Kurt Eberle

On representing the temporal structure of a natural language text

A proposal to deal with tenses in the framework of Discourse Representation Theory is presented, as it has been implemented for a fragment at the IMS for the project LILOG. It is based on the theory of tenses of H. Kamp and Ch. Rohrer. The system uses the tense and aspect information, the information about the temporal discourse structure of the preceding text stored in a specific list of possible reference times, and background knowledge. These types of information interact in order to choose a suited temporal anchor for the event of a new sentence.With respect to extended texts, choosing the right reference time for a new event is a problem which has been largely neglected in the literature.

Disambiguation by information structure in DRT

Text understanding and high quality machine translation often necessitate the disambiguation of ambigous structures or lexical elements. Drawing inferences from the context can be a means for resolving semantic ambiguities. However, often, this is an expensive strategy that, in addition, not always comes up with a clear preference for one of the alternatives. In this paper, we argue that in a number of cases deep semantic analyses can be avoided by taking into account the constraints that the alternative readings impose onto the information structure. To this end, we present a study of the ambigous German adverb erst and point out the particular circumstances under which the given information structure disambiguates the adverb without further semantic analysis.

Eventualities in a natural language understanding system

This paper focuses on the discussion of suitable representations of eventualities in a formal language and of the possibilities to draw inferences from representations. We argue in favour of the treatment of eventualities as individuals, which are structured along different lines. When reifying eventualities, there are different possibilities of individualization. This is similarily true for the domain of objects. Thus we investigate these possibilities in parallel with objects, and obtain a rather symmetric structuring of the domain of individuals, i.e. a sort hierarchy which is sensible for different kinds of eventualities and objects respectively.

Partial orderings and Aktionsarten in discourse representation theory

This paper presents an approach to deal with the underspecification of Aktionsarten in German sentences. In German the difference between an accomplishment and the associated progressive state is often not marked on the sentence level. This distinction is important for correctly interpreting texts and for translation into languages which provide morphological markings of Aktionsarten. To maintain compositionality we suggest a two-step analysis of a text with respect to the temporal relations and the classification as events or states. This analysis is guided by the Discourse Representation Theory developed by Kamp and makes use of world knowledge and an inference component.The problem of classification can be reformulated as the problem of finding an embedding function f from the representational entities onto the domain of a model. The models we use are structures built from intervals of time, events and individuals. Considering intensional models of this type will allow us to give truth-conditions for progressive states related to corresponding accomplishments. We restrict ourselves to progressive states of intentional actions and use the beliefs of the agent.

Point-Based Descriptions of Interval Relations

Temporal text understanding and question answering presupposes the availability of a component which computes the transitive closure of a set of statements about interval relations. Vilain and Kautz have addressed the question of computing the transitive closure by translating such statements into equivalent point-based descriptions. The problem is that there are basic statements about interval relations which have to be translated into disjunctions of point-based descriptions. Therefore, in order to handle the general case economically, it is necessary to have translations with a minimum number of disjuncts for each and every case. In this paper we present an algorithm which computes a best translation, i.e. a translation which consists of a minimal number of disjuncts. This number amounts at most to 5.