W3C home > Mailing lists > Public > semantic-web@w3.org > August 2008

Re: Why do you want to do that?

From: Denny Vrandečić <dvr@aifb.uni-karlsruhe.de>
Date: Tue, 12 Aug 2008 14:45:32 +0200
Message-ID: <48A185EC.5090302@aifb.uni-karlsruhe.de>
To: "Richard H. McCullough" <rhm@pioneerca.com>
CC: Pat Hayes <phayes@ihmc.us>, Semantic Web at W3C <semantic-web@w3.org>, KR-language <KR-language@YahooGroups.com>

Just a few notes from an interested reader...

In OWL DL and OWL Lite Individuals and Classes are indeed mutually
disjoint. This is not the case in RDF, OWL Full, and in the current OWL2
languages (the latter build on Boris Motik's work [1], very recommended
reading about the current topic).

Dick, you have frequently stated that you consider individuals and
classes to be mutually disjoint, and this is understood, but you have
not answered to Pat's very illustrative examples in his 13:47 Mail (ID
<p06230925c4c6cd2b0a0f@[4.246.36.83]>).

Another example: why cannot Eagle be both a class (for "Fred the eagle")
and an individual (instantiating the class "species")? (similar to
Boris' example from his paper)

It is an ontological question, and the answer depends on your definition
of the terms "individual" and "class". Whereas I think to understand
Pat's definitions of these terms (since they are in line with the W3C
standards), I am still struggling to understand your definitions. It
would be very helpful if you could give references to their definitions,
since this would resolve the discussion.

And now back to my thesis ...

Cheers,
denny


[1] "On the properties of Metamodeling in OWL", best paper ISWC 2005,
Journalversion here: http://dx.doi.org/10.1093/logcom/exm027

