File Information
File: 05-lr/acl_arc_1_sum/cleansed_text/xml_by_section/intro/99/p99-1006_intro.xml
Size: 17,091 bytes
Last Modified: 2025-10-06 14:06:49
<?xml version="1.0" standalone="yes"?> <Paper uid="P99-1006"> <Title>Discourse Relations: A Structural and Presuppositional Account Using Lexicalised TAG*</Title> <Section position="3" start_page="41" end_page="45" type="intro"> <SectionTitle> 2 Framework </SectionTitle> <Paragraph position="0"> In previous papers (Cristea and Webber, 1997; Webber and Joshi, 1998; Webber et al., 1999), we have argued for using the more complex structures (elementary trees) of a Lexicalized Tree-Adjoining Grammar (LTAG) and its operations (adjoining and substitution) to associate structure and semantics with a sequence of discourse clauses. 2 Here we briefly review how it works.</Paragraph> <Paragraph position="1"> In a lexicalized TAG, each elementary tree has at least one anchor. In the case of discourse, the anchor for an elementary tree may be a lexical item, punctuation or a feature structure that is lexically null. The semantic contribution of a lexical anchor includes both what it presupposes and what it asserts (Stone and Doran, 1997; Stone, 1998; Stone and Webber, 1998). A feature structure anchor will either unify with a lexical item with compatible features (Knott and Mellish, 1996), yielding the previous case, or have an empty realisation, though one But a similar situation can occur within the clause, with relative clause dependencies - from the verb back to the relative pronoun and forward to a trace - so the possibility is not unmotivated from the perspective of syntax.</Paragraph> <Paragraph position="2"> 2We take this to be only the most basic level of discourse structure, producing what are essentially extended descriptions of situations/events. Discourse may be further structured with respect to speaker intentions, genre-specific presentations, etc. that maintains its semantic features.</Paragraph> <Paragraph position="3"> The initial elementary trees used here correspond, by and large, to second-order predicate-argument structures - i.e., usually binary predicates on propositions or eventualities - while the auxiliary elementary trees provide further information (constraints) added through adjoining.</Paragraph> <Paragraph position="4"> Importantly, we bar crossing structural connections. Thus one diagnostic for taking a predicate argument to be anaphoric rather than structural is whether it can derive from across a structural link.</Paragraph> <Paragraph position="5"> The relation in a subordinate clause is clearly structural: Given two relations, one realisable as &quot;Although oC/ \[3, the other realisable as &quot;Because y ~5&quot;, they cannot together be realised as &quot;Although ~ because y \[3 &&quot; with the same meaning as &quot;Although oC/ \[3. Because y 8&quot;. The same is true of certain relations whose realisation spans multiple sentences, such as ones realisable as &quot;On the one hand oz. On the other hand 13.&quot; and &quot;Not only T- But also &&quot; Together, they cannot be realised as &quot;On the one hand oC/. Not only T. On the other hand 13. But also &&quot; with the same meaning as in strict sequence. Thus we take such constructions to be structural as well (Webber and Joshi, 1998; Webber et al., 1999).</Paragraph> <Paragraph position="6"> On the other hand, the p-bearing adverb &quot;then&quot;, which asserts that one eventuality starts after the culmination of another, has only one of its arguments coming structurally. The other argument is presupposed and thus able to come from across a structural boundary, as in (1) a. On the one hand, John loves Barolo.</Paragraph> <Paragraph position="7"> b. So he ordered three cases of the '97.</Paragraph> <Paragraph position="8"> c. On the other hand, because he's broke, d. he then had to cancel the order.</Paragraph> <Paragraph position="9"> Here, &quot;then&quot; asserts that the &quot;cancelling&quot; event in (d) follows the ordering event in (b). Because the link to (b) crosses the structural link in the parallel construction, we take this argument to come non- null structurally through anaphoric presupposition. 3 Now we illustrate briefly how short discourses built from LTAG constituents get their semantics.</Paragraph> <Paragraph position="10"> For more detail, see (Webber and Joshi, 1998; Webber et al., 1999). For more information on compositional semantic operations on LTAG derivation trees, see (Joshi and Vijay-Shanker, 1999).</Paragraph> <Paragraph position="11"> (2) a. You shouldn't trust John because he never returns what he borrows.</Paragraph> <Paragraph position="12"> b. You shouldn't trust John. He never returns what he borrows.</Paragraph> <Paragraph position="13"> C. You shouldn't trust John because, for example, he never returns what he borrows.</Paragraph> <Paragraph position="14"> d. You shouldn't trust John. For example, he never retums what he borrows.</Paragraph> <Paragraph position="15"> Here A will stand for the LTAG parse tree for &quot;you shouldn't trust John&quot; and a, its derivation tree, and B will stand for the LTAG parse tree for &quot;he never returns what he borrows&quot; and 13, its derivation tree. The explanation of Example 2a is primarily structural. It involves an initial tree (y) anchored by &quot;because&quot; (Figure 2). Its derived tree comes from A substituting at the left-hand substitution site of y (index 1) and B at the right-hand substitution site (index 3). Semantically, the anchor of y (&quot;because&quot;) asserts that the situation associated with the argument indexed 3 (B) is the cause of that associated with the argument indexed 1 (A).