File Information
File: 05-lr/acl_arc_1_sum/cleansed_text/xml_by_section/metho/03/w03-0909_metho.xml
Size: 24,675 bytes
Last Modified: 2025-10-06 14:08:24
<?xml version="1.0" standalone="yes"?> <Paper uid="W03-0909"> <Title>Surfaces and Depths in Text Understanding: The Case of Newspaper Commentary</Title> <Section position="3" start_page="0" end_page="0" type="metho"> <SectionTitle> 2 Sample commentary </SectionTitle> <Paragraph position="0"> Figure 1 shows a sample newspaper commentary, taken from the German regional daily &quot;M&quot;arkische Allgemeine Zeitung&quot; in October 2002, along with an English translation. To ease reference, numbers have been inserted in front of the sentences. Let us first move through the text and make some clarifications so that the reader can get the picture of what is going on. Dagmar Ziegler is the treasury secretary of the German state of Brandenburg. A plan for early retirement of teachers had been drafted collectively by her and the education secretary, whose name is Reiche. Sentence 5 points out that the plan had intended education to be exempt from the cutbacks happening all over the various ministries -- Reiche's colleagues in 6 are thus the other secretaries. While the middle part of the text provides some motivation for the withdrawal, 9-14 state that the plan nonetheless should be implemented, for the reasons given in 10-12. Our intended &quot;most concise summary&quot; then would be: a0 Topic: Treasury secretary delays decision on teacher staff plan a0 Author's opinion: Government has to decide quickly and give priority to education, thus implement the plan Notice that a statistical summarization technique (i.e., a sentence extraction approach) is very unlikely to yield (1) Dagmar Ziegler sitzt in der Schuldenfalle. (2) Auf Grund der dramatischen Kassenlage in Brandenburg hat sie jetzt eine seit mehr als einem Jahr erarbeitete Kabinettsvorlage &quot;uberraschend auf Eis gelegt und vorgeschlagen, erst 2003 dar&quot;uber zu entscheiden. (3) &quot;Uberraschend, weil das Finanz- und das Bildungsressort das Lehrerpersonalkonzept gemeinsam entwickelt hatten. (4) Der R&quot;uckzieher der Finanzministerin ist aber verst&quot;andlich. (5) Es d&quot;urfte derzeit schwer zu vermitteln sein, weshalb ein Ressort pauschal von k&quot;unftigen Einsparungen ausgenommen werden soll auf Kosten der anderen. (6) Reiches Ministerkollegen werden mit Argusaugen dar&quot;uber wachen, dass das Konzept wasserdicht ist. (7) Tats&quot;achlich gibt es noch etliche offene Fragen. (8) So ist etwa unklar, wer Abfindungen erhalten soll, oder was passiert, wenn zu wenig Lehrer die Angebote des vorzeitigen Ausstiegs nutzen. (9) Dennoch gibt es zu Reiches Personalpapier eigentlich keine Alternative. (10) Das Land hat k&quot;unftig zu wenig Arbeit f&quot;ur zu viele P&quot;adagogen. (11) Und die Zeit dr&quot;angt. (12) Der grosse Einbruch der Sch&quot;ulerzahlen an den weiterf&quot;uhrenden Schulen beginnt bereits im Herbst 2003. (13) Die Regierung muss sich entscheiden, und zwar schnell. (14) Entweder sparen um jeden Preis - oder Priorit&quot;at fuer die Bildung.</Paragraph> <Paragraph position="1"> (1) Dagmar Ziegler is up to her neck in debt. (2) Due to the dramatic fiscal situation in Brandenburg she now surprisingly withdrew legislation drafted more than a year ago, and suggested to decide on it not before 2003. (3) Unexpectedly, because the ministries of treasury and education both had prepared the teacher plan together. (4) This withdrawal by the treasury secretary is understandable, though. (5) It is difficult to motivate these days why one ministry should be exempt from cutbacks -- at the expense of the others. (6) Reiche's colleagues will make sure that the concept is waterproof. (7) Indeed there are several open issues. (8) For one thing, it is not clear who is to receive settlements or what should happen in case not enough teachers accept the offer of early retirement. (9) Nonetheless there is no alternative to Reiche's plan. (10) The state in future has not enough work for its many teachers. (11) And time is short. (12) The significant drop in number of pupils will begin in the fall of 2003. (13) The government has to make a decision, and do it quickly. (14) Either save money at any cost - or give priority to education. a result along these lines, because word frequency is of little help in cases where the line of the argument has to be pulled out of the text, and might make some synthesis necessary. Just to illustrate the point, the Microsoft Word &quot;25 percent&quot; summarization reads as follows: &quot;Uberraschend, weil das Finanz- und das Bildungsressort das Lehrerpersonalkonzept gemeinsam entwickelt hatten. Reiches Ministerkollegen werden mit Argusaugen dar&quot;uber wachen, dass das Konzept wasserdicht ist. Entweder sparen um jeden Preis - oder Priorit&quot;at f&quot;ur die Bildung.