Richard H. McCullough wrote:
> Hi Pat
> We've got a lot of confusion between us right now.
> I'll do my best to straighten it out.
> See below for my answers, comments, etc.
>  
> Dick McCullough
> Ayn Rand do speak od mKR done;
> mKE do enhance od Real Intelligence done;
> knowledge := man do identify od existent done;
> knowledge haspart proposition list;
> http://mKRmKE.org/
> ----- Original Message -----
> From: "Pat Hayes" <phayes@ihmc.us <mailto:phayes@ihmc.us>>
> To: "Richard H. McCullough" <rhm@pioneerca.com <mailto:rhm@pioneerca.com>>
> Cc: "Frank Manola" <fmanola@acm.org <mailto:fmanola@acm.org>>; "Adam
> Pease" <adampease@earthlink.net <mailto:adampease@earthlink.net>>;
> "Semantic Web at W3C" <semantic-web@w3.org
> <mailto:semantic-web@w3.org>>; "KR-language"
> <KR-language@YahooGroups.com <mailto:KR-language@YahooGroups.com>>
> Sent: Monday, August 11, 2008 11:15 PM
> Subject: Re: Why do you want to do that?
> 
>> At 10:56 PM -0700 8/11/08, Richard H. McCullough wrote:
>>>I finally got a few minutes to read OWL Guide 3.1.3
>>>I read that section as supporting my position.  The word "context"
>>>is mentioned
>>>several times, with the implication that X ismem IndividualSet; in
>>>one context,
>>>and X ismem ClassSet; in a different context.
>>
>> Is that actually asserted anywhere? Please give a citation.
>  
> I said "implication".  It is not clearly asserted.  You have to read
> between the lines.
> Here is the beginning of OWL Guide, section 3.1.3.
> Note the two fragments I have put in bold type.
> In the second one, the wording "could itself be considered" implies that
> they
> again mean "in certain contexts".
> Note also that "could be considered" is not the same as "is".
> To correlate with the mKR language, think "could be *view*ed".
>  
> QUOTE
> 
> There are important issues regarding the distinction between a class and
> an individual in OWL. A class is simply a name and collection of
> properties that describe a set of individuals. Individuals are the
> members of those sets. Thus classes should correspond to naturally
> occurring sets of things in a domain of discourse, and individuals
> should correspond to actual entities that can be grouped into these classes.
> 
> In building ontologies, this distinction is frequently blurred in two ways:
> 
>     * Levels of representation: It is well known that *in certain
>       contexts something that is obviously a class can itself be
>       considered an instance of something else.* For example, in the
>       wine ontology we have the notion of a Grape, which is intended to
>       denote the set of all grape varietals. CabernetSauvingonGrape is
>       an example instance of this class, as it denotes the actual grape
>       varietal called Cabernet Sauvignon. However,
>       *CabernetSauvignonGrape could itself be considered a class,* the
>       set of all actual Cabernet Sauvignon grapes.
> 
>     * Subclass vs. instance: It is very easy to confuse the instance-of
>       relationship with the subclass relationship. For example, it may
>       seem arbitrary to choose to make CabernetSauvignonGrape an
>       individual that is an instance of Grape, as opposed to a subclass
>       of Grape. This is not an arbitrary decision. The Grape class
>       denotes the set of all grape varietals, and therefore any subclass
>       of Grape should denote a subset of these varietals. Thus,
>       CabernetSauvignonGrape should be considered an instance of Grape,
>       and not a subclass. It does not describe a subset of Grape
>       varietals, it is a grape varietal. 
> 
> UNQUOTE
>>>
>>>But a problem arises because OWL doesn't have contexts.
>>>So, apparently, the OWL solution to mix all contexts together,
>>>and ASSUME that all the propositions are still true.
>>
>> No. The OWL methodology, like that of virtually all modern logic, is
>> to define a formal semantics for the notation, which then DETERMINES
>> what is true and false. There are no assumptions anywhere. So there
>> isn't a problem. And there are no contexts in OWL as it isn't a
>> context-dependent language.
>  
> The namespaces used (including rdf, rdfs, owl)
> contain propositions which are part of the context of
> every OWL "document".
>  
> Every OWL "statement" has a context, which includes the "document" context.
> In a list of OWL statements, each statement is part of the context of
> the OWL statements which follow it.
>  
> The mKR language explicitly says what the context is, using the format
>     at space=s, time=t, view=v { statement };
> 
>>
>>>
>>>To make this problem more apparent, we can specify the contexts
>>>
>>>    at view =  ind { X ismem IndividualSet; };
>>
>> What does that mean? What is the semantics of your formalism? Because
>> if you were to provide one, that would stop all the argument, by
>> answering the question.
>  
> "ind" is the name of a proposition list, which is the context of the
> statement "X ismem IndividualSet;".
> 
>>
>>>    at view = cls { X ismem ClassSet; };
>>>
>>>The question is: what happens when we mix the two contexts together?
>>
>> What DETERMINES what happens? How is consistency defined for your
> formal logic?
>  
> Use all the propositions to deduce what is true, and what is false.
>>
>>>
>>>Pat Hayes says
>>>    at view = mix { X ismem IndividualSet; X ismem ClassSet; };
>>
>> No, I didn't say that, as I don't speak this language. I wrote in English.
>  
> Agreed.  That is the mKR paraphrase of what you wrote in English.
> 
>>
>> Pat
>>
>>>Dick McCullough says
>>>    at view = mix { not{X ismem IndividualSet;}; X ismem ClassSet; };
>>>
>>>
>>>Dick McCullough
>>>Ayn Rand do speak od mKR done;
>>>mKE do enhance od Real Intelligence done;
>>>knowledge := man do identify od existent done;
>>>knowledge haspart proposition list;
>>>http://mKRmKE.org/
>>>
>>>----- Original Message ----- From: "Frank Manola" <fmanola@acm.org
> <mailto:fmanola@acm.org>>
>>>To: "Richard H. McCullough" <rhm@pioneerca.com <mailto:rhm@pioneerca.com>>
>>>Sent: Friday, August 08, 2008 9:01 AM
>>>Subject: Re: Why do you want to do that?
>>>
>>>>On Aug 8, 2008, at 11:21 AM, Richard H. McCullough wrote:
>>>>
>>>>>
>>>>>Over the last six years, I have suggested a number of
>>>>>"improvements" to the RDF language.  Not one of
>>>>>my suggestions was adopted.  Apparently,
>>>>>RDF is fine just the way is, thank you!
>>>>
>>>>Yep.  That doesn't imply opposition to improvements though;  some
>>>>people think the way to provide the "improvements" they want is to
>>>>define languages "on top of" RDF (like the OWL dialects) rather
>>>>than  making those changes directly in RDF.  That way, your
>>>>"improvement"  and my improvement can possibly co-exist more nicely
>>>>:-)
>>>>
>>>>>
>>>>>
>>>>>I would now like to turn the tables, and ask
>>>>>why do you want to do that?
>>>>>I'll start with two features of RDF which seem to be popular.
>>>>>
>>>>>1. X  subClassOf  X;
>>>>>A neat mathematical property, right?
>>>>>But if you do the inferences, what it means is
>>>>>   X  sameAs  X;
>>>>>We already knew that.
>>>>>Why do you want to do that?
>>>>
>>>>I need some help with this question.  Do you think being able to
>>>>say X subClassOf Y is OK?  If so, are you asking why RDFS (not RDF,
>>>>BTW) doesn't explicitly forbid the special case of X subClassOf X?
>>>>Why do  you want to do that (i.e., test for this special case all
>>>>the time)?   Or are you asking why people *write* X subClassOf X?
>>>>
>>>>>
>>>>>
>>>>>2. X  type  Y;  X  subClassOf  Z;
>>>>>Another neat property: X is an individual and a class.
>>>>>Now I can ... What?  I don't know.
>>>>>Why do you want to do that?
>>>>
>>>>How about the example in Section 3.1.3 of the OWL Guide?
>>>>
>>>>--Frank
>>
>>
>> --
>> ---------------------------------------------------------------------
>> IHMC (850)434 8903 or (650)494 3973   home
>> 40 South Alcaniz St. (850)202 4416   office
>> Pensacola (850)202 4440   fax
>> FL 32502 (850)291 0667    cell
>> http://www.ihmc.us/users/phayes      phayesAT-SIGNihmc.us
>> http://www.flickr.com/pathayes/collections
>>
>>
>>
>>
Received on Tuesday, 12 August 2008 12:46:10 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 21:45:24 GMT