Why a single unifiied ontology is impossible
John F Sowa
representation, I still had some lingering hopes for a unified
top-level ontology of everything. But those hopes were always
running into inconvenient facts.
This morning, I came across one more of those facts: the need
for collaboration among specialists with diverse backgrounds.
It came from an article about the evolution of flightless birds:
> “It’s exciting what can be done with a research team with diverse
> skill sets,” Edwards added. “Our group had developmental biologists,
> computational biologists, morphologists, statisticians, population
> geneticists — and, of course, ornithologists. Each brings a different
> perspective and the results, I think, are amazing.”
Implication: An ontology specialized for one purpose must be
supplemented with ontologies designed for other purposes.
For any particular application, unification is best performed
at the *problem level*, not at some abstract universal level.
This point explains why language-based resources, such as WordNet,
are more useful for relating multiple ontologies than any single
top-level ontology that privileges one choice over an infinity
of other options.
That doesn't prove that it's impossible to have a universal
ontology of ontologies. But it shows yet another obstacle
that none of the currently proposed TLOs have begun to address.
By the way, this is a lesson that Wittgenstein learned after
he spent a few years teaching children in an Austrian mountain
village: Children don't think or learn according to any TLO.
Children *and* adults think in terms of "language games".
Wittgenstein never rejected logic. What he rejected was the claim
of an ideal universal top-level ontology. I believe he was right.
See below for some excerpts and the URL of the article.
John
_______________________________________________________
The evolution of flightless birds
Since Darwin’s era, scientists have wondered how flightless birds like
emus, ostriches, kiwis, cassowaries, and others are related, and for
decades the assumption was that they must all share a common ancestor
who abandoned the skies for a more grounded life...
By the early 2000s, new research using genetic tools upended that story,
and instead pointed to the idea that flightlessness evolved many times
throughout history. Left unanswered, however, were questions about
whether evolution had pulled similar or different genetic levers in each
of those independent avian lineages...
Flightless birds all have similar body types, Sackton noted. “They have
reduced forelimbs [wings], to different degrees, and they all have this
loss of the ‘keel’ in their breastbone that anchors flight muscles,” he
said. “What that amounts to is a suite of convergent morphological
changes that led to this similar body plan across all these species.”
“What’s interesting about the morphological changes … is they have to
preserve their hind limbs,” he said. “There are lots of ways to stop a
limb from forming, but shrinking a forelimb without changing the hind
limb is more difficult.”
“One of the things that was exciting about this project for me
personally was how we were able to bring the computational expertise in
the Informatics Group to bear on this really important question in
evolutionary biology,” said Sackton. “This joining of computational,
statistical genetics with the natural history perspectives is important
for getting the full picture of how these birds evolved.”
https://news.harvard.edu/gazette/story/2019/05/harvard-study-explores-genetics-behind-evolution-of-flightless-birds/
Jon Awbrey
My first year at college the university was holding a cross-campus colloquium
taking its theme from C.P. Snow's Two Cultures all about the need for and the
difficulties of cross-disciplinary communication and collaboration in our day.
The university had recently created three residential colleges focused on the
arts, sciences, and government/history but designed to provide future citizens
with an integrated perspective on how these concentrations fit into the bigger
picture of the modern world.
Long time passing, I found myself returning to these issues around the turn of the
millennium, addressing the ???problem of silos??? and the ???scholarship of integration???
from the perspective of Peirce's and Dewey's pragmatism and semiotics. Here's the
papers Susan A. and I wrote about that:
Conference version:
http://www.iupui.edu/~arisbe/menu/library/aboutcsp/awbrey/integrat.htm
Published version:
https://journals.sagepub.com/doi/abs/10.1177/1350508401082013
I don't know if the brands of ontologies being cranked out today are going to be
the ultimate answer to these problems, but I do think there are applications of
logic, math modeling, and pragmatic semiotics that would certainly help a lot.
Regards,
Jon
On 5/7/2019 11:54 AM, John F Sowa wrote:
> Twenty years ago, when I finished my book on knowledge
> representation, I still had some lingering hopes for a unified
> running into inconvenient facts.
>
> This morning, I came across one more of those facts:?? the need
> It came from an article about the evolution of flightless birds:
>
>> skill sets,??? Edwards added. ???Our group had developmental biologists,
>> perspective and the results, I think, are amazing.???
>
> Implication:?? An ontology specialized for one purpose must be
>
> For any particular application, unification is best performed
> at the *problem level*, not at some abstract universal level.
>
> This point explains why language-based resources, such as WordNet,
> are more useful for relating multiple ontologies than any single
> top-level ontology that privileges one choice over an infinity
> of other options.
>
> That doesn't prove that it's impossible to have a universal
>
> By the way, this is a lesson that Wittgenstein learned after
> he spent a few years teaching children in an Austrian mountain
>
> of an ideal universal top-level ontology.?? I believe he was right.
> See below for some excerpts and the URL of the article.
>
> John
> _______________________________________________________
>
> The evolution of flightless birds
>
> more grounded life...
>
> By the early 2000s, new research using genetic tools upended that story, and instead pointed to the idea that
> flightlessness evolved many times throughout history. Left unanswered, however, were questions about whether evolution
> had pulled similar or different genetic levers in each of those independent avian lineages...
>
> and they all have this loss of the ???keel??? in their breastbone that anchors flight muscles,??? he said. ???What that amounts
> to is a suite of convergent morphological changes that led to this similar body plan across all these species.???
>
> ???What???s interesting about the morphological changes ??? is they have to preserve their hind limbs,??? he said. ???There are
> lots of ways to stop a limb from forming, but shrinking a forelimb without changing the hind limb is more difficult.???
