amount of heat here, which is 
then coming off as secondary side steam generation, 
and this water will proceed to disappear if you do 
not start replacing it. 
And the auxiliary feedwater which normally comes 
on to make sure this does not go dry came on at 
about 30 seconds into the event. And at least the 
pumps were running. So this picture here is just the 
first 30 ,~econds and-or thereabouts. 
And this figure here is indicative of the situation from 
1 to 6 minutes into the event. 
Figure i: A Typical Primary Segment 
Figure 1, the speaker closes the block with explicit 
reference to its focus: "So this picture here is just 
the first 30 seconds and-or thereabouts." The sub- 
sequent block ii~ this testimony is also well marked, 
this time by its initial sentence (which is the last 
sentence of the figure): "And this figure here is in- 
dicative of the situation from 1 to 6 minutes." The 
basic block structure is further distinguished by cer- 
tain forms of repetition, whose use is closely tied to 
the basic block structure. While a brief allusion to 
these forms is made below, a detailed discussion is 
contained in \[MCM89\]. 
2.3 CHARACTERISTICS 
OF COHERENT, HIGH-LEVEL 
STRUCTURE 
Given tile existence of a high-level structure as evi- 
denced by explicit markings and repetition, we must 
consider how such a framework is chosen. If the only 
consideration were the identification of a set of re- 
lated foci which can partition the material to be con- 
veyed, then any set of related concepts about which 
the material can be cohesively organized would suf- 
fice. Consider the motivation behind the block of 
Figure 1. On the surface, the events in this block 
all occurred within the first thirty seconds of the ac- 
cident. However, it is doubtful whether the driving 
force behind the construction of this block was to 
communicate the time frame in which these events 
occurred; rather, what is of importance is their rela- 
tive sequence in the total series of events, their cause- 
effect relations, and their impact on the resulting ac- 
cident. One could argue that the individual events in 
this block represent a cause-effect chain, and hence 
their mutual grouping; but this chain is continued 
in the next block of the testimony. Thus, this ar- 
gument alone cannot account for this segmentation. 
Apparently, other factors beyond the ability to cohe- 
sively juxtapose clauses contribute to the segmenta- 
tion and the high-level framework about which it is 
constructed. 
In all of the dialogues examined, the blocks are of 
approximately the same size. Apparently, speakers 
choose organizations that tend to be well-balanced. 
However, balance does not seem to be the only crite- 
rion taken into consideration; if it were, we could ex- 
pect to see perfectly balanced organizations in which 
each of the concepts to be conveyed is realized by its 
own basic block in addition to those well-bahmced or- 
ganizations found in our text analysis whose blocks 
are comprised of many concepts. The size of the com- 
ponent basic blocks seems to be a further consider- 
ation in determining what constitutes an acceptable 
framework. No lengthy discourse organized about a 
single primary segment has been observed, nor have 
any in which primary segments are composed of sin- 
gle concepts. The "ideal" size of a basic block seems 
to be three or four paragraphs in length for an ex- 
planation of three pages. 
We contend that the high-level structure of ex- 
tended explanations reflects the characteristics of an 
"ideal" framework in which: 
The basic block loci about which the material is 
organized reflect various aspects of the unifying 
framework. 
• Basic blocks are the same size. 
• The length of each basic block approximates the 
ideal size. 
Our basic block model of discourse posits that the 
attainment of each of these features, termed compo- 
sitional goals, is instrumental in the selection of a 
high-level framework. 
However, not all observed explanations conform to 
the basic block model as presented thus far. For ex- 
ample, in explanations that are concluded by a final 
278 
summary, one would expect tile summary to empha- 
size the major points stressed by the speaker in the 
body of the text. In fact, there is a strong correla- 
tion between the segments comprising such a sum- 
mary and the foci about which the basic blocks are 
constructed (see \[MCM89\]). Ilowever, some expla- 
nations have been examined in which material that 
appears in a summary has not been included in the 
text. Another apparent exception has been observed 
in explanations that exhibit a well-defined high-level 
structure that accounts for all of the basic blocks ex- 
cept for the final one, which is at best loosely related 
to the others. 
We postulate that a speaker attempts to identify 
a framework that is capable of coherently expressing 
all of the material he wishes to convey while satisfy- 
ing the compositional goals equally well. However, a 
speaker is rarely blessed with such an ideal situation. 
The problem appears to be one of finding a satisfac- 
tory, rather than an optimal, unifying framework. 
We hypothesize that the satisfaction of some goals 
will be sacriffced ill favor of others so that a frame- 
work that provides the best overall solution can be 
achieved. This relaxation of goal constraints explains 
the above anomalies and the variability of size and 
balance observed in basic block structure. 
3 A MODEL FOR GENERATION 
We have developed a computational ntodel of dis- 
course generation that captures the structural char- 
acteristics observed in naturally-occm:ring explana- 