</Paragraph> <Paragraph position="2"> Unexpectedly, because the ministries of treasury and education both had prepared the teacher plan together. Reiche's colleagues will make sure that the concept is waterproof. Either save money at any cost - or give priority to education.</Paragraph> <Paragraph position="3"> It includes the final sentence (most probably because it is the final sentence), but in the context of the other two extracted sentences it does not convey the author's position -- nor the precise problem under discussion.</Paragraph> </Section> <Section position="4" start_page="0" end_page="0" type="metho"> <SectionTitle> 3 Rhetorical Structure </SectionTitle> <Paragraph position="0"> Since RST (Mann, Thompson 1988) has been so influential in discourse-oriented computational linguistics, we start our analysis with a &quot;man-made&quot; RST analysis, which was produced collectively by two RSTexperienced students. See Figure 2.1 (The English reader can relatively easy map the German segments to their translations in Fig. 1 with the help of the sentence numbers added to the text in the tree).</Paragraph> <Paragraph position="1"> Some considerations motivating this analysis (in terms of segment numbers, not sentence numbers): 1 is seen as the general Background for the satellite of the over-all Concession, which discusses the problem arising from the debt situation. Arguably, it might as well be treated as Background to the entire text. The Evaluation between 2-6 and 7-12 is a relation often found in opinion texts; an alternative to be considered here is Antithesis -- in this case, however, 7-12 would have to be the nucleus, which seems to be problematic in light of the situation that 3-4 is the main portion that is being related to the material in 13-16.</Paragraph> <Paragraph position="2"> 8-12 explains and elaborates the author's opinion that the withdrawal is understandable (7). The distinctions between the relations Explanation, Elaboration, and Evidence were mostly based on surface cues, such as tats&quot;achlich ('indeed') signalling Evidence. The Elabora1Visualization by the RST Tool (O'Donnell, 1997). Notation follows Mann and Thompson (1988): vertical bars and incoming arrows denote nuclear segments, outgoing arrows denote satellites. Numbers at leaves are sentence numbers; segment numbers are given at internal nodes.</Paragraph> <Paragraph position="3"> tions, would arrive at such a tree -- more specifically, at a formal representation of it.</Paragraph> <Paragraph position="4"> What kind of information is necessary beyond assigning relations, spans and nuclei? In our representation of the summary tree, we have implicitly assumed that reference resolution has been worked out - in particular that the legislation can be identified in the satellite of the Explanation, and also in its nucleus, where it figures implicitly as the object to be decided upon. Further, an RST tree does not explicitly represent the topic of the discourse, as we had asked for in the beginning. In our present example, things happen to work out quite well, but in general, an explicit topic identification step will be needed. And finally, the rhetorical tree does not have information on illocution types (1-place rhetorical relations, so to speak) that distinguish reported facts (e.g., segments 3 and 4) from author's opinion (e.g., segment 7). We will return to these issues in Section 6, but first consider the chances for building up rhetorical trees automatically.</Paragraph> </Section> <Section position="5" start_page="0" end_page="0" type="metho"> <SectionTitle> 5 Prospects for Rhetorical Parsing </SectionTitle> <Paragraph position="0"> Major proponents of rhetorical parsing have been (Sumita et al., 1992), (Corston-Oliver, 1998), (Marcu, 1997), and (Schilder, 2002). All these approaches emphasise their membership in the &quot;shallow analysis&quot; family; they are based solely on surface cues, none tries to work with semantic / domain / world knowledge. (Corston-Oliver and Schilder use some genre-specific heuristics for preferential parsing, though.) In general, our sample text belongs to a rather &quot;friendly&quot; genre for rhetorical parsing, as commentaries are relatively rich in connectives, which are the most important source of information for making decisions -- but not the only one: Corston-Oliver, for example, points out that certain linguistic features such as modality can sometimes help disambiguating connectives. Let us now hypothesize what an &quot;ideal&quot; surface-oriented rhetorical parser, equipped with a good lexicon of connectives, part-of-speech tagger and some rough rules of phrase composition, could do with our example text.