Lynne Cahill, Natural Languages Technology Group, has been awarded an ESRC grant entitled Syntactic subcategorisation in multilingual inheritance lexicons.
PolySyn: Syntactic subcategorisation in multilingual inheritance lexicons.
The PolyLex lexicons are a set of hierarchically structured lexicons containing the 3000 most frequent words of English, German and Dutch. They differ fro many multilingual lexicons principally in the way they permit sharing of information at levels other than semantics (meaning). That is, most multilingual lexicons are linked only at the level of meaning, while the PolyLex lexicons allow words that have a similar form in different languages also to be linked, even if they don't have the same meaning.
The lexicons currently contain phonological (sound) and morphological (word structure) information and orthography (spelling) is currently being added to the lexicons (PolyOrth: ESRC grant RES-000-22-1485). The only syntactic (grammatical) information that is currently included is simply whether a word is a noun, verb etc.
The new project will develop a pilot experiment conducted during the original PolyLex project to add more detailed information about the grammatical behaviour of the words. For example, it will include the information that, while the verb "walk" in English is typically an intransitive verb ( it takes no object, as in "I walk"), the verb "give" is typically a di-transitive verb (it takes two objects as in "I give the book to John").