Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: SUO: Re: W3C approves RDF and OWL




John,
   There are many ways to map language to logic.  I've implemented (with a 
colleague) one that is described at 
<http://home.earthlink.net/~adampease/professional/Pease-NLPKE.pdf>, and 
uses SUMO in its output.  I applaud your efforts to implement a different 
system, even though details of our approaches differ.

Adam

At 02:22 PM 2/17/2004 -0500, John F. Sowa wrote:

>Adam,
>
>Every relational noun in English -- such as parent, employee,
>or client -- designates two things:  a relational pattern and
>a class of entities that fit the pattern.
>
>A simple way to generate the corresponding representation in
>FOL is to use declarations such as those that I proposed for
>CLCE.  Following is an example:
>
>Declare object as noun from SQL("SELECT ID FROM OBJECTS"),
>         shape as functional noun (x1 shape of x2)
>            from SQL("SELECT SHAPE, ID FROM OBJECTS"),
>         color as functional noun (x1 color of x2)
>            from SQL("SELECT COLOR, ID FROM OBJECTS"),
>         supporter as noun (x1 supporter of x2)
>            with relation(support)
>            from SQL("SELECT * FROM SUPPORTS"),
>         supportee as noun (x2 supportee of x1)
>            with relation(support);
>
>The first line of this declaration introduces "object" as
>a _categorial noun_, which is represented by either a type
>label in typed logic or a monadic predicate in an untyped
>logic.
>
>The next 4 lines define "color" and "shape" as functional
>nouns (a special case of relational nouns), which can be
>represented as either a dyadic relation or as a monadic
>function, depending on which kind of notation you prefer.
>But in addition to the relation, they also create type labels
>or monadic predicates for "color" and "shape" when they are
>used as categorial nouns without the preposition "of".
>
>The last 5 lines define "supporter" and "supportee" as
>relational nouns, which both map to the same dyadic relation
>named "support", which is mapped to an SQL table named
>"supports".  Note that the arguments of "supportee" are
>mapped to the relation "support" in the reverse order of
>the argument mapping of "supporter".
>
>In addition to the mappings to the relation named support,
>the two nouns "supporter" and "supportee" also correpond to
>type labels or monadic relations for the classes of objects
>that happen to be supporters or supportees.
>
>All of this is implicit knowledge that is familiar to every
>English speaker, and all the information needed for the
>CLCE translator to generate a representation in FOL can be
>derived from the above declaration.  In this case, each
>relational noun maps to a dyadic relation and to a monadic
>relation (or type label).  The monadic relation can always
>be derived from the dyadic relation by a simple mapping:
>
>   supportee(x) = (Ey)support(y,x)
>
>   supporter(x) = (Ey)support(x,y)
>
>For further discussion of this example, see Section 2 of
>the CLCE specification:
>
>    http://www.jfsowa.com/clce/specs.htm
>
>John