</Paragraph> <Paragraph position="16"> The explanation of Example 2b is primarily structural as well. It employs an auxiliary tree (y) anchored by &quot;.&quot; (Figure 3). Its derived tree comes from B substituting at the right-hand substitution site (index 3) of ),, and &quot;f adjoining at the root of A (index 0). Semantically, adjoining B to A via y simply implies that B continues the description of the situation associated with A. The general inference that this stimulates leads to a defeasible contribution of causality between them, which can be denied without a contradiction - e.g.</Paragraph> <Paragraph position="17"> (3) You shouldn't trust John. He never returns what he borrows. But that's not why you shouldn't trust him.</Paragraph> <Paragraph position="18"> Presupposition comes into play in Example 2c.</Paragraph> <Paragraph position="19"> This example adds to the elements used in Ex- null to have the same temporal relation in the absence of &quot;then&quot; just shows that tense is indeed anaphoric and has no trouble crossing structural boundaries either.</Paragraph> <Paragraph position="20"> ample 2a, an auxiliary tree anchored by &quot;for example&quot; (8), which adjoins at the root of B (Figure 4). &quot;For example&quot; contributes both a presupposition and an assertion, as described in more detail in Section 3. Informally, &quot;for example&quot; presupposes a shared set of eventualities, and asserts that the eventuality associated with the clause it adjoins to, is a member of that set. In Example 2c, the set is licensed by &quot;because&quot; as the set of causes/reasons for the situation associated with its first argument.</Paragraph> <Paragraph position="21"> Thus, associated with the derivation of (2c) are the assertions that the situation associated with B is a cause for that associated with A and that the situation associated with B is one of a set of such causes.</Paragraph> <Paragraph position="22"> Finally, Example 2d adds to the elements used in Example 2b, the same auxiliary tree anchored by &quot;for example&quot; (~5). As in Example 2b, the causality relation between the interpretations of B and A comes defeasibly from general inference. Of interest then is how the presupposition of &quot;for example&quot; is licenced - that is, what provides the shared set or generalisation that the interpretation of B is asserted to exemplify. It appears to be licenced by the causal relation that has been inferred to hold between the eventualities denoted by B and A, yielding a set of causes/reasons for A.</Paragraph> <Paragraph position="23"> Thus, while we do not yet have a complete characterisation of how compositional semantics, defeasible inference and anaphoric presupposition interact, Examples 2c and 2d illustrate one significant feature: Both the interpretive contribution of a structural connective like &quot;because&quot; and the defeasible inference stimulated by adjoining can license the anaphoric presupposition of a p-bearing element like &quot;for example&quot;.</Paragraph> <Paragraph position="24"> Recently, Asher and Lascarides (1999) have described a version of Structured Discourse Representation Theory (SDRT) that also incorporates the semantic contributions of both presuppositions and assertions. In this enriched version of SDRT, a proposition can be linked to the previous discourse via multiple rhetorical relations such as background and defeasible consequence. While there are similarities between their approach and the one presented here, the two differ in significant ways: * Unlike in the current approach, Asher and Lascarides (1999) take all connections (of both asserted and presupposed material) to be structural attachments through rhetorical relations.</Paragraph> <Paragraph position="25"> The relevant rhetorical relation may be inher-</Paragraph> <Paragraph position="27"> ent in the p-bearing element (as with &quot;too&quot;) or it may have to be inferred.</Paragraph> <Paragraph position="28"> * Again unlike the current approach, all such attachments (of either asserted or presupposed material) are limited to the right frontier of the evolving SDRT structure.</Paragraph> <Paragraph position="29"> We illustrate these differences through Example 1 (repeated below), with the p-bearing element &quot;then&quot;, and Example 5, with the p-bearing element &quot;too&quot;. Both examples call into question the claim that material licensing presuppositions is constrained to the right frontier of the evolving discourse structure.</Paragraph> <Paragraph position="30"> (4) a. On the one hand, John loves Barolo.</Paragraph> <Paragraph position="31"> b. So he ordered three cases of the '97.</Paragraph> <Paragraph position="32"> c. On the other hand, because he's broke, d. he then had to cancel the order.</Paragraph> <Paragraph position="33"> (5) (a) I have two brothers. (b) John is a history major. (c) He likes water polo, (d) and he plays the drums. (e) Bill is in high school. (f) His main interest is drama. (g) He too studies history, (h) but he doesn't like it much.</Paragraph> <Paragraph position="34"> In Example 1, the presupposition of &quot;then&quot; in (d) is licensed by the eventuality evoked by (b), which would not be on the right frontier of any structural analysis. If &quot;too&quot; is taken to presuppose shared knowledge of a similar eventuality, then the &quot;too&quot; in Example 5(g) finds that eventuality in (b), which is also unlikely to be on the right frontier of any structural analysis. 