PatH,
Deciding what is “primitive” may never
be formally resolvable in a mathematical sense, which is why the main criteria
for inclusion in the content of the base FO would be logical consistency and voting
by the consortium members. The practical goal is to develop an FO that
will support translation of representations in databases and ontologies into
each other, and that criterion also needs to be determined by whether the
translated systems will arrive at the same conclusions from the same
data. Aiming for representation of all of the semantic primitives is a practical
means to focus the initial effort on a finite goal, it is not an ideological stance
or a technical requirement. If the intended meaning of an ontology
element can be expressed axiomatically in terms of other, pre-existing ontology
elements (with the ‘disjoint’ exception), that would indicate
that it is non-primitive; whether it is nevertheless included in the base FO
will depend on whether there is support for or objections to its inclusion,
which would then have to be resolved by a vote. There should be little or
no time wasted on debating whether something is or is not primitive, and if it
turns out that the notion of being “primitive” does distract the
work without commensurate benefit, it can be ignored. Looking
for the primitives at the initial stage provides a practical guide to limit the
scope of the required agreement among the members developing the common
foundation ontology.
[PC] >> Force can be directly perceived by
the sense of touch, and things that make direct sensory impressions have, in my
view, a privileged status as prime candidates for primitiveness.
[PH] > Just as an academic aside: on what basis?
Because, at least, the linguistic description and the
logical specifications have to be comprehensible to the human developers, and
things that are directly experienced by all people are easiest to
understand. It does not logically follow, but I think it highly likely,
that terms referring to such sensory experience are the ones most commonly used
to describe everything else – i.e., most likely to be primitives in the
sense of not being describable by other things. Note that I said they
were “candidates” which means that they should be considered by the
consortium, but could be rejected if there is a good reason. I just doubt
that any of the direct perceptions would be rejected by a group aiming at
creating a practical FO useful for all of them.
[PH] > Why would this notion of what one might call
'psychological' primitiveness have any relationship at all with your earlier
notion of 'definitional' primitiveness? Science seems to have advanced by
abandoning sensorially rooted concepts in favor of much more abstract notions
such as symmetry preservation and conservation laws.
I would never confuse an ontology
intended to serve for interoperability among a highly diverse set of
applications as being equivalent to the fundamental foundations of mathematics
or physics – though they can be included.
[PH] > Both real science and naive science are fun
enterprises, but it is not a good idea to get them confused.
Yes, I did enjoy my career as a biochemist, and
this project actually applies an important lesson derived from that
experience. Progress occurs when multiple groups find experimental
support of disproof of hypotheses formulated within some *common paradigm*
that is accepted by (some) research community, and provides a framework for
interpreting the results. For the task of semantic interoperability, a
common foundation ontology serves as the *paradigm of meaning*, with multiple
sub-theories, that can be investigated experimentally (to determine if it serves
the purpose of interoperability or of specific applications).
Importantly, the paradigm can evolve and add more detail as multiple groups
discover how its parts behave in practical applications. When only
small groups use the same foundation ontology, not only is broad interoperability
hindered, but the ability of a large group to rapidly evolve the common paradigm
is forfeited because there are fewer workers in the small groups to test each
hypotheses.
[PH] > But voting is almost a worst-case way to
resolve technical debates, as it amounts to agreeing not to agree. Strongly
held views which are voted out will become secessionist blocs, and likely form
rival consortia with a large initial burden of resentment and bitterness.