</Paragraph> <Section position="1" start_page="0" end_page="0" type="sub_section"> <SectionTitle> 5.1 Segmentation </SectionTitle> <Paragraph position="0"> As we are imagining an &quot;ideal&quot; shallow analyser, it might very well produce the segmentation that is underlying the human analysis in Figure 2. The obvious first step is to establish a segment boundary at every full stop that terminates a sentence (no ambiguities in our text). Within sentences, there are six additional segment boundaries, which can be identified by considering connectives and part-of-speech tags of surrounding words, i.e. by a variant of &quot;chunk parsing&quot;: Auf Grund ('due to') has to be followed by an NP and establishes a segment up to the finite verb (hat). The und ('and') can be identified to conjoin complete verb phrases and thus should trigger a boundary. In the following sentence, weil ('because') has to be followed by a full clause, forming a segment. The next intra-sentential break is between segments 11 and 12; the oder ('or') can be identified like the und above. In segment 17-18, und zwar ('and in particular') is a strict boundary marker, as is the entweder - oder ('either - or') construction in 19-20.</Paragraph> </Section> <Section position="2" start_page="0" end_page="0" type="sub_section"> <SectionTitle> 5.2 Relations, scopes, nuclei </SectionTitle> <Paragraph position="0"> The lexical boundary markers just mentioned also indicate (classes of) rhetorical relationships. Auf Grund -when used in its idiomatic reading -- signals some kind of Cause with the satellite following in an NP. Because the und in 3-4 co-occurs with the temporal expressions jetzt ('now') and erst 2003 ('not before 2003'), it can be taken as a signal of Sequence here, with the boundaries clearly identifiable, so that the RST subtree 2-4 can be derived fully. Furthermore, 5 takes up a single adverbial &quot;uberraschend from 3, and in conjunction with the weil-clause in 6, the Elaboration can be inferred. weil ('because') itself signals some Cause, but the nuclearity decision (which in the &quot;real&quot; tree in Fig. 2 leads to choosing Result) is difficult here; since 5 merely repeats a lexeme from 3, we might assign nuclearity status to 6 on the &quot;surface&quot; grounds that it is longer and provides new material. We thus have derived a rhetorical structure for the entire span 2-6. In 7, aber ('but') should be expected to signal either Contrast or Concession; how far the left-most span reaches can not be determined, though. Both 8 and 9 provide no reliable surface clues. In 10, tats&quot;achlich ('indeed') can be taken as an adverbial indicating Evidence; again the scope towards the left is not clear. So ..</Paragraph> <Paragraph position="1"> etwa ('thus .. for instance') in 11 marks an Elaboration, and the oder in 12 a Disjunction between the two clauses.</Paragraph> <Paragraph position="2"> Span 10-12 therefore receives an analysis. In 13, dennoch ('nonetheless') is a clear Concession signal, but its scope cannot be reliably determined. Finally, the only two remaining decisions to be made from surface observations are the Elaboration 17-18 (und zwar, 'and in particular') and the Disjunction 19-20. Then, making use of RST's &quot;empty&quot; relation Join, we can bind together the assembled pieces and are left with the tree shown in Fig. 4.</Paragraph> <Paragraph position="3"> sions. First, rhetorical parsing should allow for under-specified representations as -- intermediate or final -outcome; see (Hanneforth et al., submitted). Second, text understanding aiming at quality needs to go further than surface-oriented rhetorical parsing. With the help of additional domain/world-knowledge sources, attempts should be made to fill gaps in the analysis. It is then an implementation decision whether to fuse these additional processes into the rhetorical parser, or to use a pipeline approach where the parser produces an under-specified rhetorical tree that can afterwards be further enriched. Third, probabilistic or statistical knowledge can also serve to fill gaps, but the information drawn from such sources should be marked with its status being insecure. As opposed to decisions based on lexical/linguistic knowledge (in 5.2), the tentative decisions from 5.3 may be overwritten by later knowledge-based processes.</Paragraph> </Section> </Section> <Section position="6" start_page="0" end_page="0" type="metho"> <SectionTitle> 6 Knowledge-Based Understanding </SectionTitle> <Paragraph position="0"> &quot;Understanding a text&quot; for some cognitive agent means to fuse prior knowledge with information encountered in the text. This process has ramifications for both sides: What I know or believe influences what exactly it is that I &quot;take away&quot; from a text, and my knowledge and beliefs will usually to a certain extent be affected by what I read. Naturally, the process varies from agent to agent: They will understand different portions of a text in different ways and to different degrees. Thus, when we endeavour to devise and implement models of text understanding, the target should not be to arrive at &quot;the one and only&quot; result, but rather to account for the mechanics of this variability: the mechanism of understanding should be the same, but the result depend on the type and amount of prior knowledge that the agent carries. In the end, a representation of text meaning should therefore be designed to allow for this flexibility.