<?xml version="1.0" standalone="yes"?>
<Paper uid="W06-1408">
  <Title>Generating References to Parts of Recursively Structured Objects Helmut Horacek Universit t des Saarlandes</Title>
  <Section position="4" start_page="47" end_page="47" type="intro">
    <SectionTitle>
2 Previous Work
</SectionTitle>
    <Paragraph position="0"> Within this paper, we adopt Dale's terminology (1988). A referential description (Donellan 1966) serves the purpose of letting the hearer or reader identify a particular object or set of objects in a situation. Referring expressions to be generated are required to be distinguishing descriptions, that is, descriptions of the entities being referred to, but not to any other object in the context set. A context set is defined as the set of the entities the addressee is currently assumed to be attending to -- this is similar to the concept of focus spaces of the discourse focus stack in Grosz' &amp; Sidner's (1986) theory of discourse structure. Moreover, the contrast set (the set of potential distractors (McDonald 1981)) is defined to entail all elements of the context set except the intended referents.</Paragraph>
    <Paragraph position="1"> Generating referring expressions is pursued since the eighties (e.g., (Appelt 1985), among several others). Subsequent years were characterized by a debate about computational efficiency versus minimality of the elements appearing in the resulting referring expression (Dale 1988, Reiter 1990, and several others). In the mid-nineties, this debate seemed to be settled in favor of the incremental approach (Dale and Reiter 1995) -- motivated by results of psychological experiments (e.g., Levelt 1989), certain non-minimal expressions are tolerated in favor of adopting the fast strategy of incrementally selecting ambiguity-reducing attributes from a domain-dependent preference list. Complementary activities include the generation of vague descriptions (van Deemter, 2000) and extensions to multimodal expressions (Van der Sluis 2005). Recently, algorithms have also been developed to the identification of sets of objects rather than individuals (Bateman 1999, Stone 2000, Krahmer, v. Erk, and Verweg 2001), and the repertoire of descriptions has been extended to boolean combinations of attributes, including negations (van Deemter 2002). To avoid the generation of redundant descriptions what incremental approaches typically do, Gardent (2002) and Horacek (2003) proposed exhaustive resp. best-first searches.</Paragraph>
    <Paragraph position="2"> All these procedures more or less share the design of the knowledge base which bears influence on the descriptor selection. Objects are conceived as atomic entities, which can be described in terms of sets of attributes and relations to other objects. In such a setting, a structured object can be represented, among others, by a set of relations to its components, which are themselves conceived as objects. An exception to this method is the work by Paraboni and van Deemter (2002) who use hierarchical object representations to refer to parts of a book (figures, sections., etc.). Reference to such a component is made identifiable by iteratively adding a description of embedding structures until obtaining uniqueness. There are, however, no approaches addressing identification of objects or their components when the structures in these objects are of a recursive nature. Objects of this kind are mostly abstract ones, such as formulas, but also some sorts of geometric objects. Typical applications where such objects are prominent include scientifictechnical documentation and tutoring systems. As we will see in the next section, naturally observed references to such objects have a number of particularities which are not addressed by existing generation algorithms.</Paragraph>
  </Section>
class="xml-element"></Paper>