>
> ???One of the things that was exciting about this project for me personally was how we were able to bring the
> computational expertise in the Informatics Group to bear on this really important question in evolutionary biology,???
> said Sackton. ???This joining of computational, statistical genetics with the natural history perspectives is important
> for getting the full picture of how these birds evolved.???
>
> https://news.harvard.edu/gazette/story/2019/05/harvard-study-explores-genetics-behind-evolution-of-flightless-birds/
>
Richard H. McCullough
... language-based resources, such as WordNet,
are more useful for relating multiple ontologies than any single
top-level ontology that privileges one choice over an infinity
of other options.
--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/c57dce10-1bc4-cc7b-3843-95ac0611c0fc%40bestweb.net.
For more options, visit https://groups.google.com/d/optout.
bruces...@cox.net
I found myself trying to synthesize connections between some of the big themes that emerged -- relative to my own general model of cognitive structure -- for example, how an ontology of concepts might connect to and work collaboratively with an ontology of graphics -- both guided by probabilities and "typicality", perhaps in a converging way tending towards "understanding" or "explanation". These elements in combination seem to be a primary ingredient list for a comprehensive model of cognition.
I was wondering -- how to understand the marriage of perception ("graphics") and conceptual or verbal/word-based interpretation in some integrated model? That almost seems like a thumbnail model of science. I thought of the "trellis" concept suggested by Amit Sheth -- an approximately stable general form -- that takes an increasingly specific form as it becomes populated by specific data.
Is that on the right track?
If it is -- maybe (??) this idea does indeed tend toward this tenuous notion of universal ontology -- of a *useful* universal ontology....
Bruce Schuman
Santa Barbara CA USA, 805-705-9174
Weavingunity.net
Azamat Abdoullaev
--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/00ab01d50504%24385302b0%24a8f90810%24%40cox.net.
Stingley, Patrick
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAGCbhFm7WHj-k4ALZiXPVi0JqbgWtKRsks7xEFYug37fxHVPJw%40mail.gmail.com.
bruces...@cox.net
Apologies for this too-long email.
*************************
This is a huge and very complicated subject -- but I think there are some big-picture simple things that can be said about it.
I got interested in "categories and concepts" many years ago, initially from a holistic and graphical point of view. I got into the study of "mystical symbols" when I first went back to the university, and what emerged for me was the idea that many such symbols can properly be understood as a kind of "pre-mathematical intuition” into profound ontological subjects. Clear examples might include the cross, mandala, and yin/yang, and probably some others like axis mundi. I like this model of “mandala” – taken from the “wheel of the dharma” on the national flag of India. http://originresearch.com/interval/index.cfm
Human beings have not yet advanced to the point of collective evolution where the commonly repeated symbols that arise in deep intuition are well-understood. They are seen as art, as fantasy, as emotional or simply as incomprehensible, as "mystery". Scholars like Joseph Campbell and psychologists like Jung have looked carefully at this topic, but perhaps not from the point of view of scientific interpretation. But maybe Ramon Lull was moving in this direction.
John, you said the other day in a previous thread that your experience with people trying to connect "data" and "information" with "wisdom" was that their ideas generally tended towards "mush". I know what you mean; I often feel something similar -- because it seems to me that the enterprise of human civilization has not (yet) established a clear and non-controversial understanding of holistic or graphical thinking. So the common thinking is muddy. This entire domain of analysis itself might be thought of as mush. It's uncharted, pre-scientific, unmapped, still blurry and very experimental. It's the mysterious domain of "religion". Skeptics shake their head and walk away.
This is the connection between analysis and intuition, between empiricism and holism – and how that connection is experienced in a normal human mind. At the linear/empirical/scientific end of the spectrum, things are pretty clear. We understand measurement. But what is at the absolute top of this descending cascade? This is the domain of “the absolute container” – where everybody and everything get confused….
*
As I started thinking about this comment, in the afterglow of the very interesting ontology summit just concluded, with its reference to an array of special-case ontologies -- both graphical and verbal -- perhaps seen in the light of Michael Gruninger's comment (if I understood it right) that we do indeed need broader/more inclusive ontologies -- that there might be a huge missing area in the study of intuition.
What about an ontology of symbols? Pull together 500 or 1000 primary and common symbols from various traditions around the world -- and work on interpreting their meaning. Nils Bohr had the "yin/yang' symbol on his personal coat of arms. What does it mean? What are its ontological implications? It's clearly a mathematical form.
Years ago at Berkeley, I bought an engineering book on the "Smith Chart" -- which was about radio antenna design -- because its large-page format presented multiple alternative versions and interpretations of the yin/yang diagram -- a circle with a large letter S dividing it into two sections – which turned out to have major implications for the behaviour of radio waves.
This kind of symbolism has arisen in in some form in every culture that has been studied. So, I think the explanation is -- people are struggling to understand their own deepest intuitions. These symbols are like mathematical hypothesis based on "best guess" approaches. "I don't know for sure what it means, but this is what I get".
*
Today, many years later, I am still following my guiding holistic intuition. I am continuing to take notes and gather elements on what looks to me like a potent universal container for all possibilities of semantic ontology. It's profoundly "mystical" -- and relates to common mystical symbols like "Uroboros" (snake swallowing its tail) -- but it also seems to directly plug in to the concept of "measurement" in its most basic and unquestionable level.
If this framework is making sense, and moving in the right direction -- it might be an exacting map connecting quantitative and qualitative variables -- showing how they can be mapped across a well-defined spectrum of levels.
