[topicmapmail] Re: Getty Art and Architecture Thesaurus On Line

[Murray Altheim]

kal@techquila.com wrote:
> Murray Altheim writes:
> 
> 
>>kal@techquila.com wrote:
>>
>>>Lars Marius Garshol writes:
>>>
>>>
>>>
>>>>* Murray Altheim
>>>>| 
>>>>| No, I don't think the terminology is incorrect, it's just that
>>>>| things can be instances or not within specific contexts. "species"
>>>>| is a class in one context (when it is used as a categorizing term)
>>>>| and is an "instance" in the context of being an instance of a
>>>>| zoological taxonomical category (or taxon, for short).
>>>>
>>>>I think Murray is right here. We could have fun discussing whether or
>>>>not these are the same subject "species" or not, but in my mind
>>>>it's clear that both of these points of view make sense.
>>>
>>Again, Peirce's Firstness, Secondness, and Thirdness apply here (sorry
>>to be pedantic about this). "species" would be considered Firstness,
>>i.e., defined absent an association (Secondness) within a context
>>(Thirdness). So, the subject of species is safe -- it remains the
>>subject regardless of how it's used in associations or assertions (which
>>is what I'm calling statements of Thirdness in Ceryle).
>
> OK - but that is just one possible position, right ? It happens to be one
> that I agree with, but thats neither here nor there :)

Certainly. I don't expect those that disagree with you and I to agree
with you and I. I hope they feel likewise.

>>>Actually, I think that it is important that we agree on whether a topic
>>>playing the XTM-defined role of 'class' in one association can also play
>>>the XTM-defined role of 'instance' in another association. It is *not*
>>>necessary is to enforce this as the only point of view, nor is it necessary
>>>(or even possible) to make a statement about whether those two "species"
>>>are the same subject or not. However, there should be a statement on the
>>>use of the XTM PSIs and it seems from this discussion that stating that the
>>>XTM class-instance association is not transitive is possibly not enough.
>>>I'm not in a position to check right now, but if its not in the current ISO
>>>13250 revision drafts, then I would urge the editors to make the semantics
>>>of the XTM class-instance association and the XTM subclass-superclass
>>>association as tightly defined as possible.
>>
>>Well, any logic I'm aware of defines class-instance as intransitive,
>>so I think you're safe. Mathematically, that's the case AFAIU.
> 
> Well, I would hope so. But there still remains the issue of whether it is
> an error for a topic which is being used as a class to also be an instance
> of a class. My feeling is that it is not an error, simply the result of
> representing two different type systems in a single topic map (which may be
> the result of merging two separate topic maps). But if its not an error,
> and we all agree that its not an error, perhaps it would be best to say
> that its not an error so as to avoid confusion. Or is this all just so
> obvious that it goes without saying ?

I can't remember the term, but to constrain topic maps would be something
like a category error, in application of a constraint at the wrong meta-
level. What we've *really* been talking about are semantics at a knowledge
modeling level, above the topic map level. If you note that the "semantics"
supplied with XTM were *mostly* in the form of PSIs, where absent them XTM
would be almost semantics-free, this is akin to RDF being similarly
semantics-free, only gaining them in specific applications like OWL. SO
in a sense what we've been talking about is an application layer on top
of topic maps, a set of PSIs enabling a specific model. So it would be
relatively straightforward to take the URLs available in OWL and use XTM
to express OWL statements (in a Description Logics way), or use any other
formal modeling language to express it differently. XTM is simply a
carrier for the upper level. It has the advantage of having built-in
syntax and semantics that make clear a number of things that RDF doesn't
do "natively", such as reification and most importantly, *mapping*.

Murray

...........................................................................
Murray Altheim                         http://kmi.open.ac.uk/people/murray/
Knowledge Media Institute
The Open University, Milton Keynes, Bucks, MK7 6AA, UK                    .

   The Office of Special Plans (OSP) was set up by the defence secretary,
   Donald Rumsfeld, to second-guess CIA information and operated under
   the patronage of hardline conservatives in the top rungs of the admini-
   stration, the Pentagon and at the White House, including Vice-President
   Dick Cheney. The ideologically driven network functioned like a shadow
   government, much of it off the official payroll and beyond congressional
   oversight. [...It] also forged close ties to a parallel, ad hoc intelli-
   gence operation inside Ariel Sharon's office in Israel specifically to
   bypass Mossad and provide the Bush administration with more alarmist
   reports on Saddam's Iraq than Mossad was prepared to authorise.
   http://www.guardian.co.uk/Iraq/Story/0,2763,999737,00.html

   (my longest .sig!)