Combining documentary linguistics and corpus phonetics to advance corpus-based typology

Abstract

This article argues that documentary linguistics and corpus phonetics can form a happy marriage in that corpora extracted from language documentation collections contain highly relevant data that can advance corpus phonetics by enabling broad comparative studies. To make this point, this article reviews previous research on phonetic lengthening at utterance boundaries and pause probabilities before nouns and verbs in ten languages. I then introduce the DoReCo initiative, which, based on experience gained from these studies, builds a database of time-aligned corpora from documentary collections of 50 languages for corpus phonetic research and other research purposes.