tions. Our model is based on the hypothesis that 
the high-level structure of a discourse can be de- 
termined by bottom-up processes that attempt to 
satisfy speaker, listener, and compositional goals. 
Once this organization has been established, top- 
down processes are used to organize the information 
into basic blocks and to supplenlent that information 
based on the choice of framework. Only after the ba- 
sic block structure has been established will detailed 
organization within a block and realization into text 
proceed. The remainder of this section describes our 
strategy for identifying an organizational framework 
and the resulting basic blocks given an initial set of 
concepts to be conveyed. We conclude by establish- 
ing our model within the context of a complete text 
generating system. 
3.1 IDENTIFICATION OF CANDI- 
DATE FRAMEWORKS 
We suggest that a speaker, when organizing nil ex- 
tended explanation, will be faced with one of the 
following general situations: 
®IIe already has a well-defined organizational 
structure in which the material to be conveyed 
has already been organized. 
• lie has a set of goals about which he wants to 
structure the discourse. In this situation, the 
goals serve as basic block loci about which he 
must now attempt to partition his material. 
* The speaker is starting "from scratch"; he needs 
to find a unifying framework about which he can 
structure what he wants to say. It is this situa- 
tion which our strategy intends to capture. 
Our strategy assumes that the generation process 
starts with some initial state of affairs and a com- 
municative goal to be achieved. Based on this, a 
rough set of information to be conveyed will be iden-. 
tiffed. The selection of this material is not our con- 
cern; rather, we are concerned with determining a 
reasonable high-level structure for this information. 
We assume that the material to be conveyed has been 
categorized into at least two levels of importance: 
that which must be included at all costs (tile ker- 
nel), and that which would be nice to include, but 
due to time, style, or coherence may be left out. The 
process of identifying candidate frameworks begins 
with the kernel. 
The implication of the discussion in Section 2 is 
that a unifying framework will be some feature to 
which all of the kernel concepts can be related. In 
our model, we a.ssmne that the domain knowledge is 
maintained as a hierarchical network. Such a repre- 
sentation suggests that a node to which all of these 
concepts converge could serve as the unifying frame- 
work. Tile basic block loci, which represent various 
aspects of this framework, would then correspond to 
children of this node. 
Given a set of propositions to be conveyed, one 
is randonfly selected from the kernel and a trace 
is performed upwards from each of its arguments 
through the hierarchy. The traversal of the hierarchy 
is pertbrmed using generalization links, e.g., ISA, IS- 
PART, and SUBG()AL, incrementing counters asso- 
ciated with each node that is traversed. Additionally, 
CAUSE/EFFECT links can be followed when events 
are encountered. Property links, such as color and 
size, are used only if they were included as part of 
the kernel. Once the traversal has been completed 
for each element of the kernel, the nodes are ranked 
according to their counters; this will give some idea 
of tile number of concepts that converge on each. 
At this point, the general characteristics of basic 
blocks can be used to eliminate some obvious non- 
candidate frameworks. Since no extended explana- 
tions have been observed to be composed of a single, 
massive basic block, a node that has the same value 
as its parent can be eliminated from consideration 
as a candidate framework. By the same token, since 
no organization has been observed in which a ba- 
sic block is constructed for each concept, any node 
that has a number of children on the same order 
as tlle number of concepts in the kernel carl also be 
eliminated from consideration. Furthermore, nodes 
279 
at which very few concepts converge can be disre- 
garded since an acceptable framework must account 
for most (ideally all) of the kernel. 
The remaining nodes represent the potential can- 
didates for the unifying framework. In all likelihood, 
none of these candidates will account for all of the 
kernel concepts. However, it may be possible to con- 
nect such concepts to a candidate by finding a link 
between these concepts and ones which are already 
associated with a candidate, possibly via a property 
they have in common. Additionally, it may be possi- 
ble to include such material in a final summary or in 
a "catchall" block. That speakers do introduce such 
material in this way is evidenced by our transcript 
analysis. We claim that the inclusion of such mate- 
rial in a final summary is warranted if the speaker 
initially intended to include a final summary and if 
the amount of material is of the same order as that 
comprising the individual segments of the summary; 
for the creation of a catchall block to be viable, there 
must be sufficient material to construct an adequate 
basic block and the material must form a cohesive 
unit. 6 
3.2 SELECTION OF THE UNIFY- 
ING FRAMEWORK 
Once potential fl'ameworks have been identified, each 
must be evaluated according to how well it meets the 
criteria described in Section 2.3: 
• ttow thoroughly does a candidate account for 
the selected material? 
,, How uniformly does a candidate distribute the 
concepts among the resulting basic blocks? 
,, IIow closely do the generated blocks conform to 
