<?xml version="1.0" standalone="yes"?>
<Paper uid="P88-1020">
  <Title>Rhetorical Structure Theory: Description and Construction of Text Structures, in Natural Language Generation: Nero Results in</Title>
  <Section position="3" start_page="0" end_page="163" type="metho">
    <SectionTitle>
1 The Problem of Coherence
</SectionTitle>
    <Paragraph position="0"> The example texts in this paper are generated by Penman, a systemic grammar-based generator with larger coverage than probably any other existing text generator. Penman was developed at ISI (see \[Mann &amp; Matthiessen 831, \[Mann 831, \[Matthiessen 84\]). The input to Penman is produced by PEA (Programming Enhancement Advisor; see \[Moore 87\]), a program that inspects a user's LISP program and suggests enhancements.</Paragraph>
    <Paragraph position="1"> PEA is being developed to interact with the user in order to answer his or her questions about the suggested enhancements. Its theoretical focus is the production of explanations over extended interactions in ways that are superior to the simple goal-tree traversal of systems such as TYRESIAS (\[Davis 76\]) and MYCIN (\[Shortliffe 76\]).</Paragraph>
    <Paragraph position="2"> Supported by DARPA contract MDAg03 81 C0~5.</Paragraph>
    <Paragraph position="3"> In answer to the question how does the system enhance a program~, the following text (not generated by Penman) is not satisfactory: (a). The system performs the enhancement. Before *hat, the system resolves conficts. First, the system asks the user to tell Jt the characteristic of the program to be enhanced. The system app//es transformations to the program.</Paragraph>
    <Paragraph position="4"> /t confrms the enhancement with the user. It scans the program in order to find opportunities to apply transfarmations to the program.</Paragraph>
    <Paragraph position="5"> ... because you have to work too hard to make sense of it. In contrast, using the same propositions (now rearranged and linked with appropriate connectives), paragraph (b) (generated by Penman) is far easier to understand: (b). The system as/ca ~he user to tell it the characteristic of the program to be enhanced. Then the system applies transformations to the program. In particular, the system scans the program in order to ~nd opportunities to apply transformations to the program. Then the system resolves contlicts. It con~rms the enhancement with the user. Fina//y, it performs the enhancement.</Paragraph>
    <Paragraph position="6"> Clearly, you do not get coherent text simply by stringing together sentences, even if they are related -- note especially the underlined text in (b) and its corresponding three propositions in (a). The goal of this paper is to describe a method of planning paragraphs to be coherent while avoiding unintended spurious effects that result from the juxtaposition of unrelated pieces of text.</Paragraph>
  </Section>
  <Section position="4" start_page="163" end_page="163" type="metho">
    <SectionTitle>
2 Text Structuring
</SectionTitle>
    <Paragraph position="0"> This planning work, which can be called tezt siructuring, must obviously be clone before the actual generating of language can begin. Text structuring is one of a number of pre-generation text planning tasks. For some of the other tasks Penman has special-purpose domain-specific solutions. They include: * aggregation: determining, for input elements, the appropriate level of detail (see \[Hovy 87\]), the scoping of sentences, and the use of connectives * reference: determining appropriate ways of referring to items (see \[Appelt 87a, 87b\]) * hypotheticals: determining the introduction, scope, and closing of hypothesis contexts (spans of text in which some values are assumed, as in air you want to go to the game, then ... ~) The problem of text coherence can be characterized in specific terms as follows. Assuming that input elements are sentence- or clause-sized chunks of representation, the permutation set of the input elements defines the space of possible paragraphs. A simplistic, brute-force way to achieve coherent text would be to search this space and pick out the coherent paragraphs. This search would be factorlally expensive. For example, in paragraph (b) above, the 7 input clusters received from PEA provide 7! ---- 5,040 candidate paragraphs. However, by utilizing the constraints imposed by coherence, one can formulate operators that guide the search and significantly limit the search to a manageable size. In the example, the operators described below produced only 3 candidate paragraphs. Then, from this set of remaining candidates, the best paragraph can be found by applying a relatively simple evaluation metric.</Paragraph>
    <Paragraph position="1"> The contention of this paper is that, exercising proper care, the coherence relations that hold between successive pieces of text can be formulated as the abovementioned search operators and used in a hierarchical-expanslon planner to limit the search and to produce structures describing the coherent paragraphs.</Paragraph>
    <Paragraph position="2"> The illustrate this contention, the Penman text structurer is a simplified top-down planner (as described first by \[Sacerdoti 77\]). It uses a formalized version of the relations of Rhetorical Structure Theory (see immediately below) as plans. Its output is one (or more) tree(s) that describe the structure(s) of coherent paragraphs built from the input elements. Input elements are the leaves of the tree(s); they are sent to the Penman generator . to be transformed into sentences.</Paragraph>
