File Information
File: 05-lr/acl_arc_1_sum/cleansed_text/xml_by_section/metho/96/w96-0403_metho.xml
Size: 25,442 bytes
Last Modified: 2025-10-06 14:14:24
<?xml version="1.0" standalone="yes"?> <Paper uid="W96-0403"> <Title>Paraphrasing and Aggregating Argumentative Text Using Text Structure</Title> <Section position="3" start_page="21" end_page="23" type="metho"> <SectionTitle> 2 The Macroplanner of </SectionTitle> <Paragraph position="0"> P R 0 VERB The macroplanner of PROVERB combines hierarchical planning \[13\] with local organization \[15\] in a uniform planning framework \[6\]. The hierarchical planning is realized by so-called top-down presentation operators that split the task of presenting a particular proof into subtasks of presenting subproofs.</Paragraph> <Paragraph position="1"> While the overall planning mechanism is similar to the RST-based planning approach, the plan operators resemble the schemata in schema-based planning. The output of the macroplanner is an ordered sequence of proof communicative acts (PCAs).</Paragraph> <Paragraph position="2"> PCAs are the primitive actions planned during macroplanning to achieve communicative goals. Like speech acts, PCAs can be defined in terms of the communicative goals they fulfill as well as in terms of their possible verbalizations. Based on an analysis of proofs in mathematical textbooks, there are mainly two types of goals: Conveying derivation step: In terms of rhetorical relations, PCAs in this category represent a variation of the rhetorical relation derive \[8\]. Below we examine the simple PCA called Derive as an example.</Paragraph> <Paragraph position="3"> (Derive Reasons: (a 6F, F C G) Method : def-subset Conclusion: a 6G) Depending on the reference choices, the following is a possible verbalization: &quot;Since a is an element of F and F is a subset of G, a is an element of G by the definition of subset.&quot; Updating the global attentional structure: These PCAs either convey a partial plan for the forthcoming discourse or signal the end of a subproof. PCAs of this sort are also</Paragraph> <Section position="1" start_page="21" end_page="22" type="sub_section"> <SectionTitle> 3.1 Introduction and General Structure </SectionTitle> <Paragraph position="0"> Text Structure is first proposed by Meteer \[11, 12\] in order to bridge the generation gap between the representation in the application program and the linguistic resources provided by the language. By abstracting over concrete linguistic resources, Text Structure should supply the planner with basic vocabularies, with which it chooses linguistic resources. Meteer's text structure is organized as a tree, in which each node represents a constituent of the text. In this form it contains three types of linguistic information: constituency, structural relations among constituents, and in particular, the semantic categories the constituents express.</Paragraph> <Paragraph position="1"> The main role of the semantic categories is to provide vocabularies which specify type restrictions for nodes. They define how separate Text Structures can be combined, and ensure that the planner only builds expressible Text Structures. For instance if tree A should be expanded at node n by tree B, the resulting type of B must be compatible to the type restriction attached to n. Panaget \[14\] argues, however, that Meteer's semantic categories mix the ideational and the textual dimension as argued in the systemic linguistic theory \[5\]. Here is one of his examples: &quot;The ship sank&quot; is an ideational event, and it is textually presented from an EVENT-PERSPECTIVE. &quot;The sinking of the ship&quot; is still an ideational event, but now presented from an OBJECT-PERSPECTIVE.</Paragraph> <Paragraph position="2"> On account of this, Panaget split the type restrictions into two orthogonal dimensions: the ideational dimension in terms of the Upper Model \[1\], and the hierarchy of textual semantic categories based on an analysis of French and of English. In our work, we basically follow the approach of Panaget.</Paragraph> <Paragraph position="3"> Technically speaking, the Text Structure in PROVERB is a tree recursively composed of kernel subtrees or composite subtrees: An atomic kernel subtree has a head at the root and arguments as children, representing basically a predicate/argument structure.</Paragraph> <Paragraph position="4"> Composite subtrees can be divided into two subtypes: the first has a special matrix child and zero or more adjunct children and represents linguistic hypotaxis, the second has two or more coordinated children and stands for parataxis.</Paragraph> </Section> <Section position="2" start_page="22" end_page="22" type="sub_section"> <SectionTitle> 3.2 Type Restrictions </SectionTitle> <Paragraph position="0"> Each node is typed both in terms of the Upper Model and the hierarchy of textual semantic categories. The Upper Model is a domain-independent property inheritance network of concepts that are hierarchically organized according to how they can be linguistically expressed. Figure 1 shows a fragment of the Upper Model in PRO VERB. For every domain of application, domain-specific concepts must be identified and placed as an extension of the Upper Model.