My guiding intuition is -- all language emerges as labels for distinctions within the continuum -- and those distinctions can be defined at any level of abstraction. Because we have not (yet) accurately and reliably mapped the higher levels of intuition, and do not as yet have the conceptual apparatus to do so, we find ourselves wandering in a mush of confused ideas.
***
So maybe the issue is "what is meant by the concept of a unified ontology?" I personally am not interested in building a huge dictionary. I want to understand the derivation of words and their meaning. I think it is a universal process, and that all meaning can be understood in approximately the same way -- ie, as a dimensional decomposition descending across levels of analysis – where those levels of analysis are clearly defined as the framework for a universal ontology.
In this context -- special case structures don't really matter. They are no longer logically independent objects -- as I think you are saying they are -- but rather as special-case instances a universal/general form that contains them all and from which they are all derived by a highly coherent and rational process.
My guess is -- we are going to see how this works. This does not imply that all independently evolved domains are going to plug into one another immediately. But their common derivation will provide a powerful source of insight. My guess is, we will find ways to meet at the edges, and evolve common ways to interact that span the borders.
Bruce Schuman
Santa Barbara CA USA, 805-705-9174
Weavingunity.net
-----Original Message-----
From: ontolo...@googlegroups.com <ontolo...@googlegroups.com> On Behalf Of John F Sowa
Sent: Tuesday, May 7, 2019 8:54 AM
To: ontolog-forum <ontolo...@googlegroups.com>
Cc: ontolog...@googlegroups.com
Subject: [ontolog-forum] Why a single unifiied ontology is impossible
Twenty years ago, when I finished my book on knowledge representation, I still had some lingering hopes for a unified top-level ontology of everything. But those hopes were always running into inconvenient facts.
Bruce Schuman
bruces...@cox.net
I agree with this, Patrick, thank you.
I like this notion of “null” as the common point or space. I’ve been thinking the same thing.
And I think it is short-sighted to presuppose that
“any single top-level ontology [must necessarily] privilege one choice over an infinity of other options”
I think the top level should itself implicitly contain “an infinity of options” – and that this make perfect sense when we understand that “every particular parsing of the space” is a special case instance. We are not trying to make a rigid list of word meanings – we’re trying to make language absolutely and infinitely fluent – and derived from a common source.
If we don’t have the time or situation to drive all ambiguity and uncertainty out of a shared abstract concept by dialog – I’d guess that the statistical methods discussed during the Ontolog Summit today point very strongly towards ways to make excellent guesses. Those guys are awesome. This stuff is do-able.
“Since everything in our ontologies derives from the Thing object, they must all have this in common, which means that they are not disjoint, but in fact are all part of a superset ontology.”
Yes, exactly. I want to hammer on this point until we see it very clearly – until we see how all symbolic and algebraic representation can be derived from this common starting point by a fully coherent process.
This kind of method can explain to any degree of detail the huge range of human diversity in all its dimensionality – while fully holding it together around its common source
From: 'Stingley, Patrick' via ontolog-forum <ontolo...@googlegroups.com>
Sent: Tuesday, May 7, 2019 11:58 AM
To: ontolo...@googlegroups.com
Subject: Re: [EXTERNAL] Re: [ontolog-forum] Why a single unifiied ontology is impossible
I disagree that a single unified ontology is impossible. In fact, I believe it is inevitable.
It is not impossible because SUMO exists. Maybe someone can correct me if this doesn't count as a satisfactory example disproving the thesis.
It is inevitable for the following reason. As we learn from suffering with Protege, all of the members of our ontologies derive from the "Thing" object.
In Set Theory, the only thing all sets have in common is the Null Set and nothing can be derived from the Null Set.
Since everything in our ontologies derives from the Thing object, they must all have this in common, which means that they are not disjoint, but in fact are all part of a superset ontology.
v.r.,
Patrick Stingley
On Tue, May 7, 2019 at 1:49 PM Richard H. McCullough <rhmccu...@gmail.com> wrote:
John, you said
... language-based resources, such as WordNet,
are more useful for relating multiple ontologies than any single
top-level ontology that privileges one choice over an infinity
of other options.
I think the synset definitions are an important contributor
to the success of WordNet. These definitions describe the
use of a word in different contexts, and thus define
the concept hierarchy lattice of the ontology.
Without synset definitions, the ontology can be a
very ambiguous structure.
Richard H. McCullough
http://ContextKnowledgeSystems.orgWhat is your context?
On Tue, May 7, 2019 at 8:54 AM John F Sowa <so...@bestweb.net> wrote:
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAGCbhFm7WHj-k4ALZiXPVi0JqbgWtKRsks7xEFYug37fxHVPJw%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.
--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAKKS58u-CfHKFpa30MRJ97pPVZdOiyQ6xOpwtMLS6pY2p8_GYg%40mail.gmail.com.
Nadin, Mihai
Dear and respected colleagues,
Ontology answers the question: What is…? as it applies to everything there is, which is actually in perpetual change. Telling machines what things are is the engineering aspect of ontology engineering. A single unified ontology, which would make the life of engineers so much easier, would presume that we can describe, in whichever manner (words, images, sounds, combinations of all kinds of representations), all there is in its continuous change. While the physical—the subset of all there is that is not alive—might be, within a certain framework, fully and consistently described, the living –by far the larger subset of all there is—is undecidable.
These considerations alone should make us aware that a single unified ontology is for reasons related to our relation to reality not possible. For the same reason a single unified ontology of medicine is a goal defying the nature of the entity it would fully describe if it could.
AI as practiced currently being nothing but ontology made operational, to talk about AI in any creative endeavor (medicine is such an endeavor) is nonsense.
