| Bookmark: |
Full-text
| URL: http://faculty.washington.edu/bdb/ECAL99.pdf |
| Cached: PDF-300K |
| SAVE AS an easy-to-recall long filename: |
| Filename format: author--year--title PDF-300K |
| Filename format: author--year--title--journal|proceedings|...--pages PDF-300K |
Related links
| CiteSeer: http://citeseer.ist.psu.edu/deboer99emergence.html |
| Web search: Google Web Search :: Google Scholar |
| Within this site: Cited by (11) References (21) |
Paper at a Glance
BibTexEmergence of Speech Sounds in Changing Populations Bart de Boer, Paul Vogt AILab Vrije Universiteit Brussel Pleinlaan 2 1050 Brussel {bartb,paul}@arti.vub.ac.beAbstract: This paper shows that realistic and coherent vowel systems can emerge from scratch in a population of agents that imitate each other under hu manlike constraints of production and perception. The simulation is extended so that populations can change; old agents can be removed, and new agents can be added. In these circumstances vowel systems can also emerge and be pre served. It is shown that sometimes an age structure in the population can im prove preservation of the vowel systems.
1 Introduction The study of human languages is not only concerned with describing the huge variety of phenomena that are encountered in human languages, but also with finding expla nations for similar phenomena that recur in languages that are neither related histori cally nor geographically. Such phenomena are called language universals, or, because there are always exceptions, universals tendencies. This paper is concerned with ex plaining the universal tendencies of human vowel systems. Although the human vocal tract is capable of producing a wide variety of different vowels (at least 45 different basic vowel qualities [10]) one finds that all human languages use only a very small subset of these. Furthermore, these subsets are not chosen randomly, but they exhibit remarkable regularities. When one looks at the vowel systems of widely different languages, such as, for example the 451 languages in the UPSID (UCLA Phonological Segment Inventory Database [12, 13]) one finds that the maximum number of vowel qualities in any language is 15 in Norwegian [10] while the mini mum number is 3 (although there are languages that are reported to have only two vowels [10]). One also finds that certain vowels, for example [i], [a] and [u] occur very frequently (in 87%, 87% and 82% of the languages in UPSID, respectively) ...
@inproceedings{deboer99emergenceOf,
author={B. de Boer and P. Vogt},
title={Emergence of speech sounds in changing populations},
year={1999},
pages={664-673},
editor={Floreano, D. and Nicoud, J-D and Mondada, F.},
publisher={Springer-Verlag},
booktitle={ECAL99},
url={http://www.isrl.uiuc.edu/~amag/langev/paper/deboer99emergenceOf.html}
}
| HOME :: Conference List :: Conference Paper | Comments to: junwang4 you-know-at gmail.com | Last update: 2/2/08 |