File Information
File: 05-lr/acl_arc_1_sum/cleansed_text/xml_by_section/evalu/86/p86-1029_evalu.xml
Size: 2,155 bytes
Last Modified: 2025-10-06 14:00:04
<?xml version="1.0" standalone="yes"?> <Paper uid="P86-1029"> <Title>DONNELLAN'S DISTINCTION AND A COMPUTATIONAL MODEL OF REFERENCE</Title> <Section position="9" start_page="188" end_page="188" type="evalu"> <SectionTitle> INTERACTIONS </SectionTitle> <Paragraph position="0"> When a speaker plans a speech act involving reference to an object, he must determine whether or not a conversationally relevant description is needed. However, the nature of the individuating set, on the one hand, and constraints on choices of referring expressions, on the other, may influence the speaker's planning in various ways. For example, if the individuating set contains only one item, say, the shortest spy, the definite description &quot;the shortest spy&quot; must be conversationally relevant. This is true both on formal and pragmatic grounds. From a formal standpoint, the description is conversationally relevant by default: no other functionally relevant description can be substituted because no such description is available. From a pragmatic standpoint, the description &quot;the shortest spy&quot; is very likely to be conversationally relevant in real discourse, simply because all we know about the referent is that he is the shortest spy. Thus, whatever we may have to say about that person is very likely to be related to the few facts contained in the description.</Paragraph> <Paragraph position="1"> Even if it is clear that a conversationally relevza~t description is needed for the speech act to succeed, constraints on choices of referring expressions may prevent the speaker from using this description. One such constraint results from the need to identify the referent for the hearer. If the conversationally relevant description is not suited for identification, a conflict arises. For example, in &quot;John believes Smith's murderer to be insane,&quot; the speaker may be trying simultaneously to represent the content of John's belief and to identify for the hearer whom the belief is about. Sometimes it is impossible to accomplish both goals with one and the same description.</Paragraph> </Section> class="xml-element"></Paper>