# Re: Subsumption question

From: Thomas Schneider <schneidt@cs.man.ac.uk>
Date: Thu, 30 Apr 2009 17:35:21 +0100
Message-Id: <8944BC60-9494-48C5-A074-21C8C139BD9D@cs.man.ac.uk>

```Hi Matthew,

I think there's 2 problems here:

(A) the term "inherited anonymous classes" in Protégé seems to be
misleading IMHO. It actually means "inferred anonymous superclasses".
Protégé team, please correct me if I'm getting this wrong. Otherwise,

(B) There still seems to be confusion between the subclass
relationship (IS-A) and the superclass relationship (the converse of
IS-A). Let me try to clarify this:

* You say that BYD is a subclass of X, where X is the (anonymous)
class "SO & isClass some (K or M)".
* You look for super- and subclasses of Y, where Y is "SO & isClass
some K".

Now the following things hold:

(1) BYD is a subclass of X (see above).
(2) Y is a subclass of X by the logical laws, but not vice versa!

(1) and (2) together do not say anything about the relationship
between BYD and Y: those are completely independent from each other.
If the converse of (2) held, you could infer (via "subclass of" being
transitive) that BYD is a subclass of Y, and this would then appear as
a query result. With the things you've asserted, nothing can be
inferred (neither by humans nor by machines).

HTH

Thomas

On 30 Apr 2009, at 17:20, Matthew Graham wrote:

> Hi Nick,
>
> I did not actually say that I was expecting BYD to a subclass. I am
> looking for any match to my DL Query in Protege and BYD is not
> appearing at all. My expectation is, as you say, that my query would
> be a subclass of BYD. There must be something in how BYD is defined
> that is not making it appear as the superclass of my query.
>
> 	Cheers,
>
> 	Matthew
>
> On Apr 30, 2009, at 2:27 AM, Nick Drummond wrote:
>
>> Hi,
>>
>> You say that
>>
>> BYDraconis -> StellarObject and isClassifiedAs some (StellarTypeK
>> or StellarTypeM)
>>
>>
>> StellarObject and isClassfiedAs some StellarTypeK
>>
>> Then you would not expect BYDraconis to be a subclass of your
>> query, rather the other way around
>>
>> StellaTypeK -> StellarTypeK or StellaTypeM
>>
>> Not the other way around.
>>
>> Nick
>>
>>
>> 2009/4/29 Matthew Graham <mjg@cacr.caltech.edu>
>> Hi,
>>
>> I'm new to OWL and am still trying to get my head around some stuff
>> so apologies if this email seems to be asking a dumb question. I'm
>> also sending this message to this list since I think it is about
>> OWL in general.
>>
>> I'm using the latest version of Protege to work on my ontology and
>> have the following issue:
>>
>> Classes: StellarObject (subclass of AstrObject), StellarTypeK
>> (subclass of ClassificationCategory), StellarTypeM (subclass of
>> ClassificationCategory)
>>
>> Properties: isClassifiedAs (domain: AstrObject, range:
>> ClassificationCategory)
>>
>> A class BYDraconis has an inherited anonymous class: StellarObject
>> and isClassifiedAs some (StellarTypeK or StellarTypeM).
>>
>> I am making a class query with DL Query looking for: StellarObject
>> and isClassfiedAs some StellarTypeK
>> but it is not matching BYDraconis. So why generally is "Class and
>> myProperty some A" not a subclass of "Class and myProperty some (A
>> or B)" ? How should I be modelling this so that it does what I want?
>>
>>        Cheers,
>>
>>        Matthew Graham
>>
>>
>>
>>
>>
>>
>>
>

+----------------------------------------------------------------------+
|  Dr Thomas Schneider                         schneider@cs.man.ac.uk  |
|  School of Computer Science       http://www.cs.man.ac.uk/~schneidt  |
|  Kilburn Building, Room 2.114                 phone +44 161 2756136  |
|  University of Manchester                                            |