The state you describe as a
potential result is what I consider the current state of ontology development. I
cannot envision how a project aimed at bridging some of the differences would
be able to make the current situation worse – at worst, it could
add one more FO to the set, but if no mappings could be found that are satisfactory
to at least some of the participants, the project would likely end very quickly,
no harm done. As I envision the project, the participants will enter
with a knowledge of the points on which they differ from others, based on the
whole history of this topic, preliminary email discussions, and creation
of a formal proposal by a planning meeting preliminary to the commencement of
the working phase of the project. They will know that issues that cannot
achieve consensus quickly will be resolved by voting, quickly. Those who
hold views irreconcilable to others, knowing that they are in the minority, and
unwilling to make good-faith effort to use the resulting FO, are not likely to
participate. Even so, some drop-out rate should be anticipated in the
plan formulated by the consortium before starting. The goal is to have,
at the end, a diverse group large enough to form a community of users that can
develop interesting open-source applications using ontologies all conforming to
(or translatable into) the structure of the FO. There doesn’t have
to be only one such common FO in the world, there only needs to be *at least*
one used by a wide community – so that those who do want to build
applications with wide semantic interoperability will have some confidence that
they are not developing for a dead-end standard. The numbers of users of
the existing FO’s (OpenCyc, SUMO, DOLCE, BFO) are not large enough to comprise
that type of community – yet. Perhaps their uses will expand to the
critical mass quickly, but the experience of the past five years does not give
me any optimism.
I haven’t heard of a
better process to quickly get an ontology that can serve (by being widely used)
as well to support widespread interoperability. The closest
approach I have seen thus far is the NeOn project, but their membership appears
too limited to achieve a large (> 50) user group that will be able to have
their ontology-driven systems interoperate accurately and automatically.
The mapping methodology they use is inadequate to the task of integrating multiple
ontologies (relying on OWL and horn clauses), but their experience would likely
be valuable in a project of broader scope.
[PC] >> I think that a principle
for construction of the FO should be to include everything that is desired by
someone
[PH] > Stop right there. Let us agree on that. Whether it
is logically consistent with the rest or not. Intransigent logical
inconsistencies require divisions to be drawn between theories, hopefully in
some reasonably principled way and hopefully documented. That is the best we
can hope for, and we know we can do it and that it can be useful. Let us be
content with that.
Whether or not the word “primitive”
was ever mentioned in the project , I would support any project to develop a
common FO with (a) a logically consistent core of at least 3000 elements, in
CL-compliant format; (b) a large enough set of participants (ca. 100); (c) a
goal of providing utilities or bridging axioms to translate different ways of
expressing a concept into other ways; and (d) a requirement that the
participants actually use the ontology in at least one application that does
not merely illustrate data in- data out. Do you know of one?
Pat
Patrick Cassidy
MICRA, Inc.
908-561-3416
cell: 908-565-4053
cassidy@xxxxxxxxx
From:
ontolog-forum-bounces@xxxxxxxxxxxxxxxx
[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Pat Hayes
Sent: Monday, January 12, 2009 1:15 PM
To: [ontolog-forum] ; Patrick Cassidy
Subject: Re: [ontolog-forum] FW: Next steps in using ontologies as
standards
On Jan 12, 2009, at 11:08 AM, Patrick Cassidy wrote:
Yes, F=ma is only a *theory* of motion, and
as I have mentioned in this thread, incompatible theories would have to be
maintained in some extension(s) to the basic FO. Sets of concepts that
are mutually dependent on each other for their description would have to be
located together – whether in the base FO or in some extension to
it. I was pleased that Matthew brought up the case of mathematically
described concepts, and this issue applies to many “derived” units
of measure such as acceleration. But in some cases that may appear to be
co-dependent, one of the elements such a “force” may also
correspond to a more basic concept, and be properly in the base FO as a true
primitive. Force can be directly perceived by the sense of touch, and
things that make direct sensory impressions have, in my view, a privileged
status as prime candidates for primitiveness.
Just as an academic aside: on what basis? Why would this
notion of what one might call 'psychological' primitiveness have any
relationship at all with your earlier notion of 'definitional' primitiveness?
Science seems to have advanced by abandoning sensorially rooted concepts in
favor of much more abstract notions such as symmetry preservation and
conservation laws. Both real science and naive science are fun enterprises, but
it is not a good idea to get them confused.
The physical equations relating to force are not restricted to
its effect on acceleration, but also include things like the compression of a
spring, and the intensity of pressure due to force on a compressible
fluid. So I think that “Force” would be a true primitive,
though “F=ma” would have to be in a ‘theories of
motion’ extension, along with Einstein’s equations. Although
the unit of force is a “derived” unit in the SI system, the effect
of force can be directly perceived, and this is a good reason (I think) to
consider it a primitive. “force” is also one of the words in the
Longman defining vocabulary.
