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] Problems with disjoint classes

David Makovoz davidmakovoz at
Mon May 30 07:46:40 PDT 2011

>From the ontological point of view you cannot have a subclass of two disjoint 
D subClassOf A
D subClassOfB
B disjointFrom A
It is clearly impossible to create an individual of class D since it has to 
belong to both A and B, 
but at the same time it cannot belong to both A and B.

Whereas there is no problem with declaring disjoint sublcasses of a class, in 
fact it's done very frequently:
D subClassOf A
E subClassOfA
E disjointFrom D

However, semantically, it doesn't seem to make sense to say that skeleton is the 
superclass and bone, tendon and cartilage are subclasses.
Simply take what you said below " tissues such as bone, tendon and cartilage." 
and this translates into
bone subClassOf tissue
tendon subClassOf tissue
cartilage subClassOf tissue
To "connect" them to the skeleton you would naturally use an object property, 
e.g., hasTissue
skeleton hasTissue bone
skeleton hasTissue tendon
skeleton hasTissue cartilage.
Or possibly instead of hasTissue you can call it consistsOfTissue. 

The same with "exoskeleton (an organ), a microtubule (an organelle) and a bone 
(a tissue) and say that they all provide support.  "
You would have disjoint classes exoskeleton, microtubule, and bone , an object 
property "provides" and the filler "support". 
bone  disointFrom exoskeleton , etc...
exoskeleton  provides support
microtubule provides support
bone provides support

Does it make sense?


From: Alex Shkotin <alex.shkotin at>
To: User support for Core Protege and the Protege-Frames editor 
<protege-discussion at>
Sent: Mon, May 30, 2011 8:33:04 AM
Subject: Re: [protege-discussion] Problems with disjoint classes

what about to say

Class: F EquivalentTo: A or B or C

It is better to discuss real classes and relationships. For example if we have 
bone, tendon and cartilage disjoint classes then physically_connected may be a 
good object property to connect an instances of these classes to sceleton.

2011/5/30 Julian Vincent <j.f.v.vincent at>

Sorry - I can't understand this.  I can see that I need to introduce some sort 
of connection in a disjoint-free class, but I can't see how this should be done. 
 It has the appearance of a standard technique, but I don't know what that 
technique is.  Could you possibly sketch out what class structure I should end 
up with?
>To re-iterate and expand slightly:
>I have disjoint classes A, B and C which I wish to use with a single other 
>subclass D.  I may need to use the same classes to describe another subclass E 
>which is disjoint with subclass E
>I have disjoint classes a, b and c and wish to make them all members of another 
>Or are these problems both the same??
>On 30 May 2011, at 12:50, Alex Shkotin wrote:
>> Hi Julian,
>> you should have relationships between individuals of your disjoint classes. 
>>Just describing real type of connection between bone and tendon...
>> Than an instance of sceleton should be all connected individiuals and sceleton 
>>"area" is a union of your classes.
>> Alex
>> 2011/5/30 Julian Vincent <j.f.v.vincent at>
>> I want to describe (for instance) a an organ such as a skeleton made of tissues 
>>such as bone, tendon and cartilage.   But all the tissues are disjoint.  So the 
>>skeleton is not passed by the Reasoner and is assigned to the 'Nothing' class.
>> An inverse of this, producing the same problem, is to take a number of objects, 
>>such as an exoskeleton (an organ), a microtubule (an organelle) and a bone (a 
>>tissue) and say that they all provide support.  But organ, organelle and tissue 
>>are disjoint.
>> What's the way around this?
>> Julian Vincent
>> _______________________________________________
>> protege-discussion mailing list
>> protege-discussion at
>> Instructions for unsubscribing: 
>> _______________________________________________
>> protege-discussion mailing list
>> protege-discussion at
>> Instructions for unsubscribing: 
>protege-discussion mailing list
>protege-discussion at
>Instructions for unsubscribing:
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the protege-discussion mailing list