Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[protege-discussion] Jim Likes Dogs

Timothy Redmond tredmond at stanford.edu
Mon Jun 25 12:08:21 PDT 2012


On 6/25/12 11:26 AM, Jim Tivy wrote:
> Hi Timothy
>
> OK, it is a bit more clear.  You are right that there is a larger question
> of what I wish to express.
>
> I want to say Jim likes Dogs which means for any dog ever entered into the
> ontology Jim has an assumed likes property assertion eg: for two dogs: Jim
> likes Rover, Jim likes Blackie.
> I would imagine if I did it correctly I would not have to make these
> explicit property assertions as dog individuals are added to the ontology?

This is right.  We are near the issue of the difference between asserted 
and inferred information.  If you add the assertion that jim likes all 
Dogs and you create a Dog, fido, then tools (e.g. reasoners) will be 
able to deduce that jim likes fido.  The fact that jim likes fido is 
inferred because it was not originally explicitly stated in the ontology.


>> you could say (using the functional owl syntax this time):
>>
>>          SubClassOf(:Dogs ObjectHasValue(ObjectInverseOf(:likes) :jim))
> [Jim Tivy]
> This construct is the one I don't like.  I don't like Dogs being a subclass
> of likedBy jim.  Having a class that mixes in an individual jim and a
> property seems wrong.  But I concede it is expressible so "wrong" is a bit
> ambiguous.

Liking and not liking shouldn't really stop you from saying what you 
wanted to say.  If you have a more concise or prettier way of saying the 
same thing then that is good.  But I have been in situations where I 
needed to state exactly this type of thing and this is the best way I 
found to say this.

It is true that this is an unusual construct for an ontology. 
Oftentimes, people will view such an assertion as being about Dogs where 
as the original English language statement was clearly about jim.  
Protege also displays ontologies in this manner.  But the notion of 
being "about" Dogs is not well-defined.  And again, it should not stop 
you from saying what you want to say.


-Timothy



>
> Some comments below:
>
>> -----Original Message-----
>> From: protege-discussion-bounces at lists.stanford.edu [mailto:protege-
>> discussion-bounces at lists.stanford.edu] On Behalf Of Timothy Redmond
>> Sent: June-25-12 9:46 AM
>> To: protege-discussion at lists.stanford.edu
>> Subject: Re: [protege-discussion] Jim Likes Dogs
>>
>> On 6/25/12 9:13 AM, Jim Tivy wrote:
>>> [Jim Tivy] Nothing complex just the subject predicate object "Jim
>>> likes Dogs".
>>> ObjectPropertyAssertion( :likes :Jim :Dogs) or "Dog" singular if you
> like.
>> You can certainly state this.  OWL 2 allows punning so this is fine.
>> But such an assertion would not have anything to do with whether Jim likes
>> individuals in the Dogs class.  So to me such an axiom would not be a
>> representation of the English phrase "Jim likes Dogs".  By making the
>> assertion in this way you are deliberately avoiding stating a relationship
>> between Jim and individual Dogs.  If I saw such an ontology I would
>> understand it as OWL but be unsure of your modeling intention.
>>
>>> It should be clear from "Jim likes Dogs" that I refer to the
>>> collective
>>> (all) - but just
>>> to be more explicit I do mean all dogs even the ones that bite jim:).
>> But that is just what you are trying to avoid stating by using the object
>> property assertion.  Again if you want to state that Jim likes all Dogs
> then
>> you could say (using the functional owl syntax this time):
>>
>>          SubClassOf(:Dogs ObjectHasValue(ObjectInverseOf(:likes) :jim))
> [Jim Tivy]
> This construct is the one I don't like.  I don't like Dogs being a subclass
> of likedBy jim.  Having a class that mixes in an individual jim and a
> property seems wrong.  But I concede it is expressible so "wrong" is a bit
> ambiguous.
>>
>>
>>> In OWL2 what are you saying here - is this a ClassAssertion a Property
> or...
>>>
>>> Individual: Jim
>>>        Types:
>>>            likes min 2 Dog
>> He is using the Manchester OWL syntax as I was.  It is a class assertion
> and
>> stated in the functional syntax it looks like this:
>>
>>             ClassAssertion(ObjectMinCardinality(2 :likes :Dogs) :jim)
> [Jim Tivy] This one I like - no pun intended.
> I like using the class expression.  Unfortunately I have not done the proof
> that this is better - it just seems to have more legs.
>
>>
>> It says that Jim likes at least two dogs.  He stated it this way because
> you
>> didn't say "Jim likes a dog" but "jim likes dogs" which might be
> interpreted as
>> "jim likes several (more than one) dog".
>>
>> -Timothy
>>
>>
>>
>>
>> _______________________________________________
>> protege-discussion mailing list
>> protege-discussion at lists.stanford.edu
>> https://mailman.stanford.edu/mailman/listinfo/protege-discussion
>>
>> Instructions for unsubscribing:
>> http://protege.stanford.edu/doc/faq.html#01a.03
> _______________________________________________
> protege-discussion mailing list
> protege-discussion at lists.stanford.edu
> https://mailman.stanford.edu/mailman/listinfo/protege-discussion
>
> Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03




More information about the protege-discussion mailing list