Abstract anaphora resolution in Danish

Abstract

In this paper 1 I describe the use of Danish pronouns and deictics in dialogues. Then I present an adaptation to Danish of Eckert and Strube's algorithm for resolving anaphora referring to individual NPs and abstract objects in English dialogues (Eckert and Strube, 1999b; Eckert and Strube, 1999a). The adapted algorithm is tested on four Danish dialogues from two dialogue collections and the results obtained are evaluated.

Focu, Activation, and This-Noun Phrases

“…What was missing was a way to identify references to abstract entities in the implicit focus. The problem we had to face was that while developing the GNOME scheme we had found-as others before us (Eckert and Strube, 2001;Navarretta, 2000)-that identifying the antecedents of 'discourse deictic' expressions in the broad sense (i.e., expressions referring to-typically, abstract-entities introduced in the discourse indirectly, such as propositions) is very hard, especially when the annotation produces something less than a full logical form in, say, the DRT sense (Kamp and Reyle, 1993). However, we had also found for the case of visual deixis that in some cases while identifying the antecedent of an expression is quite hard, classifying a NP as deictic is easier.…”

Annotating a broad range of anaphoric phenomena, in a variety of genres: the ARRAU Corpus

“…The term discourse deixis was introduced by Webber in [62] to indicate the reference to abstract entities which have not been introduced in the discourse through a nominal expression, 13 as in the following example from the RST TRAINS GNOME PEAR TOTAL 631 862 73 67 1633 Discourse deixis in its full form is a very complex form of reference, both to annotate [35,66] and to resolve [67]. Very few anaphoric annotation projects have attempted annotating discourse deixis in its entirety [35,66,68]; more typical is a partial annotation, as in the work of Byron and Navarretta, who annotated pronominal reference to abstract objects [69,70]; in ONTONOTES, where event anaphora was marked [47]; and in the work of Kolhatkar [68], that focused on so-called shell nouns. As a result, very few systems have attempted resolving this type of anaphors [71,72,73,67] Discourse deixis was one of the 'difficult cases of anaphora' on which the AR-RAU project focused, and a number of annotation experiments were conducted [35], resulting in guidelines according to which 1.…”

“…7.3 : so we ship one 7.4 : boxcar 7.5 : of oranges to Elmira 7.6 : and that takes another 2 hours Discourse deixis in its full form is a very complex form of reference, both to annotate (Artstein and Poesio 2006;Dipper and Zinsmeister 2012) and to resolve (Marasović et al 2017). Very few anaphoric annotation projects have attempted annotating discourse deixis in its entirety (Artstein and Poesio 2006;Dipper and Zinsmeister 2012;Kolhatkar 2014); more typical is a partial annotation, as in the work of Byron and Navarretta, who annotated pronominal reference to abstract objects (Byron and Allen 1998;Navarretta 2000); in ONTONOTES, where event anaphora was marked (Pradhan et al 2007); and in the work of Kolhatkar Kolhatkar (2014), which focused on so-called shell nouns. As a result, very few systems have attempted resolving this type of anaphors (Eckert and Strube 2000;Byron 2002;Kolhatkar and Hirst 2012;Marasović et al 2017) Discourse deixis was one of the "difficult cases of anaphora" on which the ARRAU project focused, and a number of annotation experiments were conducted (Artstein and Poesio 2006), resulting in guidelines according to which 1.…”

“…r Nissim et al's annotation of the Switchboard corpus (Nissim et al 2004) only provided the information status of mentions, not their antecedents if any. Navarretta had annotated pronominal anaphora in Danish and Italian dialog (Navarretta 2000) and Poesio et al had annotated Italian MapTask dialogs (Poesio et al 2004c).…”

Proceedings of the First Workshop on Computational Models of Reference, Anaphora and Coreference

“…Very few anaphoric annotation projects have attempted annotating discourse deixis in its entirety (Artstein and Poesio, 2006;Dipper and Zinsmeister, 2012). More typical is a partial annotation, as in (Byron and Allen, 1998;Navarretta, 2000), who annotated pronominal reference to abstract objects; in ONTONOTES, where event anaphora was marked (Pradhan et al, 2007b); and in the work of Kolhatkar (2014), that focused on so-called shell nouns. In ARRAU, 1.…”