Title

Annotation Driven Concordancing: The PAX Toolkit

Authors

Thorsten Trippel (Universität Bielefeld)

Dafydd Gibbon (Universität Bielefeld)

Session

SP3 Annotation Tools: From Speech Segments To Dialogues

Abstract

We describe PAX, "Portable Audio Concordance System", a proof-of-concept prototype of a multipurpose, multilingual audio concordance toolkit. The primary goal is to support efficient grammar and lexicon construction in the documentation of unwritten languages; languages currently included are Ega, Anyi, and Koulango (Ivory Coast), additional samples in German and English. The approach combines methods from corpus linguistics, annotation theory and practice, phonetics and lexicography.

Keywords

Annotation, XML, SAMPA, Spoken language, Concordancing, Corpus, Lexicons, Multimodal systems

Full Paper

171.pdf