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 interre-lated components: the structure of the sequence of utterances (called the linguistic structure), a struc-ture of purposes (called the intentional structure), and the state of focus of attention (called the attentional state). The linguistic structure consists of segments of the discourse into which the utter-ances naturally aggregate. The intentional structure captures the discourse-relevant purposes, expressed in each of the linguistic segments as well as relationships among them. The attentional state is an abstraction of the focus of attention of the participants as the discourse unfolds. The attentional state, being dynamic, records the objects, properties, and relations that are salient at each point of the discourse. The distinction among these components is essential to provide an adequate explanation of such discourse phenomena as cue phrases, referring expressions, and interruptions. The theory of attention, intention, and aggregation of utterances is illustrated in the paper with a number of example discourses. Various properties of discourse are described, and explanations for the behavior of cue phrases, referring expressions, and interruptions are explored. This theory provides a framework for describing the processing of utterances in a discourse. Discourse processing requires recognizing how the utterances of the discourse aggregate into segments, recognizing the intentions expressed in the discourse and the relationships among intentions, and track-ing the discourse through the operation of the mechanisms associated with attentional state. This processing description specifies in these recognition tasks the role of information from the discourse and from the participants ' knowledge of the domain. 1

TextTiling is a technique for subdividing texts into multi-paragraph units that represent passages, or subtopics. The discourse cues for identifying major subtopic shifts are patterns of lexical co-occurrence and distribution. The algorithm is fully implemented and is shown to produce segmentation that corresponds well to human judgments of the subtopic boundaries of 12 texts. Multi-paragraph subtopic segmentation should be useful for many text analysis tasks, including information retrieval and summarization. 1.

This paper describes TextTiling, an algorithm for partitioning expository texts into coherent multi-paragraph discourse units which reflect the subtopic structure of the texts. The algorithm uses domain-independent lexical frequency and distribution information to recognize the interactions of multiple simultaneous themes.

This article reviews research on the use of situation models in lnguage comprehension and memory retrieval over the past 15 years. Situation models are integrated mental representations of a described state of affairs. Significant progress has been made in the scientific understanding of how situation models are involved in language comprehension and memory retrieval. Much of this research focuses on establishing the existence of situation models, often by using tasks that assess one dimension of a situation model. However, the authors argue that the time has now come for researchers to begin to take the multidimensionality of situation models seriously. The authors offer a theoretical framework and some methodological observations that may help researchers to tackle this issue.

This paper describes a method for linear text segmentation that is more accurate or at least as accurate as state-of-the-art methods (Utiyama and Isahara, 2001

We propose a method for segmentation of expository texts based on hierarchical agglomerative clustering. The method uses paragraphs as the basic segments for identifying hierarchical discourse structure in the text, applying lexical similarity between them as the proximity test. Linear segmentation can be induced from the identified structure through application of two simple rules. However the hierarchy can be used also for intelligent exploration of the text. The proposed segmentation algorithm is evaluated against an accepted linear segmentation method and shows comparable results.

Speakers often repeat the first word of major constituents, as in, ‘‘I uh I wouldn’t be surprised at that.’ ’ Repeats like this divide into four stages: an initial commitment to the constituent (with ‘‘I’’); the suspension of speech; a hiatus in speaking (filled with ‘‘uh’’); and a restart of the constituent (‘‘I wouldn’t...’’). An analysis of all repeated articles and pronouns in two large corpora of spontaneous speech shows that the four stages reflect different principles. Speakers are more likely to make a premature commitment, immediately suspending their speech, as both the local constituent and the constituent containing it become more complex. They plan some of these suspensions from the start as preliminary commitments to what they are about to say. And they are more likely to restart a constituent the more their stopping has disrupted its delivery. We argue that the principles governing these stages are general and not specific to repeats.

We present a computational theory of the paragraph. Within it we formally define coherence, give semantics to the adversative conjunction "but " and to the Gricean maxim of quantity, and present some new methods for anaphora resolution. The theory precisely characterizes the relationship between the content of the paragraph and background knowledge needed for its understanding. This is achieved by introducing a new type of logical theory consisting of an object level, corresponding to the content of the paragraph, a referential level, which is a new logical level encoding background knowledge, and a metalevel containing constraints on models of discourse (e.g. a formal version of Gricean maxims). We propose also specific mechanisms of interaction between these levels, resembling both classical provability and abduction. Paragraphs are then represented by a class of structures called p-models. 1.

The proposal examined here is that speakers use uh and um to announce that they are initiating what they expect to be a minor (uh), or major (um), delay in speaking. Speakers can use these announcements in turn to implicate, for example, that they are searching for a word, are deciding what to say next, want to keep the floor, or want to cede the floor. Evidence for the proposal comes from several large corpora of spontaneous speech. The evidence shows that speakers monitor their speech plans for upcoming delays worthy of comment. When they discover such a delay, they formulate where and how to suspend speaking, which item to produce (uh or um), whether to attach it as a clitic onto the previous word (as in "and-uh"), and whether to prolong it. The argument is that uh and um are conventional English words, and speakers plan for, formulate, and produce them just as they would any word.

We consider the problem of constructing a directed acyclic graph that encodes temporal relations found in a text. The unit of our analysis is a temporal segment, a fragment of text that maintains temporal coherence. The strength of our approach lies in its ability to simultaneously optimize pairwise ordering preferences and global constraints on the graph topology. Our learning method achieves 83 % F-measure in temporal segmentation and 84 % accuracy in inferring temporal relations between two segments. 1