ontac-dev
[Top] [All Lists]

[ontac-dev] Too many axioms spoil the broth

To: ONTAC Taxonomy-Ontology Development Discussion <ontac-dev@xxxxxxxxxxxxxx>
Cc: CG <cg@xxxxxxxxxx>
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Tue, 24 Jan 2006 10:00:43 -0500
Message-id: <43D6411B.20805@xxxxxxxxxxx>
Barry,    (01)

Following is an example that illustrates the kinds
of inconsistencies that are created by having too
many axioms.  (I changed the subject line to
reflect that point.)    (02)

BS> Certainly processes have parts (subprocesses).
 > And which subprocesses of a process are unfolding
 > changes from one time to another.    (03)

Up to this point, nobody would disagree.  But the
following discussion depends on axioms inherited
from theories at a very high level of the ontology:    (04)

BS> But neither a process nor its subprocesses change
 > over time.  This is because a process/subprocess
 > IS a change over time.  It is continuants (objects)
 > which change over time, by undergoing processes.    (05)

What would you call Hurricane Katrina?  It certainly
retained its identity for a considerable period of
time.  During that time, it was a clearly recognizable
entity, whose identity was sufficiently stable that
it could be named, discussed, perceived, and identified
by millions of observers.    (06)

Like most hurricanes, Katrina was first noticed as
a "tropical depression".  As its winds increased, it
became a "tropical storm", at which point it was given
the name Katrina.  It strengthened to become a hurricane,
and eventually it moved across land, weakened, and
dissipated as a storm.    (07)

By the axiom that processes cannot change, Katrina
would be classified a continuant (object).  But it
certainly had many process-like characteristics,
and it falls on a continuum with storms, dust devils,
whirlpools, waves, ripples, and many other phenomena
that are not usually considered objects.    (08)

Examples like these are the reasons why I recommend
an upper level with very few axioms:    (09)

  1. The people who study hurricanes and the people
     who experience the effects on their homes are
     not about to consult an ontologist to determine
     whether they are being blown away by a continuant
     or an occurrent.    (010)

  2. Axioms are certainly important for detailed
     reasoning, but the people who need a theory of
     hurricanes have vastly more data, hypotheses,
     and verified theories about such phenomena than
     anything they would inherit from an upper level.    (011)

  3. The kinds of general theories that are relevant
     to reasoning about hurricanes and other phenomena
     belong to mathematics and physics.  Those theories
     are usually very specialized, there are large numbers
     of them, and they tend to change frequently (on at
     least a yearly basis).  They don't belong at an
     upper level, but they do belong in a large collection
     of optional modules that can be introduced, as needed,
     for any particular problem.    (012)

As I said before, I am not against having some axioms
at the upper levels, but they should mainly be used for
classification rather than detailed reasoning.  And even
for classification, there are many serious questions
created by borderline cases, such as hurricanes and
other physical phenomena.    (013)

Summary:  The physicians' motto, "Primum non nocere"
(First, do no harm), implies that in cases of uncertainty,
the physician should take no action that would worsen the
patient's condition.  I am afraid that many axioms and
distinctions that are being proposed for an upper-level
ontology violate that principle.  Instead of clarifying
the problem, they often introduce distractions, confusions,
and inconsistencies that make it worse.  That doesn't mean
we should have no upper-level axioms, but they should be
used mainly for classification, not for detailed reasoning.    (014)

John Sowa    (015)


_________________________________________________________________
Message Archives: http://colab.cim3.net/forum/ontac-dev/
To Post: mailto:ontac-dev@xxxxxxxxxxxxxx
Subscribe/Unsubscribe/Config: http://colab.cim3.net/mailman/listinfo/ontac-dev/
Shared Files: http://colab.cim3.net/file/work/SICoP/ontac/
Community Wiki: 
http://colab.cim3.net/cgi-bin/wiki.pl?SICoP/OntologyTaxonomyCoordinatingWG    (016)
<Prev in Thread] Current Thread [Next in Thread>