Chapter 12. Annotating Texts for Language Documentation with Discourse Profiler’s Metatagging System

Abstract

This chapter introduces a systematic and robust way to annotate (or ‘tag’) texts with discourse information. To date there has not been a method for annotating texts for language documentation with discourse-text information. This is the first paper to systematically describe the capabilities and the annotating methodology of the Discourse Profiler’s metatagging system as a means of annotating endangered languages’ texts in a Toolbox database. Since there is a division of labor between Toolbox and Discourse Profiler, the Toolbox database can be the basis for the archival tasks, whereas the Discourse Profiler software is a computer assisted discourse-text analytical tool that mines the Toolbox discourse-text annotated database in order to produce two primary capabilities: 1) to create a representative interactive compressed representation or ‘map’ of the structure and elements of a text, and 2) to quantify texts based on this special metatagging system with an array of sixteen different possible statistical outputs (including both referential distance and topic persistence statistics). Although the main focus of this chapter is on the multipurpose annotation system, I will introduce the basics of the Discourse Profiler software in order to illustrate the range of analytical possibilities that this annotation system incorporates.