  </Section>
  <Section position="5" start_page="163" end_page="164" type="metho">
    <SectionTitle>
3 Previous Approaches
</SectionTitle>
    <Paragraph position="0"> The heart of the problem is obviously coherence.</Paragraph>
    <Paragraph position="1"> Coherent text can be defined as text in which the hearer knows how each part of the text relates to the whole; i.e., (a) the hearer knows why it is said, and (b) the hearer can relate the semantics of each part to a. single overarching framework.</Paragraph>
    <Paragraph position="2"> In 1978, Hobhs (\[Hobhs 78, 79, 82\]) recognized that in coherent text successive pieces of text are related in a specified set of ways. He produced a set of relations organised into four categories, which he postulated as the four types of phenomena that occur during conversation. His argument, unfortunately, contains a number of shortcomings; not only is the categorization not well-motivated, but the llst of relations is incomplete.</Paragraph>
    <Paragraph position="3"> In her thesis work, McKeown took a different approach (\[McKeown 82\]). She defined a set of relatively static schemas that represent the structure of stereotypical paragraphs for describing objects. In essence, these schemas are paragraph templates; coherence is enforced by the correct nesting and 6\]llng.in of templates. No explicit theory of coherence was offered.</Paragraph>
    <Paragraph position="4"> Mann and Thompson, after a wide-ranging study involving hundreds of paragraphs, proposed that a set of 20 relations suffice to represent the relations that hold within the texts that normally occur in English (\[Mann &amp; Thompson 87, 86, 83\]). These relations, called RST (rhetorical structure theory), are used recursively; the assumption (never explicitly stated) is that a paragraph is only coherent if all its parts can eventually be made to fit into one overarching relation. The enterprise was completely descriptive; no formal definition of the relations or justification for their completeness were given. However, the relations do include most of Hobbs's relations and support McKeown's schemas.</Paragraph>
    <Paragraph position="5"> A number of similar descriptions exist. The description of how parts of purposive text can relate goes back at least to Aristotle (\[Aristotle 54 D. Both Grimes and Shepherd categorize typical intersentential relations (\[(\]rimes 75\] and \[Shepherd 26\]). Hovy (\[Hovy 86\]) describes a program that uses some relations to slant text.</Paragraph>
  </Section>
  <Section position="6" start_page="164" end_page="164" type="metho">
    <SectionTitle>
4 Formalizing RST Relations
</SectionTitle>
    <Paragraph position="0"> As defined by Mann and Thompson, RST relations hold between two successive pieces of text (at the lowest level, between two clauses; at the highest level, between two parts that make up a paragraph} 1. Therefore, each relation has two parts, a aucle~ and a satell~te. To determine the applicability of the relation, each part has a set of constraints on the entities that can be related.</Paragraph>
    <Paragraph position="1"> Relations may also have requirements on the combination of the two parts. In addition, each relation has an effect field, which is intended to denote the conditions which the speaker is attempting to achieve.</Paragraph>
    <Paragraph position="2"> In formalizing these relations and using them generatively to plan paragraphs, rather than analytically to describe paragraph structure, a shift of focus is required. Relations must be seen as plans the operators that guide the search through the permutation space. The nucleus and satellite constraints become requirements that must be met by any piece of text before it can be used in the relation (i.e., before it can be coherently juxtaposed with the preceding text}. The effect field contains a description of the intended effect of the relation (i.e., the goal that the plan achieves, if properly executed}. Since the goals in generation are communicative, the intended effect must be seen as the inferences that the speaker is licensed to make about the bearer's knowledge after the successful completion of the relation.</Paragraph>
    <Paragraph position="3"> Since the relations are used as plans~ and since their satellite and nucleus constraints must be reformulated as subgoais to the structurer, these constraints are best represented in terms of the communicative intent of the speaker. That is, they are best represented in terms of what the hearer will know -- i.e., what inferences the hearer would run -- upon being told the nucleus or satellite filler.</Paragraph>
    <Paragraph position="4"> As it turns out, suitable terms for this purpose are provided by the formal theory of rational interaction currently being developed by, among others, Cohen, Levesque, and Perrault. For example, in ICohen ~z Levesque 851, Cohen and Levesque present a proof that the indirect speech act of requesting can be derived from the following bask  such as Seqtlence, relate more than two pieces of text. However, for ease of use, they have been implemented as binary relations in the structurer.</Paragraph>
