C Erler
7/5/2005 2:31:00 PM
>
> On 04/07/05, Abu Abdullah <alawi2@eim.ae> wrote:
>
> Ryan Leavengood wrote:
>
> > Nikolai Weibull wrote:
> >
> >> Paolo Capriotti wrote:
> >>
> >>
> >>> On 7/4/05, Abu Abdullah <alawi2@eim.ae> wrote:
> >>>
> >>>
> >>>> C:\>inference
> >>>> > all dogs are mamals
> >>>> OK
> >>>> > all mamals are hairy animals
> >>>> OK
> >>>> > all mamals are dogs
> >>>> OK ------------------ ? ? ?
> >>>
> >>
> >>
> >>> why not?
> >>
> >>
> >>
> >> A ? B doesn't imply B ? A,
> >> nikolai
> >>
> >
> > But here the rules are being set up, and the above relationship is
> > being defined for dogs and mammals. This is not a conflicting rule. I
> > think Abu sees it as a conflicting rule.
> >
> > Ryan
> >
> >
> using the quiz web page as a source for testing. and selecting from the
> available sample there, i selected the facts *2* , *1* and *6*
> respectivly.
> and after supplying the *6th* rule, the engine should say *"Sorry, that
> contradicts what I already know."*
You know that not all mammals are dogs. How does the computer ? Let's say I
enter :
All prime numbers are noncomposite.
All prime numbers are nonfactorable.
All noncomposites are prime numbers.
This is of the same form as the example given and would not be a
contradiction. The computer does not know anything about dogs or prime
numbers except what you tell it. All it can look at is the form of what
you've typed in. If the forms don't clearly indicate contradiction, the
computer does not know there is one.
Also, making the computer say "Contradiction" with that small amount of
information would mean that you could never effectively say that two names
(dogs and mammals or prime numbers and noncomposites) refer to the same
thing, even though that happens in real life.