[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Combinatorial Categorial Grammar / Re: Question about grad schools.



On 4/16/02 at 9:20 AM, noelwelsh@yahoo.com (Noel Welsh) wrote:

> Edinburgh is the birthplace of ML and they still do a
> lot of ML work.  It's also the top rated CS department
> in the most recent UK research assessment.

    If you are interested in functional programming and computational
linguistics, you might want to look at the work of Mark Steedman, Professor of
Cognitive Science in the Division of Informatics at the University of Edinburgh,
formally of The University of Pennsylvania.

    His natural language processing formalism of choice is Combinatory
Categorial Grammar which directly maps natural language surface structures to an
underlying Logical Form composed of functional applications.

    So one sees derivations like:
    
    Keats               eats                                     apples
    NP3SM:keats'        (S\NP3S)/NP:lambda x.lambda y.eat'xy     NP3PL:apples'
    ------> T                                                    --------<T
    T/(T\NP3SM)                                                  T\(T/NP3PL)
     :lambda f.f keats'                                           : lambda f.f
                                                                  apples'
                                                                  
                        --------------------------------------------------<
                           S\NP3S: lambda y. eat' apples' y
                           
    ---------------------------------------------------------------------->
           S: eat' apples' keats'
          
    The notation takes a bit of getting used to and can't be readily explained
herein, but the parsing approach is rather intriguing and might give light
weight language designers a few ideas for sweetening their syntax without
watering down their intellectual rigour.

    For more see: "Surface Structure and Interpretation" ISBN 0-262-69193-0 and
"The Syntactic Process" ISBN 0-262-19420-1.

Regards,

Peter