</Paragraph> <Paragraph position="1"> The hierarchy of textual semantic categories is also a domain-independent property inheritance network. The concepts axe organized in a hierarchy based on their textual realization. For example, the concept clause-modifier-rankingl t is realized as an adverb, clause-modifier-rankingll as a prepositional phrase, and clause-modifier-embedded as an adverbial clause. Fig. 2 shows a fragment of the hierarchy of textual semantic categories.</Paragraph> </Section> <Section position="3" start_page="22" end_page="23" type="sub_section"> <SectionTitle> 3.3 Mapping APOs to UMOs </SectionTitle> <Paragraph position="0"> The mapping from the content to the linguistic resources now happens in a two-staged way. While Meteer associates the application program objects (APOs) directly with so-called resources trees, we map APOs into Upper Model objects, which in turn are expanded to the Text Structures. It is worth noting that there is a practical advantage of this two-staged process. Instead of having to construct resource trees for APOs, the user of our system only needs to define a mapping from the APOs to Upper Model objects (UMOs).</Paragraph> <Paragraph position="1"> When mapping APOs to UMOs, the microplanner must choose among available alternatives. For example, the application program object para that stands for the logical predicate denoting the parallelism relation between lines may map in five different Upper Model concepts. In the 0-place case, para can be mapped into object leading to the noun &quot;parallelism,&quot; or quality, leading to the adjective &quot;parallel.&quot; In the binary case, the choices are property-ascription that may be verbalized as &quot;x and y are parallel,&quot; quality-relation that allows the verbalization as &quot;x is parallel to y&quot;, or process-relation, that is the formula &quot;x II Y.&quot; The mapping of Upper Model objects into the Text Structure is defined by so-called resource trees, i.e. reified instances of text structure subtrees. The resource trees of an Upper Model concept are assembled in its realization class.</Paragraph> <Paragraph position="2"> ~Concepts of the hierarchy of textual semantic categories are noted in sans-serif text.</Paragraph> <Paragraph position="3"> concept &quot; modified-concept f- conscious-being - object ---q ..</Paragraph> <Paragraph position="4"> t- non-concious-Uung \[- relational-processes- process t. mental-processes ._~- modal-qualio' quality t. material-word-quality r&quot; logical - arbitraty-place-relation-~ ,- sequence - generalized-possession - quantifications ., .</Paragraph> <Paragraph position="5"> discrete-place-relation ~ . . r&quot; taenttty r&quot; property-ascription tntenstve ---'1 . . t- ascription -1_ circumstantial class-ascription Figure 1: A Fragment of Upper Model in PROVERB - text - sentence - clause</Paragraph> </Section> </Section> <Section position="4" start_page="23" end_page="25" type="metho"> <SectionTitle> 4 Architecture and Control </SectionTitle> <Paragraph position="0"> The main tasks of our microplanner include aggregation to remove redundancies, insertion of cue words to increase coherence, and reference choices, as well as lexical choices. Apart from that, the microplanner also handles sentence scoping and layout. An overview of the microplanner's architecture is provided in Figure 3.</Paragraph> <Paragraph position="1"> Our microplanner takes as input an ordered sequence of PCAs, structured in an attentional hierarchy. The first module, the derivation reference choice component (DRCC), suggests which parts of a PCA are to be verbalized. This is done based on the hierarchical discourse structure as well as on the textual distance. PCAs annotated with these decisions annotated are called preverbal messages (PMs).</Paragraph> <Paragraph position="2"> Starting from a list of PMs as the initial Text Structure, the microplanner progressively maps application program concepts in PMs into text structure objects of some textual semantic type by referring to Upper Model objects as an intermediate level. The Text Structure evolves by the expansion of leaves top-down and left to right. This process is controlled by the main module of our microplanner, the Text Structure Generator (TSG), which carries out the following algorithm: null * When the current node is an APO with more than one son, apply ordering and aggregation, in order to produce more concise and more coherent text. The application of an aggregating rule before the expansion of a leaf node may trigger the insertion of cue words.</Paragraph> <Paragraph position="3"> * An APO is mapped into an UMO, which is in turn expanded into a Text Structure by choosing an appropriate resource tree.