[sc34wg3] Topics and Subjects clarification

[Bernard Vatant]

> * Bernard
> | I disagree completely with the assertion "subjects are real-world things". It
> | is restrictive, misleading and metaphysically biased, in the sense that it
> | supposes that real-world is divided into things *before* we speak about
> | them, which is obviously wrong.
> | Definition of things is always ad hoc, arbitrary, and - in best cases - agreed
> | upon in a community through provisional consensus. In the worse and most
> | frequent cases, disagreement triggers religious wars. And BTW those wars
> | are triggered exactly because people confuse their subjects of conversation
> | with real-world eternal absolute things :o)
> | Main tool to agree on how to divide the world into those arbitrary "things" is
> | conversation, that's why I always insisted that subjects are strictly speaking
> | "subjects of conversation" - they are created, agreed upon and maintained
> | through conversation.

* Marc
> I am not sure what this 'clarification' is Bernard. It certainly is not a
> clarification of the way 'subject' is defined in 13250, XTM and SAM, because
> you obviously disagree with what is said there.

I agree with definition of subject in ISO13250.
But not with reference to real-world in SAM, which is not employed in 13250,
and not in XTM either, except last reference to it in Section 2. of XTM 1.0 that looks
like a bug, because all others were removed.

> You speak as if there were
> an objective truth, accessible for everybody, on what subjects reallly are,
> and that you are clarifying the notion of 'subject' by pointing at those
> truths.

Sorry if it looked that way. It's more a question of clarification of the definition we
stand on - always arbitrary - than of "truth".

> I do not believe there is such a universally accepted notion of what
> subjects are, and frankly I cannot believe you would actually believe that.

Of course there is no and can't be any universal agreement on that, and that's why we have
to clearly state on which definition we stand, or which we don't accept.

> So then apparently you are proposing to change the definition of 'subject'.

No. To stick to the original one in ISO 13250, which is clearly "whatever the topic map
author has in mind when creating the topic" ... "regardless of whether it exists".

> When I use Topic Maps I do not want to commit myself to statements such as:
> - It is false that the world is divided into things *before* we speak about them.

 ... Neither to the opposite "It is true that the world is divided ..."

> I think ISO13250 did a reasonable job in saying as little as possible about
> the 'real world' and the relation between it and Topic Maps, while realizing
> one cannot say nothing on the issue.

Sure. So don't mention it. That's my point.

> I think a Topic Map standard should
> bear as few philosophical implications as possible, and what you propose on
> the issue does not fulfill that condition.

Well "as few as possible" does not mean "none at all" ( remember Einstein's "as simple as
possible, but not simpler") ... you can't pretend to make a model of any kind without
standing on some philosophical principles, implicit or explicit, and I think better to
explicit them than assume there are none. I asked the same question in SUO forum last
year, "Don't you think building an upper ontology is culturally and philosophically
biased"? And I was answered by people-who-know-what-the-truth-is: "Why? First Order Logic
is beyond all cultural differences" ... well ... I flew away.
I think we should avoid the same kind of arrogance, and not go as far as pretending that
topic map paradigm is beyond all philosophical views, and therefore can support any one of
them. Topic Maps do convey some philosophical principles that should be more explicited -
not as absolute "truth" of course, but as "axioms" - which is different. And I'm pretty
sure that some views of the world won't be supported by topic maps, because they don't
agree with those axioms. This is not a problem to me - every tool and representation,
however generic, has its limits - but my hunch is it might be a problem for some people in
this community, that really *believe* that *anything* can be expressed in topic maps.

> Lars and Marc are standing in the corner of the Ontopia booth at a major XML
> conference, having a beer and are deeply engaged in discussion. Bernard
> enters and asks: 'Well, what is the subject?'. Lars answers: 'Nikita'.

Well ... "Nikita" in my mind is not necessarily the same subject than in yours' or Lars' -
depending among other things on how deep everyone is engaged in the beer :))

Bernard