Automating tasks we associate with intelligence (even so we do not understand yet what intelligence is) is not the same with making available a machine that “produces intelligence”waiting to be deployed.
Best wishes.
Mihai Nadin
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/007501d5051c%24e72df6b0%24b589e410%24%40cox.net.
doug foxvog
ontology of everything" to whether a "single unified ontology" is possible
or inevitable. If the top level ontology includes definitions of
"everything", then it would seem not to be possible. If, on the other
hand, it only defines what is necessary to define anything that a
specialized ontology includes, it seems quite possible.
A "top-level" ontology that could cover "everything" would describe a
limited number of types of things. It would describe individuals, types
of individuals, types of types of individuals, predicates, functions,
representations of time and space, ways of representing and inter-relating
4D and 3+1 D objects, tangible vs. intangible things, temporally rigid vs.
variable types and properties, quantities and units of measure,
information structures, how to temporally qualify statements, how to
define subcontexts, how to inter-related contexts, and how to use any of
these sorts of things as arguments to predicates.
It may be useful for the concepts of purposeful agent, artifact, and a few
of its most general sub-types (conceptual work, physical artifact) in the
top level ontology as well.
Note that such a top-level ontology would not decree whether the world is
4D or 3D -- it would provide vocabulary for describing both and allow
lower level ontologies to choose that for themselves.
Such a top-level ontology would not include living things, astronomical
objects, artifacts, chemical elements, agreements, money, biological
species, etc. These would be left for lower level ontologies.
High, but lower level, ontologies would be defined as contexts under the
top level ontology.
John's concern about combining different ontologies ends up being able to
use multiple sub-contexts of the top-level ontology (i.e., lower level
ontologies):
>> "It's exciting what can be done with a research team with diverse
>> skill sets," Edwards added. "Our group had developmental biologists,
>> computational biologists, morphologists, statisticians, population
> Implication: An ontology specialized for one purpose must be
> supplemented with ontologies designed for other purposes.
> For any particular application, unification is best performed
> at the *problem level*, not at some abstract universal level.
ontology of everything" -- just that most terms need to be defined below
that top level.
Multiple ontologies may be unified if they are based in the same way on
the same top-level ontology. Developmental biologists, computational
biologists, morphologists, statisticians, population geneticists, and
ornithologists all would have their own mid-level ontologies. By
performing problem-level analysis in a context that inherits multiple
ontologies from multiple fields, the big win of combining these skill sets
becomes available. Some questions may not need to inherit all of these
ontologies -- just the ones whose synergy allows novel solutions.
-- doug foxvog
On Tue, May 7, 2019 18:15, Nadin, Mihai wrote:
> Dear and respected colleagues,
doug foxvog
only by sub-class hierarchies, but by other relations as well. They
certainly point out that words are poor choices as ontological terms as
the same word is used in multiple synsets. One could use WordNet synsets
as a first pass for terms in an ontology. But if you look closely, you
will discover that multiple words in the same synset often have different
meanings -- some may be more or less specific than others and others could
be represented by Venn Diagram circles that greatly overlap others, but
include instances that others don't and exclude instances that others
include.
But an ontology is far more than just a hierarchy of subclasses. It needs
to include predicates with fixed meanings and properties that allow
statements to be made about terms in the ontology.
An ontology may certainly provide mappings between words and phrases and
terms in the ontology -- with multiple denotations of the same word and
multiple ways of denoting the same term in the ontology.
-- doug foxvog
Paola Di Maio
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/007501d5051c%24e72df6b0%24b589e410%24%40cox.net.
Godfrey Rust
Paola
I share your instinct, but for me its rather like having two conflicting postmodern views of truth: one says that there is no absolute truth, only relative; the other says that there is an absolute truth, but we are trapped in limited contexts and can never know what it is because unlike a GUT it would not be verifiable. Either way the practical outcome is the same.
Godfrey
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAMXe%3DSr4xreNgGA%2Byv_Tr7iqOCYTvtiuDN25rt2WqFjCfvj5HA%40mail.gmail.com.
Azamat Abdoullaev
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/003601d50518%24080de520%241829af60%24%40cox.net.
John F Sowa
Jon
> All ???
I don't understand what you're criticizing. Nothing you wrote
contradicts anything I wrote.
Jon
> and pragmatic semiotics that would certainly help a lot.
my 1984 book, my 2000 book, and the papers and lectures I've
been writing and presenting over the years.
Azamat
> "Why a single unified ontology is possible" has strong evidence
> from hard sciences, as fundamental physics. It is about the
> general theory of everything (gTOE)...
What physicists call a theory of everything is just a theory of
physics. If we had such a theory, it would explain many aspects
of physics. But that won't solve any hard problems of ontology.
Just look at quantum electrodynamics (QED), which is the most
general theory that is available today. But the mathematics
associated with QED is so complex, that nobody uses QED if they
can solve their problems with the old QM of 1926. And nobody
uses QM or relativity if they can solve their problems with good
old-fashioned Newtonian mechanics.
Conclusion: Pure physics has a goal of discovering the fundamental
principles of the universe. But every new discovery opens up more
questions than it answers.
Net result: Applied physics is a hodge-podge of thousands of
special-purpose approximations for every imaginable special case.
If some fairy godmother would magically give the world a grand TOE
tomorrow, it would have no effect on those applications.
Patrick
> In fact, I believe it is inevitable. It is not impossible
> because SUMO exists.
on a project that has to communicate with other systems that
use Cyc, DOLCE, BFO, ISO 15936, or no ontology at all. How
can all the components communicate if they're using different
ontologies?
