[sc34wg3] Topic Maps Reference Model posted (new draft)

Murray Altheim sc34wg3@isotopicmaps.org
Tue, 01 Mar 2005 01:01:06 +0000


Michel Biezunski wrote:
> Bernard, Murray
> 
> I agree with most of what you say. I am advocating to
> use the word "Perspective" to designate a TMA (or an ontology,
> which is basically the same).
> 
> The reason why I prefer the word "perspective" is that it
> immediately conveys the notion that it is a particular
> point of view rather than being the absolute, true, essential
> nature of things (that's why the term "ontology" is so misleading).
> 
> However, I wouldn't consider a totally relativistic perspective.
> Here is how I see things working. In the worlds there are:
> 
> - things (such as mountains, they just *are* and maybe some of them have
>           never ever been seen by any human being and therefore don't
>           have even names)

Michel,

Interesting you should choose "mountains" as an example of
something that just "is", as it turns out this is an excellent
example of something that "isn't" necessarily. People look up
and see a peak, and certainly a peak (as the highest visible
crest of an individual outcropping at a given distance) seems
to exist, but even that is quite tenuous.

 From a geological perspective, Mt. Everest doesn't really exist
as distinct from the two peaks that are near it, Lotse and
Nuptse, as the boundaries between them are (at least to my
knowledge) completely undefined, i.e., when one is in the bowl
of the Khumbu icefall it's hard to say which of the "three"
mountains on is standing on. There is a "top" of Mt. Everest,
but it's not exactly a point -- once up there it's actually a
small, relatively flattened plain that just happens to have a
part that is higher than the rest, and even that is measured
at a certain resolution -- at an even higher resolution we'd
have to get to what part of that bump is the highest, but it's
unfortunately usually covered with snow (six feet of which are
usually included in the measurement of its height). And we'd
and also have to take into account where the moon is in her
cycle, since that alters the shape of the earth.

Resolution in this case forms just one part of the context by
which we understand that Mt. Everest is a "peak." Peaks form
part of a continuous surface, with various points of relatively
lower-than and higher-than areas linked by apparent ridges and
other features, much like the surface of a blanket thrown upon
a bed. If one were to measure from the bottom of the ocean, the
Hawaiian volcano Mauna Kea has the greatest relief on earth at
10,200 meters (only 4,203 meters of which are above sea level).
So we have to also be sure we're talking about land-based
measurements. Where does the mountain begin?

Nobody would think to designate the peaks and valleys of a
blanket as individual mountains or valleys, and in the same way,
the surface of the earth has peaks and valleys only insofar as
humans have identified them, at a given resolution, and even
within a given culture or group of people. I'm certain that
the differentiated and named features in the Kentucky backwoods
are only known to those who live there, and don't show up on
any maps. Some features are only known to a family, some only
to the person who has named them. And different families over
time have had different factors and names. Et cetera.

Barry Smith actually has a paper that covers the mereotopological
aspects of this:

   Do Mountains Exist?
   http://ontology.buffalo.edu/smith/articles/Mountains.htm

where he states

  "But whether they are considered as individuals (tokens) or as
   kinds (types), mountains do not exist in quite the same
   unequivocal sense as do such prototypical everyday objects as
   chairs or people."

I think the point I'd make here is that mountains only exist
insofar as *people* begin to make distinctions about them
(often ignoring any factors about boundaries or epistemology
of boundaries), such that they don't *necessarily* exist, but
exist *certainly* as the topics of conversations, within the
realm of active human communication. (I think the jury is still
out on whether people communicate with computers, or computers
with computers, so I'll leave that idle for now)

Then there is a matter of who are the participants in the
communication, and under what circumstances they are making
the communication. Even an individual creates contexts for
the ways they talk about things, as dependent upon who they
are conversing with, and under what circumstances.

Everest is also known by the Tibetans as Chomolangma (short
for " Jomo Miyolangsangma", Goddess Mother of the Snows), and
by the Nepalese as Sagarmatha (Mother of the Universe). There
even seems to be variant translations of the latter, as I've
heard it as "Summit Over the Sky" and "Brow of the Ocean"
(which seem a bit hard to reconcile). You might imagine by
these names that certain people might think of the mountain
in a very different way than say, Edmund Hillary, the Sherpa
Tensing Norgay, or that French guy who ran to the top and
jumped off with a hang glider. This was considered a huge
affront by not only the Sherpas but large parts of the
climbing community (I won't go into what constitutes the
boundaries of communities...).