The exact inventory of physical units of measure
that should be in the basic FO would have to be decided by the consortium
constructing it. I have not gone through the list to determine whether
all of the current SI units (meter, kilogram, second, ampere, kelvin, mole,
candela) are the only units that belong in the base FO. The coulomb, for
example, used to be considered as one of the base units (of electrical charge),
but is now considered as a derived unit (one amp-second). I would want
the coulomb to be in the base FO, but there may be arguments against that. If
they are considered non-controversial, I would hope that the consortium decides
to include derived units as well, for convenience. Perhaps if all the
non-controversial derived units were maintained in one extension, it would be
almost as convenient.
All this imagined debate and decision-making about what is
more 'primitive' than what else, illustrates to me the futility of embarking on
this enterprise. Since being 'primitive' is important-sounding but formally
meaningless, and almost certainly ambiguous when stated with more precision,
there will be no way to resolve these debates, and no way other than voting to
decide them. But voting is almost a worst-case way to resolve technical
debates, as it amounts to agreeing not to agree. Strongly held views which are
voted out will become secessionist blocs, and likely form rival consortia with
a large initial burden of resentment and bitterness. The resulting process is
more likely to resemble the history of religion than the development of a
science, complete with doctrinal disputes, high priests and fractal divisions
among sub-congregations with differing agendas. And all of this will be
disputes about nothing, as the entire notion of some concepts being more
primitive than others has no basis in science or engineering, and has never
even been stated with enough clarity to be applied in practice (so its very
definition will give rise to debates as arcane as those over the true meaning
of the Eucharist, and cause splits and divisions between congregations.) This is
a complete waste of time.
I think that a principle for construction of
the FO should be to include everything that is desired by someone
Stop right there. Let us agree on that. Whether it is
logically consistent with the rest or not. Intransigent logical inconsistencies
require divisions to be drawn between theories, hopefully in some reasonably
principled way and hopefully documented. That is the best we can hope for, and
we know we can do it and that it can be useful. Let us be content with that.
, as long as it is logically consistent with the rest. If
it is clearly a derived concept (not primitive as defined by the consortium),
and is clearly relevant to some restricted domain, then the consortium *may
or may not* decide to move it to that domain extension, even over the
objections of some contributors. This reflects the practical need to
resolve disputes quickly. The issue of what is or is not a primitive may
have to be resolved by majority vote in some cases.
Pat Hayes and Pat Cassidy,
These statements are only true for "Newtonian"
conceptualizations of time.
Relativistic models of "space-time" require the
specification of both spatial and
temporal locations. Relativistic temporal models arise in
astronomy, astrophysics,
In a purely computer science context see the classic paper by
Leslie Lamport
on Time, Clocks, and the Ordering of Events in a Distributed System,
Communications of the ACM, vol. 21, no. 7, p. 558-565, July 1978
On Jan 11, 2009, at 10:18 AM, Pat Hayes wrote:
When you say “Those
various temporal theories can all be expressed in terms of three
concepts: time-point, time-interval and duration.” , what do you mean by
that? Is being “expressed in terms of” used only for
necessary and sufficient definitions?
No, I mean only that each theory
uses only those three terms, or can be reformulated using only those (or in
some cases only two of them.) Of course, since the theories have different
axioms, they assign somewhat different meanings to them.
Thinking more, this isn't really
accurate. They all use only these three classes,
but they also use various different relations, such as the timepoint and
subinterval orderings and the various relations between points and intervals.
So perhaps the typical number of concepts is more like six or seven than two or
three.
_________________________________________________________________
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
------------------------------------------------------------
IHMC
(850)434
8903 or (650)494 3973
40 South Alcaniz St. (850)202
4416 office
Pensacola
(850)202 4440 fax
FL 32502
(850)291 0667 mobile
|
_________________________________________________________________
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 (01)
|