This paper summarizes work over the past ve years on the automated planning and generation of multisentence texts using discourse structure relations, placing it in context of ongoing eeorts by Computational Linguists and Linguists to understand the structure of discourse. Based on a series of studies by the author and others, the paper describes how the orientation of generation toward communi...

There is a number of English or French verbs which take as semantic arguments eventualities or facts1. An example is the verb precede illustrated in (1a)2: its semantic arguments are the CL rise and the rise of blood leucocytes. The content conveyed by (1a) can be alternatively expressed in (1b); in other words, (1a) and (1b) are periphrastic. In (1b), the temporal relation between the CL and l...

Relations that hold between discourse segments can, but need not, be made explicit by means of discourse connectives. Even though the explicit signaling of discourse relations is optional, not all relations can be easily conveyed implicitly. It has been proposed that readers and listeners have certain expectations about discourse and that discourse relations that are in line with these expectat...

In this paper we explore a new theory of discourse structure that stresses the role of purpose and processing in discourse. In this theory, discourse structure is composed of three separate but interrelated components: the structure of the sequence of utterances (called the linguistic structure), a structure of purposes (called the intentional structure), and the state of focus of attention (ca...

This paper describes a new discourse module within our multilingual NLP system. Because of its unique data-driven architecture, the discourse module is language-independent. Moreover, the use of hierarchically organized multiple knowledge sources makes the module robust and trainable using discourse-tagged corpora. Separating discourse phenomena from knowledge sources makes the discourse module...

The overall sentiment of a text is critically affected by its discourse structure. By splitting a text into text spans with different discourse relations, we automatically train the weights of different relations in accordance with their importance, and then make use of discourse structure knowledge to improve sentiment classification. In this paper, we utilize explicit connectives to predict d...

This paper investigates the complexity of dependencies at the discourse level, in particular the dependencies between discourse connectives and their arguments. Our study is based on data from the Penn Discourse Treebank (PDTB) and is therefore an exploration into the ways treebanks can inform linguistic issues. We observe that, unlike in syntax, there is more uncertainty and flexibility with r...