the ideal size? 
Additionally, a candidate may be evaluated ~ to how 
well it meets the needs of a user model. 
The actual blocks are constructed around a can- 
didate's immediate descendents, not the candidate 
itself. So, for example, if a candidate were the node 
workstation.s, its children - actual workstations at 
which things occurred - will become the basic block 
fool. The ideal case is one in which each child of the 
candidate accounts for approximately the same num- 
ber of concepts. The balance each candidate achieves 
can be determined by comparing the counters of its 
children. At the same time, the candidate will be 
rated according to how closely its blocks conform to 
the ideal size. Lastly, we can rate the candidates on 
how well they meet the demands, if any, imposed by 
a user model. For example, if we know the user is fa- 
miliar with the location in which the events occurred, 
basic blocks based on location should be given higher 
ratings; organizing an explanation around a frame- 
work with which the listener is familiar will fadlitate 
his assimilation of the information. 
6See \[MCM89\] for a more complete discussion. 
3.3 GENERATION OF TEXT 
At this point, the high-level structure of the expla- 
nation has been determined. In the remainder of 
this paper, we will attempt to place our strategy in 
the context of a complete system by giving a brief 
overview of the remaining phases of our model. 
Once a fl'amework has been selected, the concepts 
to be conveyed can be partitioned about each of the 
resulting foci. At this point, there may be some con- 
cepts for which the framework does not account. The 
system must determine whether to include them as 
a separate, trailing basic block, as part of a final 
summary, or to eliminate them altogether (refer to 
Section 3.1). In addition, depending on several fac- 
tors (e.g., verbosity constraints, choice of framework, 
etc.), those concepts of moderate import may be inte- 
grated into the framework by performing a traversal 
of the networks or by finding links to concepts al- 
ready incorporated by the framework (as described 
in Sections 3.1 and 3.2). These steps complete the 
process of partitioning the material about the unify- 
ing framework. 
The next step is to determine the order of presen- 
tation for the blocks. The order in which the basic 
blocks are presented may be inlposed by the mate- 
rim itself or by some external factor. If no overriding 
considerations exist, the material may present obvi- 
ous "natural" choices, such as sequential ordering by 
time or by cause-effect. On the other hand, there 
may be a predetermined preference for a particular 
order, e.g., by their importance as determined by 
some external process. 
Once the blocks have been ordered, text generation 
can proceed. We intend to use an established, low- 
level text structuring strategy (e.g., RST or schemas) 
for this phase. We anticipate that additional infor- 
mation may be included in order to satisfy low-level 
discourse goals (e.g., supplying information requisite 
for the user's comprehension of the material to be 
presented); thus, we do not claim that the entire con- 
tents of a text have been determined before structur- 
ing is performed. Rather than organizing and realiz- 
ing text for the blocks en mass, we have adopted a 
strategy in which these steps are performed for each 
block individually. Text for one block is presented to 
the user before proceeding to the organization and 
generation of text for the next block. This is to pre- 
clude wasted effort which may result from reorgani- 
zation necessitated by a user's questions. 
4 CONCLUSIONS 
This paper has presented a computational model for 
generating the high-level structure for extended ex- 
planations. It is being implemented as part of DIALS 
(Delaware Intelligent Advisory Language System), 
using the NIKL knowledge representation system. 
No claims are made that the methodology represents 
a psychological model of composition. However, it 
280 
is maintained that the text produced by this model 
is typical of that encountered in many naturally-- 
occurring extended explanations. The model does 
not presume to generate all acceptable organizations; 
indeed, there are many welLstruetured explanations 
for which it does not account. We have presented 
the simplest ease in which a high-level organization 
is constructed from a parent and its children in a 
taxonomic hierarchy. However, many other possibil- 
ities exist, several of which are currently being ex- 
amined within the semantic network paradigm. In 
addition, we are investigating the generation of ap- 
propriate recovery strategies in response to interrup- 
tions and their impact on the planning strategy. We 
are also incorporating several forms of repetition that 
are widely used in extended explanations. 
The unit of discourse which is the major moti- 
vation tbr this model is the basic block, a primary 
segment of text which occupies the tirst level of the 
discourse hierarchy. It consists of an organizational 
focus and text constructed about that focus. Wilile 
methods such as RST and schemas are adequate for 
local organization of text, they are insufficient for the 
determination of high-level structure, providing nei- 
ther criteria which describe what constitutes an ac- 
ceptable framework for a discourse, nor any clues as 
to how this framework should be constructed. This 
paper provides both. We believe that our model 
provides strategies essential for bridging the gap be- 
tween the generation of short explanations and the 
successful generation of extended explanations. 

