Situation Entities and Discourse Modes

Alexis Palmer

As readers we are sensitive to various levels of structure within texts. The line of research described in this talk addresses two levels: the contribution of individual clauses and that of an intermediate level of text structure -- passages of several sentences or more. We follow Smith (2003) in discussing the Discourse Modes of individual text passages. Smith introduces the five modes of Narrative, Description, Report, Information, and Argument/Commentary and further identifies specific linguistic correlates of these different modes. One group of correlates points to the contributions made to the discourse by individual clauses of text. As a first step toward Discourse Mode classification, we address the problem of identifying and classifying references to situations expressed in written English texts. The situation entity (SE) classification task as construed here is the classification of clauses according to the type of entity they introduce to the discourse. The classification scheme we use includes, among others, events, states, abstract entities, and generic sentences.

In this talk I will discuss ongoing work on SE annotation and classification. We find that a feature-driven approach to annotation both improves annotation consistency and enriches the annotated data with interesting and useful semantic information (Friedrich and Palmer 2014a). One feature relevant for SE classification is the inherent lexical aspect of verbs -- whether verbs are stative or dynamic. I will present recent work on automatically classifying inherent lexical aspect (Friedrich and Palmer 2014b). Finally, I'll say something about plans to extend this work to German.

Note: this talk presents joint work with Annemarie Friedrich (Saarland University)

References

Carlota Smith. 2003. Modes of Discourse. Cambridge University Press.
Annemarie Friedrich and Alexis Palmer. Situation entity annotation. Linguistic Annotation Workshop, 2014 (a).
Annemarie Friedrich and Alexis Palmer. Automatic prediction of aspectual class of verbs in context. ACL 2014 (b).