Patrick
> object, they must all have this in common, which means that
> they are not disjoint, but in fact are all part of a superset
> ontology.
irrelevant whether you call it Thing, Entity, or whatever.
That top axiom provides no information at all. In effect,
it says, "For all x, if x exists, then x is a Thing."
Good luck in using that axiom to integrate the data coming
from all those independently developed sources.
Bruce,
> I think the top level should itself implicitly contain
> “an infinity of options” – and that this make perfect sense
> when we understand that “every particular parsing of the space”
> is a special case instance.
systems that use that "infinity of options" and make them
work together (interoperate) without getting mired down in
contradictions?
Bruce
> we’re trying to make language absolutely and infinitely fluent
> – and derived from a common source.
speak thousands of other languages. Formal ontologies specified
in formal logics were designed to enable those systems to work
together. But what if they're using different definitions for
the same words?
Nadin
That is certainly true. And philosophers, linguists, computer
scientists, engineers... have been trying to define such a thing
for centuries. The database people have been trying to develop
such standards since the 1970s. The DB, AI, and programming
language people got together to unify all those approaches in
1980. The AI people tried again in 1991 and they collaborated
with the DB people. The best thing they got was Common Logic (CL)
which was inspired by SWeLL (Semantic Web Logic Language) by
Tim Berners-Lee. But the description logic gang derailed that
foundation and foisted OWL upon us.
In 1998, the year before I finished my KR book, Klaus Tschira
hosted a grand ontology gathering in Heidelberg. There were
people drom WordNet, Cyc, and participants from in all the above
projects. But we broke up without any consensus. Klaus T. vowed
that he would never again host another ontology gathering.
Nobody has ever proposed an ontology that was so good that
anyone with a rival ontology would accept it. For a survey
with URLs for the sources, see http://jfsowa.com/ikl/ .
Doug
> able to use multiple sub-contexts of the top-level ontology (i.e.,
> lower level ontologies)
to the Cyc strategy. But it is still a TLO. And as we have
found, nobody with a TLO that is their pride and joy will accept
anybody else's pride and joy.
Doug
> not only by sub-class hierarchies, but by other relations as well.
> They certainly point out that words are poor choices as ontological
> terms as the same word is used in multiple synsets.
It is closer to the lowest common denominator of all the words in
that collection.
Doug
> It needs to include predicates with fixed meanings and properties
> that allow statements to be made about terms in the ontology.
terms that are aligned to the same synset can share data if and
only if the conflicting features are not relevant to the application.
For example, the WordNet synsets for times and dates say nothing
about a 4D ontology or a 3+1 D ontology. Therefore, times and
dates can be shared among ontologies that make different choices.
Paola and Godfrey
> I share [Paola's] instinct, but for me its rather like having
> is no absolute truth, only relative; the other says that there
> is an absolute truth, but we are trapped in limited contexts
won't know it until every possible scientific question of any kind
has been asked and answered. That is a very close approximation
to never.
On the other hand, we have been getting better approximations
over the years. Collaboration among them is possible, but any
attempt to force a unification would block further research.
Ian
> Great to see that stated ... as a lesson from Wittgenstein,
> a lesson that few (ie zero) logical positivists got.
Yes. Wittgenstein met with some of the Vienna Circlers.
Rudolf Carnap thought that LW was on his side (from his way
of reading the Tractatus). But LW totally rejected everything
that Carnap said or did. Carnap was trying to get funding for
his vision of "unified science". But fortunately, his lack of
funding saved the world from a hopelessly misguided project.
See http://jfsowa.com/pubs/signproc.pdf
Eric
> The most elaborated "Ontologies" are Mathematical Axiomatics !
> All mathematical objects are defined with axioms.
I certainly agree. And mathematics is the most glaring omission
from nearly all the proposed TLOs. Without mathematics, you can't
have science, engineering, finance, business... You can't design
a house, drive a car, talk on the phone, or count anything.
Azamat
> "Niels Bohr had the "yin/yang' symbol on his personal coat of arms.
> clearly a mathematical form".
unity of two opposites. In his case, that meant that both the
wave theory and the particle theory of light, although contradictory,
are both true. It also covers the Heisenberg uncertainty principle
and other aspects of the many puzzles about quantum mechanics.
For more, google "Niels Bohr yin yang".
In fact, that's a good theme to end on. It shows that a unified
ontology that has fixed, precise, formal definitions is incompatible
with the fundamental principles of physics.
The alternative: A supra level based on an ontology of ontologies.
And with mathematics as the "Queen of the Sciences".
John
Jon Awbrey
dashes and quotation marks. I checked and found my T-bird sent the
right unicodes but somehow the comedy of errors appropriately named
Yahoo! was losing them in transit. Looks like it was a transient,
at least as of a minute ago, so let's hope this works this time.
Re: John Sowa
At: https://groups.google.com/d/topic/ontolog-forum/xRa-_vxwdp4/overview
John, All —
My first year at college the university was holding a cross-campus colloquium
taking its theme from C.P. Snow's Two Cultures all about the need for and the
difficulties of cross-disciplinary communication and collaboration in our day.
The university had recently created three residential colleges focused on the
arts, sciences, and government/history but designed to provide future citizens
with an integrated perspective on how these concentrations fit into the bigger
picture of the modern world.
