Modelling acquisition of verb semantics

[I.Bhula at cs.bham.ac.uk]

Dear list members, For my PhD I am looking at constructing a 
psychologically plausible computational model of how children acquire the 
"root meanings" and subcategorisation frames of simple motion verbs (e.g. 
push, slide, roll, hit) using visual data and linguistic input.

I would be immensely grateful if anyone could suggest a corpus or other 
data source which is specifically targeted at documenting verb acquisition, 
or that provides correlated visual (e.g actual video of parent-child 
interactions, or some kind of relevant visual stimulus used in a lab 
setting) and linguistic data involving such simple verbs; both of these 
would make great starting points. I have been looking around and now 
suspect I may have to create my own data and video clips, which will take 
rather more time than I have!

Any other suggestions or recommendations for this work would be much 
appreciated too.

Thanks for reading,
IB