John F. Sowa wrote: (01)
> 7. Therefore, we should make a clear distinction between the
> vocabularies or terminologies, which have very few axioms,
> and the problem-oriented reasoning and computational
> systems. For general purposes, sharing should be based on
> the terminology. For reasoning and computation, the axioms
> should be introduced at the lower, problem-oriented levels. (02)
> Some things, such as a star or a vortex could be listed without
> any commitment to whether they're an object or a process. Then
> one could talk (or reason) about the sun as an object for some
> applications or a process for others. A hurricane could be
> listed as a vortex, and it would be possible to reason about a
> hurricane with either object-oriented or process-oriented axioms. (03)
To accept the idea of a simple upper ontology w/ very few axioms,
the distinction between vocabularies vs. problem-solving needs more
clarification
to understand how this might work. (04)
For example, without knowing the details of the problem,
I am clueless w.r.t. what difference there may be between
"a-hurricane-as-an-object" vs. "a-hurricane-as-a-process".
By design, the upper ontology is similarly non-informative
about the difference between "object" and "process".
The semantics of the upper ontology may be so impoverished
that I may not be able to tell if it is possible -- or not --
for some X to be both "object" and "process" at the same "time". (05)
However, if we had some kind of DOLCE-like "description" of the problem
of interest
+ a DOLCE-like "situation" where that problem is applicable to "hurricane",
then we could "disambiguate" the tell which of
"a-hurricane-as-an-object" vs. "a-hurricane-as-a-process"
is the right interpretation of "hurricane" in the context of that
problem by looking at the
role that "hurricane" has in the problem situation. (06)
For example, if the problem description is specified in DOLCE,
then an alignment of the upper ontology to DOLCE for that particular
problem context could be: (07)
"object" -> DOLCE-Lite#Endurant
"process" -> DOLCE-Lite#process
(see: http://www.loa-cnr.it/Files/DLP_description.txt) (08)
The problem description could be specified using PSL instead of DOLCE.
A possible alignment might be: (09)
"object" -> psl_core#object
"process" -> psl_core#activity
(see: http://www.mel.nist.gov/psl/psl-ontology/psl_core.html) (010)
According to how much axiomata is needed to "describe" the problem,
we might have to refine the alignment towards greater specificity
which adds more semantic information, e.g.: (011)
"process" -> state_precond#markov_precond
(see:
http://www.mel.nist.gov/psl/psl-ontology/part42/state_precond.def.html) (012)
Without a way to "understand" what the notion of "state" is,
then this mapping is no more informative than a semantically different
one,e.g: (013)
"process" -> state_precond#partial_state (014)
> That approach would accommodate both Whitehead's ontology, which
> makes processes fundamental, and an ontology that makes objects
> fundamental. In W's ontology, all objects, stars, and vortices
> would be defined as types of processes, but there would be no
> need to make that assumption in general. (015)
If all we care to align/map are "objects" and "processes" then this
might be OK.
However, we we have an important term, e.g., "function" that is not part
of the vocabulary of the upper
ontology, then its meaning would have to be specified relative to
alignments to semantically richer ontologies.
This complexity seems unecessary and it would make sense to add
"function" to the vocabulary of the upper ontology
which would allow users to specify its meaning by alignment to a
suitable ontology. (016)
Whenever we add an extra term to the upper ontology, we correspondingly
increase the complexity of mapping
the upper ontology to a richly axiomatized ontology: each separate term
needs a semantically distinct alignment
(otherwise the linguistic distinctions allowed at the upper ontology are
not preserved when mapped to a supposedly
richly axiomatized ontology). (017)
Eventually, we need to find the right balance between expressiveness (#
of terms in the upper ontology)
and practicality (intuitiveness and ease of mapping the terms of the
upper ontology to a richly axiomatized ontology). (018)
> For any kind of detailed work in science and engineering,
> Whitehead's ontology is more realistic, but for reasoning
> about everyday things, it might be convenient to assume that
> objects are the participants that constitute processes. For
> some problems, one or another of those assumptions might be
> preferable, but those detailed axioms should only be assumed
> at the problem-oriented level, *not* at the upper levels.
>
> Another example is the question whether a vase and the lump
> of clay from which it is made are one object or two. That is
> another assumption that is very much problem dependent, and
> it should *not* be a requirement enforced at the upper levels.
> The only people who worry about such issues might be pottery
> workers, and they have much more detailed problems to think about. (019)
To me, this argument points to the need to specify the "signature" of
the theory with which we care to talk/reason about kinds of problems.
That signature specifies a set of entities, relations, etc... that
formally or informally,
we need to talk/reason about problems in a way that allows us to adjucate
which statements are true/false. (020)
Once we have this signature, we can determine whether we can align it
to a known upper ontology if we that upper ontology has a vocabulary
structure rich enough to map the signature to it. If the vocabulary is
too poor,
then the upper ontology lacks the linguistic expressiveness we need to
talk/reason
about our problems. (021)
>
> The SUO work over the past five years has been interesting,
> and we all learned a lot. But the most important thing we
> learned is that assuming a fixed and frozen set of upper-level
> axioms does not promote interoperability. Instead, the axioms
> introduce irrelevant contradictions that are a major barrier
> to communication and sharing. The solution is to minimize the
> axioms at the top levels and to introduce them as needed at
> the problem-oriented lower levels. (022)
This sounds fine but it is unclear to me what that means on a practical
level. (023)
-- nicolas. (024)
>
> John Sowa
>
>
> _________________________________________________________________
> Message Archives: http://colab.cim3.net/forum/ontac-forum/
> To Post: mailto:ontac-forum@xxxxxxxxxxxxxx
> Subscribe/Unsubscribe/Config:
> http://colab.cim3.net/mailman/listinfo/ontac-forum/
> Shared Files: http://colab.cim3.net/file/work/SICoP/ontac/
> Community Wiki:
> http://colab.cim3.net/cgi-bin/wiki.pl?SICoP/OntologyTaxonomyCoordinatingWG
>
> (025)
_________________________________________________________________
Message Archives: http://colab.cim3.net/forum/ontac-forum/
To Post: mailto:ontac-forum@xxxxxxxxxxxxxx
Subscribe/Unsubscribe/Config:
http://colab.cim3.net/mailman/listinfo/ontac-forum/
Shared Files: http://colab.cim3.net/file/work/SICoP/ontac/
Community Wiki:
http://colab.cim3.net/cgi-bin/wiki.pl?SICoP/OntologyTaxonomyCoordinatingWG (026)
|