</Paragraph> <Paragraph position="4"> * A fully expanded Text Structure will be traversed again: -- to choose the appropriate lexical items. -- to make sentence scoping decisions by singling out one candidate textual semantic category for each constituent.</Paragraph> <Paragraph position="5"> This in turn may trigger the execution of a cue word rule. For instance, the choice of the category sentence for a constituent may lead to the insertion of the cue word &quot;furthermore&quot; in the next sentence.</Paragraph> <Paragraph position="6"> --to determine the layout parameters, which will be realized later as ~TEXcommands in the final output text.</Paragraph> <Paragraph position="7"> A Text Structure constructed in this way is the output of our microplanner, and will be transformed into the input formalism of TAG-GEN \[10\], our linguistic realizer.</Paragraph> <Paragraph position="8"> In the next two sections, we concentrate on two major tasks of the Text Structure generator: to choose compatible paraphrases of application program concepts, and to improve the textual structure by applying agture generator chooses among paraphrases and avoids building inexpressible text structures via type checking.</Paragraph> <Paragraph position="9"> Example We examine a simple logic formula derive(para(C1,C2),B). Note that B stands for a conclusion which will not be examined here. We will also not follow the procedure in detail.</Paragraph> <Paragraph position="10"> In the current implementation, the rhetorical relation derive is only connected to one Upper Model concept derive, a subconcept of cause.relation. The realization class associated to the concept, however, contains several alternative resource trees leading to different patterns of verbalization. We only list two variations below: * B, since A.</Paragraph> <Paragraph position="11"> * Because of A, B.</Paragraph> <Paragraph position="12"> The resource tree of the first alternative is given in Fig. 4.</Paragraph> <Paragraph position="13"> The logic predicate para(C1, C2) can be mapped to one of the following Upper Model concepts, where we always include one possible verbalization:</Paragraph> <Paragraph position="15"> Textually, the property-ascription version can be realized in two forms, represented by the two resource trees in Fig. 5.</Paragraph> <Paragraph position="16"> Type checking during the construction of the Text Structure must ensure, that the realization be compatible along both the ideational and the textual dimension. In this example, the combination of the tree in Fig. 4 and the first tree in Fig. 5 is compatible and will lead to the verbalization: &quot;B, since C1 and C2 are parallel.&quot; source Trees The second tree in Fig. 5, however, can only be combined with another realization of derive, resulting in: &quot;Because of the parallelism of line C1 and line C2, B.&quot; In our current system we concentrat on the mechanism and are therefore still experimenting with heuristics which control the choice of paraphrases. One interesting rule is to distinguish between general rhetorical relations and domain specific mathematical concepts. While the former should be paraphrased to increase the flexibility, continuity of the latter helps the user to identify technical concepts.</Paragraph> </Section> <Section position="5" start_page="25" end_page="26" type="metho"> <SectionTitle> 6 Semantic Aggregation </SectionTitle> <Paragraph position="0"/> <Section position="1" start_page="25" end_page="26" type="sub_section"> <SectionTitle> Rules </SectionTitle> <Paragraph position="0"> Although the handling of paraphrase generation already increases the flexibility in the text, the default verbalization strategy will still expand the Text Structure by recursively descending the proof and formula structure, and thereby forced to keep these structures. To achieve the second verbalization of equation (1) in the introduction, however, we have to combine Set(F) and Subset(F, G) to form an embedded structure Subset(Set(F), G). Clearly, although still in the same format, this is no more an Upper Model object, since Set(F) is an Upper Model process, not an object. Actually, this documents a textual decision that no matter how Subset and Set should be instantiated, the argument F in Subset(F, G) will be replaced by Set(F). This textual operation eliminates one of the duplicates of F. This section is devoted to various textual reorganisations which eliminate such redundancies. Following the tradition, we call them aggregation rules.</Paragraph> <Paragraph position="1"> As it will become clear when handling concrete aggregation rules, such rules may narrow the realization choices of APOs by imposing additional type restrictions. Furthermore, some realization choices block desirable textual reorganisation. On account of this we carry out aggregations before concrete resources for the APOs like object and class-ascription are chosen.</Paragraph> <Paragraph position="2"> APOs, before they are mapped to UMOs, can be viewed as variables for UMOs (for convenience, we continue to refer to them as APOs). In this sense, our rules work with such variables at the semantic level of the Upper Model, and therefore differ from those more syntactic rules reported in the literature. For a comparison see Sec. 6.4.