</Paragraph> <Section position="1" start_page="0" end_page="0" type="sub_section"> <SectionTitle> 6.1 KB Design </SectionTitle> <Paragraph position="0"> In line with many approaches to using knoweldge for language processing, we adopt the framework of terminological logic as the vehicle for representing both the background knowledge necessary to bootstrap any understanding process, and the content of the text. Thus the basic idea is to encode prior, general knowledge in the TBox (concepts) and the information from the text in the ABox (instances). For our example, the subworld of government, ministries and legislation has to be modelled in the TBox, so that entities referred to in the text can instantiate the appropriate concepts. We thus map the rhetorical tree built up by shallow analysis to an ABox in the LOOM language (MacGregor, Bates, 1987); for a sketch of representing rhetorical structure in LOOM, see (Stede, 1999, ch. 10).</Paragraph> </Section> <Section position="2" start_page="0" end_page="0" type="sub_section"> <SectionTitle> 6.2 &quot;Ideal&quot; text understanding </SectionTitle> <Paragraph position="0"> Each leaf of the tree is now subject to detailled semantic analysis and mapped to an enriched predicate/argument structure that instantiates the relevant portions of the TBox (quite similar to the 'Text Meaning Representation' of (Mahesh, Nirenburg, 1996)). &quot;Enriched&quot; indicates that beyond the plain proposition, we need information such as modality but also the type of illocution; e.g., does the utterance represent a factual statement, the author's opinion, or a proposal? This is necessary for analyzing the structure of an argument (but, of course, often it is very difficult to determine).</Paragraph> <Paragraph position="1"> One central task in text understanding is reference resolution. Surface-based methods can perform initial work here, but without some background knowledge, the task can generally not be completed. In our sample text, understanding the argument depends on recognizing that Kabinettsvorlage in (2), Lehrerpersonalkonzept in (3), Konzept in (6), and Reiches Personalpapier in (9) all refer to the same entity; that Ziegler in (1) and Finanzministerin in (4) are co-referent; that Finanz- und Bildungsressort in (3), Reiches Ministerkollegen in (6), and die Regierung in (13) refer to portions of or the complete Brandenburg government, respectively. Once again, hints can be derived from the surface words (e.g., by compund analysis of Lehrerpersonalkonzept), but only background knowledge (an ontology) about the composition of governments and their tasks enables the final decisions.</Paragraph> <Paragraph position="2"> Knowledge-based inferences are necessary to infer rhetorical relations such as Explanation or Evaluation.</Paragraph> <Paragraph position="3"> Consider for example segment 15-16, where the relationship between 'time is short' (a subjective, evaluative statement) and 'begin already in the fall of 2003' (a statement of a fact), once recognized, prompts us to assign Explanation. Similarly, the Elaboration between this segment and the preceeding 14 can be based on the fact that 14 makes a statement about the 'future situation' in Brandenburg, which is made more specific by time being short and the fall of 2003. More complex inferences are necessary to attach 14-16 then to 13 (and similarly in the segment 7-12).</Paragraph> </Section> <Section position="3" start_page="0" end_page="0" type="sub_section"> <SectionTitle> 6.3 &quot;Realistic&quot; text understanding </SectionTitle> <Paragraph position="0"> Even if it were possible to hand-code the knowledge base such that for our present sample text the complete representation can be constructed -- for the general text analysis situation, achieving a performance anywhere near the &quot;complete and correct solution&quot; is beyond reach. As indicated at the beginning of the section, though, this is not necessarily bad news, as a notion of partial understanding, or &quot;mixed-depth encoding&quot; as suggested by Hirst and Ryan (1992), should be the rule rather than the exception. Under ideal circumstances, a clause at a leaf of the rhetorical tree might be fully analyzed, with all references resolved and no gaps remaining. In the worst case, however, understanding might fail entirely. Then, following Hirst and Ryan, the text portion itself should simply be part of the representation. In most cases, the representation will be somewhere in-between: some aspects fully analyzed, but others not or incompletely understood.