Another way of looking at this is that the whole issue of what
constitutes a mountain may be seen either as a matter of human
communication or as a matter of mereotopology, i.e., an issue
of what constitutes borders and boundaries of topological
landscapes. If Mt. Fiji exists, one must either simply talk
about it as if it does, or begin to define the actual boundaries
of where the mountain begins and its surrounding landscape ends,
whether one is talking solely about its surface or about its
interior, whether objects on the surface of the mountain (e.g.
trees or boulders) constitute a part of the mountain, whether
a fissure that entirely separates it from the main mass now
constitutes a new mass or should be considered part of the main,
etc. If a mine entrance begins at one location but travels
outside of the surface boundaries, is the mine still a part of
the mountain? All these questions and answers are only askable
and answerable within a given context, such that a man and woman
taking a holiday there will treat these matters differently
than a geologist, a miner, or a real estate agent might.

Smith and Jean Petitot have a related paper that might also be
of interest:

   Physics and the Phenomenal World
   http://ontology.buffalo.edu/smith/articles/ppw.pdf.

This provides an emergentist approach to ontology that might be
a way out of this mess, by providing a mathematically-sound (or
so it is claimed) way of describing the macroscopic world in
terms of its microscopic boundary patterns. An interesting notion,
but I'm not sure if epistemologically it's not simply passing off
the problem to simply a different level, i.e., if the problem is
truly recursive, then it's only our ability to perceive at a
given resolution that provides us with a foundation upon which
to stand. As our resolution increases (as per our needs and
abilities) that pediment crumbles. But I'm open to argument on
this one...

> - subjects (units of meaning or understanding. That's what I as a 
>             human am trying to convey to another human )

Providing that there is an understanding that the "unit" is
left undefined? I think it would be very difficult to actually
define such units, as they are very tightly bound to the context
in which something is expressed. If, for example, someone is
conversing with another person who shares a great deal in common,
those shared assumptions are part of the communication, the "in"
language. If one is communicating with someone whose culture and
language are different, the amount of "in" language, the set of
basic, shared content might be enough to enable a grasp of the
essential communication (e.g., "toilet?" followed by nod of head)
but miss the details ("is your toilet clean?"). Sometimes these
missed details are relatively unimportant, sometimes they cause
people to die (e.g., misinterpreted directions in hospitals).

Then there's also the linguistics perspective on this, such that
we begin to break down what constitutes language units, etc.

> - proxies (expression of subjects in a given language or syntax)

Language itself is a proxy, a symbol system that allows the relation
of symbol to experience or idea. "Proxy" is merely a statement of a
level of meta- over its originating layer of meaning. So we have
some sort of perception that we label with a word. We can create
a proxy for that word. We can create a proxy for that proxy, with
turtles, turtles, all the way down...

   Stephen Hawking's A Brief History of Time tells the story of a
   cosmologist whose speech is interrupted by a little old lady who
   informs him that the universe rests on the back of a turtle. "Ah,
   yes, madame," the scientist replies, "but what does the turtle
   rest on?" The old lady shoots back: "You can't trick me, young man.
   It's nothing but turtles, turtles, turtles, all the way down."

[I know this comes from a Maidu cosmology but I wonder how many
others?]

> - perspectives (set of rules to which proxies must comply)

The rules themselves are recursively "perspective-ized", sometimes
under similar constraints to the statement they are a proxy for,
sometimes from an entirely different perspective (as the proxy's
purpose might *be* to provide such an altered perspective or
context).

> - views (set of proxies treated as a unit)

This seems to me to essentially be a meta-level of context, since
contexts are recursive. So a "view" is one set of contexts. A
collection of views (at a "higher" level) would itself form another
view. Like those mountain tops, these levels only exist in terms
of their use within an interpretive context, not innately or
objectively.

> I have tried this recently in a semantic-web-oriented audience.
> People get it when I used the word "perspective". It's so much
> simpler to understand!

I can certainly see how people might seem to "get it," by nodding
their heads and stating that they understand what you're saying
since the language is less obtuse and more understandable, but
are you sure they're getting exactly what you're saying, or only
doing as all humans do, interpreting what you're saying in their
own way, which might not actually agree with what you are saying?
Such incongruities sometimes only come to light after the misinter-
pretations are brought to light ("hey, I thought you said your
bathroom was CLEAN!").

> Let me know what you think.

I hope this isn't seen as too critical. I think we're all in a
big conversation about this, and it's all pretty early on.

My 4p on this one,

Murray

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

  Sometimes things are so obvious that they merely need pointing out:

       LATEST NEWS : MAJORITY OF TEENS DONíT WANT TO HAVE SEX
       Abstinence Clearinghouse
       http://www.abstinence.net/library/index.php?entryid=1808