Prediction Experiment for Western Kho-Bwa language data: dataset
Description
Prediction Experiment on Western Kho-Bwa languages
Timotheus A. Bodt (SOAS, London) and Johann-Mattis List (Max Planck Institute, Jena)
This database includes all the sound files and the transcriptions of the prediction experiment for Western Kho-Bwa. This experiment was registered as:
Bodt, Timotheus A., Nathan W. Hill and Johann-Mattis List. 2018. Prediction experiment for missing words in Kho-Bwa language data. Open Science Framework Preregistrations October 5. https://osf.io/evcbp/
The data and code can be found on:
Timotheus A. Bodt, Nathan W. Hill, & Johann-Mattis List. (2018, October 8). Prediction experiment for missing words in Kho-Bwa language data (Version v1.0.1). Zenodo. http://doi.org/10.5281/zenodo.1451176
A paper explaining the experiment is under review:
Bodt, Timotheus A. and Johann-Mattis List. 2019 (under review). Testing the predictive force of the comparative method: An ongoing experiment on unattested words in Western Kho-Bwa languages. Papers in Historical Phonology Volume 1: 1–21.
The results of the experiment will be presented at the International Conference on Historical Linguistics 24: 01-Jul-2019 - 05-Jul-2019, Canberra, Australia.
The uncut sound files, cut sound files, original field notes and preliminary transcriptions have been saved as:
Bodt, Timotheus Adrianus. (2019). 'Retrodiction' experiment Western Kho-Bwa languages: data [Data set]. Zenodo. http://doi.org/10.5281/zenodo.2529727
How to use these files?
- Download the zip folder soundfiles_prediction_experiment.zip
- Extract the files in a separate folder
- Search for the required sound file(s)
Searching sound files can best be done using the English CONCEPTS from the predictions_results.csv file. For example, searching for BACK will give all the sound files that contain the English gloss ‘back’ (including ‘backwards’, ‘back’ as body part, turn ‘back’ etc.).
Another option is the select all the sound files of a given linguistic variety / doculect by searching for the original sound file number.
I would advise against using a certain attested form in the predictions_results.csv file and search for that (e.g. p a ŋ + b u ‘chest’), because the cut sound files have been saved without spaces and morpheme breaks and because the actual transcriptions of the sound files may have changed after analysis, but were not updated in the name of the cut sound files.
If you cannot find a certain sound file, then it may simply not have been recorded or not cut from the main sound file. If you are really interested, please mail me at timintibet@hotmail.com and I will attempt to find it or record it.