4 With respect to the interpretation of &quot;too&quot;, Asher and Lascarides take it to presuppose a parallel rhetorical relation between the current clause and something on the right frontier. From this instantiated rhetorical relation, one then infers that the related eventualities are similar. But if the right frontier constraint is incorrect and the purpose of positing a rhetorical relation like parallel is to produce an assertion of similarity, then one might as well take &quot;too&quot; as directly presupposing shared knowledge of a similar eventuality, as we have done here. Thus, we suggest that the insights presented in (Asher and Lascarides, 1999) have a simpler explanation.</Paragraph> <Paragraph position="35"> Now, before embarking on more detailed analyses of two quite different p-bearing adverbs, we should clarify the scope of the current approach in terms of the range of p-bearing elements that can create non-structural discourse links.</Paragraph> <Paragraph position="36"> We believe that systematic study, perhaps starting with the 350 &quot;cue phrases&quot; given in (Knott, 1996, Appendix A), will show which of them use presupposition in realising discourse relations. It is likely that these might include: * temporal conjunctions and adverbial connectives presupposing an eventuality that stands in a particular temporal relation to the one currently in hand, such as &quot;then&quot;, &quot;later&quot;, &quot;meanwhile&quot;, &quot;afterwards&quot;, &quot;beforehand&quot;'; * adverbial connectives presupposing shared knowledge of a generalisation or set, such existing SDRT analysis in response to a p-bearing element, would seem superfluous if its only role is to re-structure the right frontier to support the claimed RF constraint.</Paragraph> <Paragraph position="37"> as &quot;for example&quot;, &quot;first...second...&quot;, &quot;for instance&quot;; null * adverbial connectives presupposing shared knowledge of an abstraction, such as &quot;more specifically&quot;, &quot;in particular&quot;; * adverbial connectives presupposing a complementary modal context, such as &quot;otherwise&quot;; * adverbial connectives presupposing an alternative to the current eventuality, such as &quot;instead&quot; and &quot;rather&quot;. 5 For this study, one might be able to use the structure-crossing test given in Section 2 to distinguish a relation whose arguments are both given structurally from a relation which has one of its arguments presupposed. (Since such a test won't distinguish p-bearing connectives such as &quot;meanwhile&quot; from non-relational adverbials such as &quot;at dawn&quot; and &quot;tonight&quot;, the latter will have to be excluded by other means, such as the (pre-theoretical) test for relational phrases given in (Knott, 1996).) * 3 For example We take &quot;For example, P&quot; to presuppose a quantified proposition G, and to assert that this proposition is a generalisation of the proposition rc expressed by the sentence P. (We will write generalisation(rt, G.) A precise definition of generalisation is not necessary for the purposes of this paper, and we will assume the following simple definition: proposition of the form Q I (x, a(x), b(x)); (ii) it allows the inference of a proposition G r of the form Q2 (x, a(x), b(x) ); and (iii) G' is inferrable from G (through having a weaker quantifier).</Paragraph> <Paragraph position="38"> The presupposed proposition G can be licensed in different ways, as the following examples show: (6) a. John likes many kinds of wine. For example, he likes Chianti.</Paragraph> <Paragraph position="39"> b. John must be feeling sick, because, for example, he hardly ate a thing at lunch.</Paragraph> <Paragraph position="40"> c. Because John was feeling sick, he did not for example go to work.</Paragraph> <Paragraph position="41"> d. Why don't we go to the National Gallery.</Paragraph> <Paragraph position="42"> Then, for example, we can go to the White House.</Paragraph> <Paragraph position="43"> Example 6a is straightforward, in that the presupposed generalisation &quot;John likes many kinds of wine&quot; is presented explicitly in the text. 6 In the remaining cases, the generalisation must be inferred. In Example 6b, &quot;because&quot; licenses the generalisation that many propositions support the proposi6Our definition of generalisation works as follows for this example: the proposition n introduced by &quot;for example&quot; is likes(john, chianti), the presupposed proposition G is many(x, wine(x),likes(john,x), and the weakened proposition G I is some(x, wine(x),likes(john,x). ~ allows G I to be inferred, and G also allows G ~ to be inferred, hence generalisation(rc, G) is true.</Paragraph> <Paragraph position="44"> tion that John must be feeling sick, while in Example 6c, it licences the generalisation that many propositions follow from his feeling sick. We can represent both generalisations using the meta-level predicate, evidence(rt, C), which holds iff a premise rc is evidence for a conclusion C.</Paragraph> <Paragraph position="45"> In Example 6d, the relevant generalisation involves possible worlds associated jointly with the modality of the first clause and &quot;then&quot; (Webber et al., 1999). For consistency, the semantic interpretation of the clause introduced by &quot;for example&quot; must make reference to the same modal base identified by the generalisation. There is more on modal bases in the next section.</Paragraph> </Section> class="xml-element"></Paper>