Matthew,
A few clarifications - it seems I have not been sufficiently explicit on
some points: (01)
>
> [MW] Unfortunately there is no limit to the number of primitive
> concepts.
>
This is a question whose answer no one knows for sure, but one that can be
experimentally investigated, by starting with a plausible foundation
ontology of primitives and seeing whether it needs supplementation to create
representations of things in many different fields. It is not a question
that I think anyone can honestly answer by gut instincts or from some
mathematical derivation from first principles. It is a question about a
natural phenomenon - just what are the units of thought that people use, and
how can they be formalized for use by computers? I am suggesting that a
serious project be organized that can answer that question.
I don't know the answer. I think that the experience of dictionary
makers with defining vocabularies strongly suggests that there *is* a limit
to the number of primitives needed to describe everything people talk about.
But it is possible that the principle may not apply to logical
representation of information. It is in my view immensely important for
those interested in interoperability to discover whether any such set of
logically specified primitives can be identified.
I can imagine at least two kinds of behavior, as one uses a starting
foundation ontology to specify the meanings of things in diverse specialized
domains: (1) the number of primitives required may approach an asymptote -
this can be determined by statistical analysis of the required new
primitives for each increment of specialized terms that are logically
described; or (2) the number of new primitives required may increase
indefinitely, in a manner analogous to way prime numbers increase as the
total number of integers increase, i.e. as a decreasing fraction, but
without limit.
The difference will be in how stable the foundation ontology will be at
any given time. Ideally, in the asymptote case, it will be very stable
after some initial period of use. That will be ideal. But if the number
of primitives continues to rise, like prime numbers, that will require more
sophisticated versioning - a need to build a utility so that the
representations in old applications can be updated as each new version with
new primitives is adopted. The primitive-based FO can still be the optimal
solution to interoperability, even if the number of primitives does in fact
have no limit.
But in either case, this still provides a means to have a common paradigm
of meaning that is as stable at any given time as nature permits. This
approach seems more likely than any other to create a stable paradigm of
meaning, supporting the most accurate interoperability that is possible for
automatic interpretations over multiple independently developed
applications.
That does not mean that local standards for specialized purposes would
have to be abandoned. Quite the opposite. Existing ontologies will be very
important in providing the test cases to determine which semantic primitives
are required. Since the foundation ontology will have all of the primitives
required to translate any **existing** local ontology data into the
representation of any other local ontology that is represented using (or
mapped to) the FO, local groups can and should use any representation that
they consider optimal for their purposes, but they can preserve global
interoperability by either specifying their terms using the primitives of
the FO, or retroactively mapping to the FO. IF any new primitives are
required to represent terms in some new application ontology, those new
primitive will be added when the need is recognized. (02)
>
> [MW] So what you are really proposing then is that you create your own
> favourite ontology, and then map every other ontology into it and out
> of it, or persuade others to do it for you.
>
No, I am proposing that a large *consortium* be organized including anyone
with an interest in interoperability, to create a starting FO based on
semantic primitives that can be used to translate among all of the
('favorite') domain ontologies of interest to the participants. After that,
it can evolve to include enough semantic primitives to translate among any
other ontologies whose custodians find a benefit in being interoperable with
others. I do not foresee any technical barriers, and I feel confident that
enough participants can be found to form a credible critical mass of users
to support a viable common paradigm of meaning. What I don't know is how we
can get the money (probably #30 million) to create the first version of such
an FO by such a process. (03)
Pat (04)
Patrick Cassidy
MICRA, Inc.
908-561-3416
cell: 908-565-4053
cassidy@xxxxxxxxx (05)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (06)
|