Long time passing, I found myself returning to these issues around the turn of the
papers Susan A. and I wrote about that:
Conference version:
http://www.iupui.edu/~arisbe/menu/library/aboutcsp/awbrey/integrat.htm
Published version:
https://journals.sagepub.com/doi/abs/10.1177/1350508401082013
I don't know if the brands of ontologies being cranked out today are going to be
Jon
inquiry into inquiry: https://inquiryintoinquiry.com/
academia: https://independent.academia.edu/JonAwbrey
oeiswiki: https://www.oeis.org/wiki/User:Jon_Awbrey
isw: http://intersci.ss.uci.edu/wiki/index.php/JLA
facebook page: https://www.facebook.com/JonnyCache
bruces...@cox.net
JS
> If we had such a theory, it would explain many aspects of physics.
> But that won't solve any hard problems of ontology.
As I understand it, a physical theory is defined in concepts. A mathematical model is a concept. It's symbol, their manipulation, their equations, the foundational definitions that support the mathematics in those equations -- are concepts. Just this morning on the very interesting "curiosity stream" web site where there are many fascinating programs on physics, the speaker mentioned the famous equation "E=mc(squared)". That is a highly compositional structure -- what is energy, what is mass, what is speed -- that equation is an exercise in "compositional semantics". I am not challenging that equation. But is a highly derivative concept -- and very simple compared to many modern physical theories -- like the one I was looking at this morning, on "dark energy".
> Just look at quantum electrodynamics (QED), which is the most general theory
> that is available today. But the mathematics associated with QED is so complex,
> that nobody uses QED if they can solve their problems with the old QM of 1926.
> And nobody uses QM or relativity if they can solve their problems with good
> old-fashioned Newtonian mechanics.
JS
> of the universe. But every new discovery opens up more questions than it answers.
JS
> approximations for every imaginable special case.
> If some fairy godmother would magically give the world a grand TOE tomorrow,
> it would have no effect on those applications.
But what I want to talk about is -- "a grand TOC" -- a Grand Theory of Concepts that defines the medium and substrate and inherent conceptual principles on which or underlying which all physics is defined ("what is our common blackboard?")
And I don’t mean to be doubting physics -- or its methods of correlation with empirical measurement. What I am trying to eliminate is inherent/implicit ambiguity in conceptual form, that is introduced by unconscious assumptions embedded in fundamental and perhaps "axiomatic" principles and methods.
Bruce,
> I think the top level should itself implicitly contain “an infinity of
> options” – and that this make perfect sense when we understand that
> “every particular parsing of the space”
> is a special case instance.
> that use that "infinity of options" and make them work together
> (interoperate) without getting mired down in contradictions?
I'd say that this process has to go through a series of growth and development stages that involve first of all building a body of shared principles. That alone might be very challenging. These principles should not be defined at the level of particular sciences and disciplines -- but at a general and "universal" level. I'd probably start out by suggesting ways to define all abstract concepts in ways that are tightly grounded in measurement. That's a big general principle that should exist at the heart of any universal semantic ontology hoping to fan out like the branches of "the single tree of science and humanities" that this concept implies. Once we are solidly grounded, we can then see what defines the top -- or the "trunk" of the tree, from which all this diversity emerges.
What is abstraction? What is generalization? How are "the disciplines" organized along some single spectrum of levels -- if they can be -- and if so, what would that spectrum look like? I'd say the dimensionality of such a structure can be defined -- perhaps by defining the "types of measurement", that characterize each discipline as per Stanly Smith Stevens.
If this structure was defined on a shared spectrum, there would be something like a universal version of the "trellis" that Amit Sheth discussed yesterday. We would have a shared framework of underlying assumptions and principles on which to grow and align the vines of our diversified inquiries.
Bruce
> We are not trying to make a rigid list of word meanings – we’re trying
> to make language absolutely and infinitely fluent – and derived from a
> common source.
> of other languages. Formal ontologies specified in formal logics were
> designed to enable those systems to work together.
> But what if they're using different definitions for the same words?
But powerful general-purpose definitions CAN emerge -- which then see all special languages as diversified instances of common principles.
That's the kind of TOC I'd like to see emerge.
There's a lot more to say, thanks for the comment and this discussion.
- BRS
Richard H. McCullough
--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/d4f98936-4b06-32dd-865d-a38108268153%40bestweb.net.
Patrick Stingley
Adam Pease
Many thanks for the mention. Others have different opinions but I
see SUMO (http://www.ontologyportal.org) as a clear existence proof that
a common ontology is possible. After nearly two decades of encoding a
wide variety of domains, each new project has required elaboration but
not rework of any significant portion of the ontology.
Much like conventional procedural programming, we've collectively
realized that with some considerable effort, reusable generalizations
are possible.
all the best,
Adam
On 5/8/19 7:04 PM, Patrick Stingley wrote:
> I have invited Adam Pease (the author of SUMO) to weigh in on this
> because after 20 years of working to establish a master ontology
> (including WordNet among others), he’s found a lot of considerations
> that are not apparent until well after the work has begun.
>
> Respectfully,
>
> Patrick Stingley
>
> On May 7, 2019, at 2:58 PM, Stingley, Patrick <psti...@blm.gov
>
>> I disagree that a single unified ontology is impossible. In fact, I
>> believe it is inevitable.
>> It is not impossible because SUMO exists. Maybe someone can correct
>> me if this doesn't count as a satisfactory example disproving the thesis.
>>
>> It is inevitable for the following reason. As we learn from suffering
>> with Protege, all of the members of our ontologies derive from the
>> "Thing" object.
>> In Set Theory, the only thing all sets have in common is the Null Set
>> and nothing can be derived from the Null Set.
>> Since everything in our ontologies derives from the Thing object, they
>> must all have this in common, which means that they are not disjoint,
>> but in fact are all part of a superset ontology.