</Paragraph> <Paragraph position="3"> So far, we have investigated three types of aggregation which will be addressed in the next two subsections. A categorization of the aggregation rules is given in Fig. 6.</Paragraph> <Paragraph position="5"/> </Section> <Section position="2" start_page="26" end_page="26" type="sub_section"> <SectionTitle> 6.1 Semantic Grouping </SectionTitle> <Paragraph position="0"> We use semantic grouping to characterize the merge of two parallel Text Structure objects with the same top-concept by grouping their arguments. Two APOs are parallel in the sense that they have the same parent node.</Paragraph> <Paragraph position="1"> The general form of this type of rules can be characterized by the pattern as given below:</Paragraph> </Section> <Section position="3" start_page="26" end_page="26" type="sub_section"> <SectionTitle> Rule Pattern A </SectionTitle> <Paragraph position="0"/> <Paragraph position="2"> The syntax of our rules means that a text structure of the form above the bar will be transformed into one of the form below the bar. Viewing Text Structure as a tree, P\[a\] and P\[b\] are both sons of +, they are merged together by grouping the arguments a and b under another operator ~. In the first rule below, + and ~ are identical.</Paragraph> <Paragraph position="3"> where + can be either a logical A or a logical V, and P stands for a logical predicate. The following example illustrates the effect of this rule.</Paragraph> </Section> </Section> <Section position="6" start_page="26" end_page="26" type="metho"> <SectionTitle> Set(F) A Set(G) </SectionTitle> <Paragraph position="0"> &quot;F is a set. G is a set.&quot; are aggregated to:</Paragraph> </Section> <Section position="7" start_page="26" end_page="28" type="metho"> <SectionTitle> Set(F A G) </SectionTitle> <Paragraph position="0"> &quot;F and G are sets.&quot; The rule covers the predicate grouping rule reported in \[3\]. This is also the best place to explain why we apply aggregation before choosing concrete linguistic resources. If the two occurrences of Set are instantiated differently, this rule will be blocked.</Paragraph> <Paragraph position="1"> Now let us examine another semantic grouping rule, where + and ~ are no longer identical.</Paragraph> <Paragraph position="2"> Here +, ~, and P are instantiated to A, V, and ~, respectively. By instantiating +, E\[~ and P in pattern A to different logical connectives and derivation relations, we have alltogether five rules in this category. The correctness of the rules in this category with respect to the information conveyed is guaranteed by the semantics of the Upper Model concerned. In the case of rule A.2 for instance, (PiVP2) ~ C is a logical consequence of (P1 ~ C) A (P2 ~ C).</Paragraph> <Section position="1" start_page="26" end_page="26" type="sub_section"> <SectionTitle> 6.2 Semantic Embedding </SectionTitle> <Paragraph position="0"> The next category of aggregation rules handles parallel structures which are not identical. In this case, some of them may be converted to embedded structures, as is done by the following rule.</Paragraph> <Paragraph position="1"> concepts the argument T of f may take, * concepts(P) denotes the Upper Model concept P may result in.</Paragraph> <Paragraph position="2"> We require also that PIT\] is realized as an object T with modifiers. It is this intuitive explanation which guarantees the correctness of this rule with respect to meaning.</Paragraph> <Paragraph position="3"> The following example illustrates this rule, in particular, how the decision made here narrows the choices of linguistic resources for both P and T as an argument of Q. We begin with the two APOs in a conjunction below, containing a common APO F.</Paragraph> <Paragraph position="4"> Set(F) A Subset(F, G) &quot;F is a set. F is a subset of G.&quot; Since F is directly governed by Subset, f and Q in our rule above coincide here.</Paragraph> <Paragraph position="5"> concepts(Subset, F) = {object), while concepts(Set) = (class-ascription, object). Therefore, their intersection is {object). This not only guarantees the expressibility of the new APO, but also restricts the choice of linguistic resources for Set, now restricted to object. The result as well as its verbalization is given below: Subset(Set(F), G) &quot;The set F is a subset of G.&quot; Actually, for mathematical texts we have only used two embedding rules, with the other being the dual of rule B.1 where P and Q change their places.</Paragraph> </Section> <Section position="2" start_page="26" end_page="28" type="sub_section"> <SectionTitle> 6.3 Pattern-based rules Optimization </SectionTitle> <Paragraph position="0"> Rules in the third category involve more complex changes of the textual structure in a way which is neither a grouping nor an embedding. They could be understood as some domain-specific communicative conventions, and must be explored in every domain of application. In PRO VERB, currently four such rules are integrated. Three of them build a sequence of some transitive relations into a chain.