ALAN'S WEB PAGES

Note, this site is still under constructions and many of the links won't work

I am working with John Barnden and Mark Lee on modelling metaphor. See the ATT-Meta page for a description of the project.

However, I have also worked and dabbled in a number of other areas that broadly involve language.

• My thesis, from the Centre for Computational Linguistics at UMIST, was entitled Arguments and the Resolution of Local Syntactic Ambiguity. I took the psycholinguistic data concerning how humans resolve structural ambiguities and investigated whether it could be given a linguistic or a computational-linguistic characterization. A description of my thesis can be read here.
• After completing my thesis I worked in the Experimental Psychology Laboratory at the University of Sussex, in broadly the same area, but with an emphasis on the role reference resolution plays in ambiguity resolution.
• I stayed at Sussex, but switched department to Cognitive and Computer Science where I worked on a project that used the search technique of simulated annealing to seek out the best tree structure that could be assigned to examples of real text.
• Another shift of direction took me back to the Centre for Computational Linguistics at UMIST, where I worked on the 'Trans-European Employment' or TREE project in collaboration with a number of industrial and academic partners. Our role was to adapt a 'Example-Based Machine Translation' tool for the purpose of extracting information from job ads.
• I stayed on at UMIST to work on the Parasite project.
• Most recently, I have worked at Sail Labs in Munich.

My c.v can be found here.

A list of conference papers and publications can be found here.

If I had more time I would pursue the idea of treating a 'combinatory categorial grammar' as a partial specification of a tree; a process I call dendrification. This appears to solve certain problems having to do with parsing with combinatory categorial grammars, and also imposes constraints -probably far too many- on the form a category can take, which may be linguistically interesting. This is a hobby and not an area of expertise. If there are any experts in the area who would like to discuss the issue, please contact me.