    <Paragraph position="5">  events after action a as well as from a few other operators such as AND and OR. They then define suture,ties as, essentiaUy, speech act operators with activating conditious (g~tes) and e~ectz. These summaries closely resemble, in structure, the RST plans described here, with gates corresponding to satellite and nucleus constraints and effects to intended effects.</Paragraph>
  </Section>
  <Section position="7" start_page="164" end_page="166" type="metho">
    <SectionTitle>
5 An Example
</SectionTitle>
    <Paragraph position="0"> The RST relation Purpose expresses the relation between an action and its intended result:  1. (BMB S H (STATE ?state-l)) 2. (BMB S H (GOAL ?a~-I ?state-l)) s. (B~ S H (RESULT Zact-1 ?~t-2)) 4. (BMB S H (OBJ ?act-2 ?state-I)) Intended EEectss 1. (BMB S H (BEL ?ag~-I (RESULT ?act-1 ?state-l))) 2. (BMB S H (PURPOSE ?act-I ?state-l))  For example, when used to produce the sentence The system scans the program in order to find opportunltJes to apply ~ansformatlons to t~e program, this relation is instantiated as  The elements SCAN-l, OPP-1, etc., are part of a network provided to the Penman structurer by PEA. These elements are defined as propositions in a property-inheritance network of the usual kind written in NIKL (\[Schmolze &amp; Lipkis 83\], \[Kaczmarek et aL 86\]), a descendant of KL-ONE (\[Brachman 78\]). Some input for this example sentence is:  The relations are used as plans; their intended effects are interpreted as the goals they achieve. In other words, in order to bring about the state in which both speaker and hearer know that OPP-1 is the purpose of SCAN-I (and know that they both know it, etc.), the structurer uses Purpose as a plan and tries to satisfy its constraints.</Paragraph>
    <Paragraph position="1"> In this system, constraints and goals are interchangable; for example, in the event that (RESULT SCAN-I FIND-I) is believed not known by the hearer, satellite constraint 3 of the Purpose re= lation simply becomes the goal to achieve (BHB S H (RESULT SCAN-I FIND-I)). Similarly, the propositions (B~ S H (RESULT SCAN-1 ?ACT-2)) (BMB S H (0BJ ?ACT-2 0PP-I)) are interpreted as the goal to find some element that could legitimately take the place of ?ACT-2.</Paragraph>
    <Paragraph position="2"> In order to enable the relations to nest recursively, some relations' nucleuses and satellites contaln requirements that specify additional relations, such as examples, contrasts, etc. Of course, these additional requirements may only be included ff such material can coherently follow the content of the nucleus or satellite. The question of ordering such additional constituents is still under investigation. The question of whether such additional material should be included at all is not addressed; the structure,&amp;quot; tries to say everything it is given. The structurer produces all coherent paragraphs (that is, coherent as defined by the relations) that satisfy the given goal(s) for any set of input elements. For example, paragraph (b) is produced to satiny the initial goal (BMB S e (SEQUENCE ASK-1 ?l~E~r)). This goal is produced by PEA, together with the appropriate representation elements (ASK-1. SCAM-I, etc.) in response to the question hoto a~oes ~e system enhance a progr~m~. Di~erent initial goals will result in di~erent parsgraphs. null Each paragraph is represented as a tree in which branch points are RST relations and leaves are input elements. Figure 1 is the tree for paragraph (b). It cont~n, the relations Sequence (signalled by &amp;quot;then&amp;quot; and &amp;quot;finally'i, Elaboration ('in particular'), and Purpose ('in order to').</Paragraph>
    <Paragraph position="3"> In the corresponding paragraph produced by Penman, the relations' characteristic words or phrases (boldfaced below) appear between the blocks of text they relate: \[The system asks the user to tell it the character~stlc of the program to be enhanced.l(6) Then \[the system applies transformations to the program.\](b) In particular, \[the system scans the program\](c) in order to \[f~nd opportunitlea to apply ~ranaformations to the program.\]{a) Then \[the system resolves conflicts.\](e) lit confu'ms the enhancemeng with the user.\](/) Finally, \[it performs the enhancement.\](g)  As stated above, the structurer is a simplified top-down hierarchical expansion planner (see Figure 2). It operates as follows: given one or more communicative goals, it find s RST relations whose intended effects match (some of) these goals; it then inspects which of the input elements match the nucleus and subgoal constraints for each relation. Unmatched constraints become subgoals which are posted on an agenda for the next level of planning. The tree can be expanded in either depth-first or breadth-first fashion. Eventually, the structuring process bottoms out when either: (a) all input elements have been used and unsatisfied subgoais remain (in which case the structurer could request more input with desired properties from the encapsulating system); or (b) all goals axe satisfied. If more than one plan (i.e., para. graph tree structure) is produced, the results axe ordered by preferring trees with the minimum unused number of input elements and the minimum number of remaining unsatisfied subgoals. The best tree is then traversed in left-to-right order; leaves provide input to Penman to be generated in English and relations at branch points provide typical interclausal relation words or phrases. In this way the structurer performs top-down goal refinement clown to the level of the input elements.</Paragraph>
  </Section>
class="xml-element"></Paper>