Finding schwa: Comparing the results of an automatic aligner with human judgments when identifying schwa in a corpus of spoken French

Authors

  • Peter M. Milne Dept. of Linguistics, University of Ottawa, 70 Laurier Ave East, ON K1N 6N5, Canada

Keywords:

Linguistics, Empirical data, Human judgments, Labor intensive, Linguistic data, Natural languages, Test research

Abstract

The article compares the results of an automatic aligner with human judgments when identifying schwa in a corpus of spoken French. The value in working with natural language corpora is the ability to collect large volumes of empirical data with which to test research hypotheses. The challenge is to generate these data quickly and accurately. Accumulating the linguistic data required to test and evaluate hypotheses can be a time consuming and labor intensive job. All data was systematically coded for presence or absence of schwa by trained researchers. The data was also time aligned at both the word and phone level by a forced aligner. The results of the two methods of coding were statistically compared to determine their level of agreement. Results show a significant correlation between the two methods and a high likelihood of overall agreement. Possible effects of dialect or phonetic context were investigated using a two-way, between subjects analysis of variance.

Additional Files

Published

2011-09-01

How to Cite

1.
Milne PM. Finding schwa: Comparing the results of an automatic aligner with human judgments when identifying schwa in a corpus of spoken French. Canadian Acoustics [Internet]. 2011 Sep. 1 [cited 2025 Feb. 13];39(3):190-1. Available from: https://jcaa.caa-aca.ca/index.php/jcaa/article/view/2475

Issue

Section

Proceedings of the Acoustics Week in Canada