>>
>> v.r.,
>>
>> Patrick Stingley
>>
>>
>> On Tue, May 7, 2019 at 1:49 PM Richard H. McCullough
>> <rhmccu...@gmail.com <mailto:rhmccu...@gmail.com>> wrote:
>>
>> John, you said
>>
>> top-level ontology that privileges one choice over an infinity
>> of other options.
>> https://groups.google.com/d/msgid/ontolog-forum/c57dce10-1bc4-cc7b-3843-95ac0611c0fc%40bestweb.net.
>> For more options, visit https://groups.google.com/d/optout.
>>
>> --
>> All contributions to this forum are covered by an open-source license.
>> For information about the wiki, the license, and how to subscribe or
>> unsubscribe to the forum, see http://ontologforum.org/info/
>> ---
>> You received this message because you are subscribed to the Google
>> Groups "ontolog-forum" group.
>> To unsubscribe from this group and stop receiving emails from it,
>> send an email to ontolog-foru...@googlegroups.com
>> <https://groups.google.com/d/msgid/ontolog-forum/CAGCbhFm7WHj-k4ALZiXPVi0JqbgWtKRsks7xEFYug37fxHVPJw%40mail.gmail.com?utm_medium=email&utm_source=footer>.
--------
Adam Pease
http://www.ontologyportal.org
http://www.adampease.org
@apease_ontology on Twitter
Kingsley Idehen
Hi Patrick, Many thanks for the mention. Others have different opinions but I see SUMO (http://www.ontologyportal.org) as a clear existence proof that a common ontology is possible. After nearly two decades of encoding a wide variety of domains, each new project has required elaboration but not rework of any significant portion of the ontology. Much like conventional procedural programming, we've collectively realized that with some considerable effort, reusable generalizations are possible. all the best, Adam
Hi Adam,
It has been a while!
Anyway, I quickly loaded your SUMO->Wordnet and SUMO rendition in OWL to our URIBurner instance.
Results:
1. Entity
Type Sampling -- just click !
Kingsley
-- Regards, Kingsley Idehen Founder & CEO OpenLink Software Home Page: http://www.openlinksw.com Community Support: https://community.openlinksw.com Weblogs (Blogs): Company Blog: https://medium.com/openlink-software-blog Virtuoso Blog: https://medium.com/virtuoso-blog Data Access Drivers Blog: https://medium.com/openlink-odbc-jdbc-ado-net-data-access-drivers Personal Weblogs (Blogs): Medium Blog: https://medium.com/@kidehen Legacy Blogs: http://www.openlinksw.com/blog/~kidehen/ http://kidehen.blogspot.com Profile Pages: Pinterest: https://www.pinterest.com/kidehen/ Quora: https://www.quora.com/profile/Kingsley-Uyi-Idehen Twitter: https://twitter.com/kidehen Google+: https://plus.google.com/+KingsleyIdehen/about LinkedIn: http://www.linkedin.com/in/kidehen Web Identities (WebID): Personal: http://kingsley.idehen.net/public_home/kidehen/profile.ttl#i : http://id.myopenlink.net/DAV/home/KingsleyUyiIdehen/Public/kingsley.ttl#this
Adam Pease
Thanks for doing this! The SUMO OWL file is a bit out of date so
I'll see about producing the current version for you over the coming
weeks. It also doesn't include the bulk of the combined SUMO that has
dozens of domain ontologies. But is a very welcome start!
all the best,
Adam
On 5/8/19 11:36 PM, Kingsley Idehen wrote:
> On 5/8/19 3:58 PM, Adam Pease wrote:
>> Hi Patrick,
>> Many thanks for the mention. Others have different opinions but I
>> see SUMO (http://www.ontologyportal.org) as a clear existence proof that
>> a common ontology is possible. After nearly two decades of encoding a
>> wide variety of domains, each new project has required elaboration but
>> not rework of any significant portion of the ontology.
>> Much like conventional procedural programming, we've collectively
>> realized that with some considerable effort, reusable generalizations
>> are possible.
>>
>> all the best,
>> Adam
>
>
> Hi Adam,
>
> It has been a while!
>
> Anyway, I quickly loaded your SUMO->Wordnet and SUMO rendition in OWL to
> our URIBurner instance.
>
> Results:
>
> 1. Entity Type Sampling
>
> 2. SPARQL Query Definition
>>>> You received this message because you are subscribed to the
>>>> Google Groups "ontolog-forum" group.
>>>> To unsubscribe from this group and stop receiving emails from
>>>> it, send an email to
>>>> ontolog-foru...@googlegroups.com
>>>> <mailto:ontolog-forum%2Bunsu...@googlegroups.com>.
>>>> To view this discussion on the web visit
>>>> https://groups.google.com/d/msgid/ontolog-forum/c57dce10-1bc4-cc7b-3843-95ac0611c0fc%40bestweb.net.
>>>> --
>>>> All contributions to this forum are covered by an open-source license.
>>>> For information about the wiki, the license, and how to subscribe or
>>>> You received this message because you are subscribed to the Google
>>>> Groups "ontolog-forum" group.
>>>> To unsubscribe from this group and stop receiving emails from it,
>>>> To view this discussion on the web visit
>>>> https://groups.google.com/d/msgid/ontolog-forum/CAGCbhFm7WHj-k4ALZiXPVi0JqbgWtKRsks7xEFYug37fxHVPJw%40mail.gmail.com
>>>> <https://groups.google.com/d/msgid/ontolog-forum/CAGCbhFm7WHj-k4ALZiXPVi0JqbgWtKRsks7xEFYug37fxHVPJw%40mail.gmail.com?utm_medium=email&utm_source=footer>.
