[Attempto] inverse names of relations

[Nagarjuna G]

On Thu, Sep 1, 2011 at 7:06 PM, Norbert E. Fuchs <fuchs at ifi.uzh.ch> wrote:
>
> On 1 Sep 2011, at 10:00, Nagarjuna G wrote:
>
> >  ... How to tell that the predicate "part of" and "consists of" are inverse to each other.
>
>
> On Thu, Sep 1, 2011 at 2:25 PM, Norbert E. Fuchs <fuchs at ifi.uzh.ch> wrote:
>
> > Are "part of" and "consists of" really inverse to each other? Though an engine is a part of a car the car does not consist of the engine.
>
>
> On 1 Sep 2011, at 14:54, Nagarjuna G wrote:
>
> > I assumed them to be.  Aren't they?
>
>
> As my car/engine example shows they aren't in general. To express the exact inverse in English you would have to say something like
>
>  If an engine is a part of a car then the car consists only of the engine.
>
> ACE does not provide "only" but provides the special determiners "nobody but", "nothing but", and their variant "no ... but" that are used with bare plural noun phrases, bare mass noun phrases or proper names. Unfortunately these constructs do not allow for variables. Thus you cannot say
>
>  If X is a part of Y then Y consists-of nothing but X.
>
> You would need a more complex construct, for instance
>
>  If X is a part of Y then if Y consists-of Z then Z is X.
>

I do not have a usecase for the complex example you gave. But it is
instructive.

The example I have is: Given "Milk consists of fat, carbohydrates,
proteins, minerals and water", I should be able to say "Water is a
part of Milk" as a valid inference. Or if a query is raised "Is water
a part of Milk?" I expect yes.

(In some biology books "consists of' and "composed of" are used
interchangeably.)

I understood that an if-then construct can be used to tell RACE that a
relation is inverse.

Thanks again
--
GN