"The Development of an Automatic Parsing System Using the ICE Corpus as Linguistic Knowledge Base"

Researchers: 
Alex Fang and Sean Wallis 

Started: 
1st October 1995 (AF), 1st November 1995 (SW) 

2 years duration

The research is the development of a computer system that automatically parses the syntactic structure of English. The approach relies on linguistic knowledge extracted inductively from the ICE Corpus, consisting of syntactic rules which are then applied to new data. This is then complemented by a lexical database to ease decisions for situations where these inductive rules appear problematic. A combination of approaches can make parse decisions at different levels. Explicit lexical rules may not be empirically trustworthy. Induced implicit rules are incapable of semantic differentiation without lexical typing. Our natural language systems disambiguate with a variety of sources of knowledge. The proposed system is inspired by that intuition.

The Survey Parser will shortly be available for academic researchers in a package together with the tagger AUTASYS and the lexical database mentioned above.