</Paragraph> <Paragraph position="1"> Rule C. 1 below addresses the problem that every step of derivation is mapped to a separate sentence in the default verbalization. It reflects the familiar phenomenon that when several derivation steps form a chain, they are verbalized in a more connected way. To accommodate the phenomenon of a chain, we have also added a slot called next in the domain model concept derive-chain. Now suppose that we have two consecutive derivations with R1,M1,C1 and R2, M2, C2 as its premises (called reasons), the rule of inference (called method), and the conclusion.</Paragraph> <Paragraph position="2"> They form part of a chain if the conclusion C1 is used as a premise in the second step, namely C1 E R2. In this case, the following rule combines them into a chain by putting the second derivation into the next slot of the chain. At the same time, C1 is removed from</Paragraph> <Paragraph position="4"> The following example illustrates how this rule works. We will only give the verbalization and omit the Text Structure. Given a sequence of two derivation steps which can be verbalized as: &quot;0 C_ o'*, by the definition of transitive closure.&quot; and &quot;Since (x, y) E a and o C o*, (x, y) E c* by the definition of subset.&quot; Rule C.1 will produce a chain which will be verbalized as &quot;a C 0&quot; by the definition of transitive closure, thus establishing (x, y) E 0&quot; by the definition of subset, since (x,y) E 0.&quot; Note that the rule above is only a simplification of a recursive definition, since chaining is not restricted to two derivation steps. Readers are referred to \[4\]. Although this rule inserts the second derive into another Text Structure, the resulting structure is now a chain, no longer a plain derive. Therefore it distinguishes clearly f;om the rules in Section 6.2.</Paragraph> <Paragraph position="5"> There are two more chaining rules for the logical connectors implication and equivalence. A further rule removes redundancies in some case analyses (see \[4\]).</Paragraph> </Section> <Section position="3" start_page="28" end_page="28" type="sub_section"> <SectionTitle> 6.4 Discussion </SectionTitle> <Paragraph position="0"> While many systems have some. aggregation rules implemented \[9, 2\], there are comparatively few detailed discussions in the literature. The most structured categorization we found is the work of Dalianis and Hovy \[3\], where they define aggregation as a way of avoiding redundancy. Some of their rules, nevertheless, make decisions which we would call reference choice. Since this is treated in another module, we define our aggregation at the semantic level. The following are several significant features of our aggregation rules.</Paragraph> <Paragraph position="1"> The first difference is that our aggregation rules are defined in terms of manipulations of the Upper Model. They remove redundancies by combining the linguistic resources of two adjacent APOs, which contain redundant content. They cover the more syntactic rules reported in the literature at a more abstract level.</Paragraph> <Paragraph position="2"> Second, Text Structure provides us stronger means to specify textual operations.</Paragraph> <Paragraph position="3"> While rules reported in the literature typically aggregate clauses, our rules operate both above and beneath the level of clause constituents.</Paragraph> <Paragraph position="4"> Third, while most investigations have concentrated on general purpose microplanning operations, we came to the conclusion that microplanning needs domain-specific rules and patterns as well.</Paragraph> </Section> </Section> <Section position="8" start_page="28" end_page="28" type="metho"> <SectionTitle> 7 A Running Example </SectionTitle> <Paragraph position="0"> The following example illustrates the mechanism of aggregation and its effect on resulting text. We start with the following sequence of PMs:</Paragraph> <Paragraph position="2"> Without aggregation, the system produces: &quot;Let F be a set. Let G be a set. Let F C G.</Paragraph> <Paragraph position="3"> Let a E F. Let b E F. Since a E F and F C G, a E G. Since b E F and F C G, b E G.&quot; Aggregation of the assume-PMs results in:</Paragraph> <Paragraph position="5"> whereas the application of the grouping rule for independent derive-PMs provides: derive( ( element( a, F) A element(b, F) A Subset(F, G)), e, (element(a, G) A element(b, G) ) ) After that, the predicate grouping rule A.1 is applied to the arguments of assume, which are grouped to:</Paragraph> <Paragraph position="7"> A element(a A B, F A F))) Note that F A F is later reduced to F. Predicate grouping applies to the arguments of derive in a similar way. Finally, the system produces the following output: &quot;Let F and G be sets, F C G, and a, b E F. Then a, b E G.&quot;</Paragraph> </Section> class="xml-element"></Paper>