I am looking for a spoken natural language corpus which has the
following properties:
1. The stressed words are marked.
2. The text has least 20.000 words.
3. Since we want to investigate syntax, there should be some non-trivial
syntactic structure in the text.
4. Preferred language is English.
5. The text is tagged.
We want to use the corpus for connectionist experiments on natural
language (synatx) acquisition of children.
Can someone tell me how I could obtain a corpus of the described type?
Thanks
Stephan
____________________________________________________________
Stephan Chalup e-mail:chalup@fit.qut.edu.au
MLRC, Neurocomputing Laboratory, School of Computing Science
FIT, QUT, 2 George Street, Brisbane Qld 4001, Australia
ph:+61 (0)7 3864 2575 http://sky.fit.qut.edu.au/~chalup/
____________________________________________________________