References 

L. A. Arena. Linguistics and Composition. 
Georgetown University Press, 1975. 

R,. W. Daniel. A Contemporary Rhetoric. 
Little, Brown and Company, 1967. 

B. J. Grosz and C. L. Sidner. Attention, 
intentions, and the structure of discourse. 
Computational Linguistics, 12, 1986. 

E. lIovy and K. McCoy. Focusing your 
rst: a step toward generating coherent 
maltisentential text. In Proceedings of the 
Eleventh Annual Meeting of the Cognitive 
Science Society, Cognitive Science Soci- 
ety, 1989. 

E. Hovy. Approaches to the planning of 
coherent text. Fourth International Work- 
shop on Natural Language Generation, 
i988. 

T. S. Kane and L. J. Peters. A Practi- 
cal Rhetoric of Expository Prose. Oxford 
University Press, 1966. 

K. F. McCoy. Correcting Object-Related 
Misconceptions. PhD thesis, University of 
Pennsylvania, 1985: 

K. MeKeown. Text Generation. Cam- 
bridge University Press, 1985. 

D. Mooney, M. S. Carberry, and K. Mc- 
Coy. 7'he Identification of a Unifying 
Framework for the Oryanization of Exo 
tended, Interactive Explanations. Techni- 
cal Report 90-1, University of Delaware, 
1989. 

J. Moore and C. Paris. Constructing co- 
herent text using rhetorical relations. In 
Proceedings of the Tenth Annual Meeting 
of the Cognitive Science Society, Cogni- 
tive Science Society, 1988. 

J. 13. Moore and W. R. Swartout. A re- 
active approach to explanation. Fourth 
International Workshop on Natural Lain 
guage Generation, 1988. 

W. C. Mann and S. A. Thompson. Rheto> 
ical structure theory: toward a functional 
theory of text organization. Text, 8, 1988. 

C. Paris. 7'he Use of Explicit User Models 
in Text Generation: Tailoring to a User's 
Level of Expertise. PhD thesis, Columbia 
University, 1987. 

L. Polanyi. The Linguistic Discourse 
Model: 7bwards a Yormal Theory of Dis° 
course Structure. Technical Report 6409, 
BBN, 1986. 

R. Reiehman. Conversational coherency. 
Cognitive Science, 2, 1978. 

E. Sacerdoti. A Structure for Plans and 
Behavior. Elsevier North-llolland, 1977. 

W. N. Thompson. Fundamentals of Como 
munication. McGraw ttill, 19.57. 

C. V. Wicker and W. P. Albrecht. The 
American Technical Writer. American 
Book Company, 1960. 
