[Paper Review] Marking-up multiple views of a Text: Discourse and Reference

Dan Cristea, Nancy Ide+1 more|ArXiv.org|Sep 15, 2009

Multi-Agent Systems and Negotiation13 references3 citations

TL;DR

This paper presents a TEI and CES-based encoding scheme for representing multiple discourse and reference views in a single text, enabling structured analysis of discourse cohesion. It integrates a tool architecture that supports Veins Theory, demonstrating how multi-view markup enhances discourse analysis through standardized, extensible markup for linguistic annotation.

ABSTRACT

We describe an encoding scheme for discourse structure and reference, based on the TEI Guidelines and the recommendations of the Corpus Encoding Specification (CES). A central feature of the scheme is a CES-based data architecture enabling the encoding of and access to multiple views of a marked-up document. We describe a tool architecture that supports the encoding scheme, and then show how we have used the encoding scheme and the tools to perform a discourse analytic task in support of a model of global discourse cohesion called Veins Theory (Cristea & Ide, 1998).

Motivation & Objective

• To develop a standardized, extensible method for encoding multiple discourse and reference perspectives in a single text.
• To support discourse analytic tasks by enabling structured access to various views of textual structure.
• To integrate the encoding scheme with a tool architecture for practical annotation and analysis.
• To validate the approach through application to Veins Theory, a model of global discourse cohesion.
• To demonstrate the feasibility of multi-view markup in linguistic corpora using TEI and CES guidelines.

Proposed method

• The authors design a data architecture based on the Corpus Encoding Specification (CES) to represent multiple views of discourse and reference in a single document.
• They extend the Text Encoding Initiative (TEI) guidelines to support structured markup of discourse relations and referential chains.
• A dedicated tool architecture is implemented to facilitate the creation, storage, and querying of multi-view annotations.
• The approach enables independent but co-occurring encoding of discourse structure and reference tracking within the same document instance.
• The system supports versioning and modular annotation, allowing different analysts to work on distinct views without conflict.
• The framework is evaluated through its application to discourse cohesion analysis using Veins Theory as a theoretical model.

Experimental results

Research questions

• RQ1 How can multiple, co-existing views of discourse and reference be encoded within a single text representation?
• RQ2 What architectural principles enable extensible and interoperable markup of linguistic annotations across different discourse dimensions?
• RQ3 How can a tool architecture support the creation and querying of multi-view discourse annotations?
• RQ4 To what extent can this approach enhance the analysis of global discourse cohesion?
• RQ5 Can the integration of TEI and CES standards enable scalable and reusable discourse markup in linguistic corpora?

Key findings

• The proposed encoding scheme successfully supports the representation of multiple, independent views of discourse and reference within a single text instance.
• The CES-based data architecture enables structured, modular, and versioned annotation of discourse elements.
• The tool architecture facilitates collaborative and extensible annotation, supporting complex discourse analysis tasks.
• The approach was effectively applied to analyze discourse cohesion using Veins Theory, demonstrating its utility in theoretical discourse modeling.
• The integration of TEI and CES provides a robust, standards-compliant foundation for multi-view linguistic annotation.
• The system enables fine-grained access to discourse structure and reference relations, supporting advanced linguistic analysis.