</Paragraph> <Paragraph position="1"> For example, a sentence adverbial might be unknown and thus the modality of the sentence not be determined. The ABox then should reflect this partiality accordingly, and allow for appropriate inferences on the different levels of representation.</Paragraph> <Paragraph position="2"> The notion of mixed depth is relevant not only for the tree's leaves: Sometimes, it might not be possible to derive a unique rhetorical relation between two segments, in which case a set of candidates can be given, or none at all, or just an assignment of nucleus and satellite segments, if there are cues allowing to infer this. In (Reitter and Stede, 2003) we suggest an XML-based format for representing such underspecified rhetorical structures.</Paragraph> <Paragraph position="3"> Projecting this onto the terminological logic scheme, and adding the treatment of leaves, we need to provide the TBox not only with concepts representing entities of &quot;the world&quot; but also with those representing linguistic objects, such as clause or noun group, and for the case of unanalyzed material, string. To briefly elaborate the noun group example, consider Reiches Ministerkollegen ('Reiche's colleagues') in sentence 6. Shallow analysis will identify Reiche as some proper name and thus the two words as a noun group. An ABox istance of this type is created, and it depends on the knowledge held by the TBox whether additional types can be inferred. Reiche has not been mentioned before in the text, because from the perspective auf the author the name is prominent enough to be identified promptly by the (local) readers.</Paragraph> <Paragraph position="4"> If the system's TBox contains a person of that name in the domain of the Brandenburg government, the link can be made; otherwise, Reiche will be some un-identified object about which the ABox collects some information from the text.</Paragraph> <Paragraph position="5"> Representations containing material with different degrees of analysis become useful when accompanied by processes that are able to work with them ('mixed-depth processing'). For summarization, this means that the task becomes one of fusing extraction (of unanalyzed portions that have been identified as important nuclei) with generation (from the representations of analyzed portions). Of course, this can lead to errors such as dangling anaphors in the extracted portions, but that is the price we pay for robustness -- robustness in this refined sense of &quot;analyse as deeply as you can&quot; instead of the more common &quot;extract something rather than fail.&quot;</Paragraph> </Section> </Section> <Section position="7" start_page="0" end_page="0" type="metho"> <SectionTitle> 7 Implementation Strategy </SectionTitle> <Paragraph position="0"> Finally, here is a brief sketch of the implementation work that is under way in the Computational Linguistics group at Potsdam University. Newspaper commentaries are the genre of choice for most of our current work. We have assembled a corpus of some 150 commentaries from &quot;M&quot;arkische Allgemeine Zeitung&quot;, annotated with rhetorical relations, using the RST Tool by O'Donnell (1997).</Paragraph> <Paragraph position="1"> It uses an XML format that we convert to our format of underspecified rhetorical structure ('URML' Reitter & Stede 2003).</Paragraph> <Paragraph position="2"> This data, along with suitable retrieval tools, informs our implementation work on automatic commentary understanding and generation. Focusing here on understanding, our first prototype (Hanneforth et al., submitted) uses a pipeline of modules performing 1. tokenization 2. sentence splitting and segmentation into clauses 3. part-of-speech tagging 4. chunk parsing 5. rhetorical parsing 6. knowledge-based processing The tagger we are using is the Tree Tagger by Schmid (1994); the chunk parser is CASS (Abney 1996). The remaining modules, as well as the grammars for the chunk parser, have been developed by our group (including student projects).2 The rhetorical parser is a chart parser and uses a discourse grammar leading to a parse forest, and is supported by a lexicon of discourse markers (connectives). We have started work on reference resolution (in conjunction with named-entity recognition). Addition of the knowledge-based component, as sketched in the previous section, has just begun. The main challenge is to allow for the various kinds of underspecification within the LOOM formalism and to design appropriate inference rules.</Paragraph> <Paragraph position="3"> As implementation shell, we are using GATE (http://www.gate.ac.uk), which proved to be a very useful environment for this kind of incremental system construction. null</Paragraph> </Section> class="xml-element"></Paper>