>
> --
> Regards,
>
> Kingsley Idehen
> Founder & CEO
> OpenLink Software
> Home Page:http://www.openlinksw.com
> Community Support:https://community.openlinksw.com
> Weblogs (Blogs):
> Company Blog:https://medium.com/openlink-software-blog
> Virtuoso Blog:https://medium.com/virtuoso-blog
> Data Access Drivers Blog:https://medium.com/openlink-odbc-jdbc-ado-net-data-access-drivers
>
> Personal Weblogs (Blogs):
> Medium Blog:https://medium.com/@kidehen
> Legacy Blogs:http://www.openlinksw.com/blog/~kidehen/
> http://kidehen.blogspot.com
>
> Profile Pages:
> Pinterest:https://www.pinterest.com/kidehen/
> Quora:https://www.quora.com/profile/Kingsley-Uyi-Idehen
> Twitter:https://twitter.com/kidehen
> Google+:https://plus.google.com/+KingsleyIdehen/about
> LinkedIn:http://www.linkedin.com/in/kidehen
>
> Web Identities (WebID):
> Personal:http://kingsley.idehen.net/public_home/kidehen/profile.ttl#i
> :http://id.myopenlink.net/DAV/home/KingsleyUyiIdehen/Public/kingsley.ttl#this
>
> All contributions to this forum are covered by an open-source license.
> For information about the wiki, the license, and how to subscribe or
> unsubscribe to the forum, see http://ontologforum.org/info/
> ---
> You received this message because you are subscribed to the Google
> Groups "ontolog-forum" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to ontolog-foru...@googlegroups.com
> <mailto:ontolog-foru...@googlegroups.com>.
> To view this discussion on the web visit
> <https://groups.google.com/d/msgid/ontolog-forum/5aed9d7f-388b-a505-c393-95713f8b01e7%40openlinksw.com?utm_medium=email&utm_source=footer>.
Azamat Abdoullaev
The basis for the merge is ;; is the upper level of Sowa's ontology. The definitions and axioms of the other sources ;; have been mapped into this ontology. Thus far, the merge incorporates Russell and Norvig's ;; ontology, Casati and Varzi's theory of holes, Allen's temporal axioms, the relatively ;; noncontroversial elements of Smith's and Guarino's respective mereotopologies, and the KIF ;; formalization of CPR. Note that this file does not not, as of yet, include Sowa's ;; upper-level ontology.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/9f461d80-5266-b2fa-b7e6-1440f22bc11d%40articulatesoftware.com.
Adam Pease
The latest version of SUMO is in
https://github.com/ontologyportal/sumo. I've tried to keep all prior
versions of SUMO so people can see how it has evolved.
all the best,
Adam
On 5/9/19 11:01 AM, Azamat Abdoullaev wrote:
> Adam,
> I found 52 versions of SUMO, the first one as below.
> Now wonder if the last one makes a final TLO, or there is some real
> world schema behind as a trade secrete. Thanks
>
> The basis for the merge is
> ;; is the upper level of Sowa's ontology. The definitions and axioms of the other sources
> ;; have been mapped into this ontology. Thus far, the merge incorporates Russell and Norvig's
> ;; ontology, Casati and Varzi's theory of holes, Allen's temporal axioms, the relatively
> ;; noncontroversial elements of Smith's and Guarino's respective mereotopologies, and the KIF ;; formalization of CPR. Note that this file does not not, as of yet, include Sowa's ;; upper-level ontology.
>
>
> On Thu, May 9, 2019 at 11:04 AM Adam Pease
> >>>
> >>>> I disagree that a single unified ontology is impossible. In
> fact, I
> >>>> believe it is inevitable.
> >>>> It is not impossible because SUMO exists. Maybe someone can
> correct
> >>>> me if this doesn't count as a satisfactory example disproving
> the thesis.
> >>>>
> >>>> It is inevitable for the following reason. As we learn from
> suffering
> >>>> with Protege, all of the members of our ontologies derive from the
> >>>> "Thing" object.
> >>>> In Set Theory, the only thing all sets have in common is the
> Null Set
> >>>> and nothing can be derived from the Null Set.
> >>>> Since everything in our ontologies derives from the Thing
> object, they
> >>>> must all have this in common, which means that they are not
> disjoint,
> >>>> but in fact are all part of a superset ontology.
> >>>>
> >>>> v.r.,
> >>>>
> >>>> Patrick Stingley
> >>>>
> >>>>
> >>>> On Tue, May 7, 2019 at 1:49 PM Richard H. McCullough
> >>>> <rhmccu...@gmail.com <mailto:rhmccu...@gmail.com>
> >>>>
> >>>> John, you said
> >>>>
> >>>> ...language-based resources, such as WordNet,
> >>>> are more useful for relating multiple ontologies than any
> single
> >>>> top-level ontology that privileges one choice over an
> infinity
> >>>> of other options.
> >>>> **//___^
> >>>> I think the synset definitions are an important contributor
> >>>> to the success of WordNet. These definitions describe the
> >>>> use of a word in different contexts, and thus define
> >>>> the concept hierarchy lattice of the ontology.
> >>>> Without synset definitions, the ontology can be a
> >>>> very ambiguous structure.
> >>>> Richard H. McCullough
> >>>> http://ContextKnowledgeSystems.org
> >>>> What is your context?
> >>>>
> >>>>
> >>>> On Tue, May 7, 2019 at 8:54 AM John F Sowa
> <so...@bestweb.net <mailto:so...@bestweb.net>
> >>>> You received this message because you are subscribed
> to the
> >>>> Google Groups "ontolog-forum" group.
> >>>> To unsubscribe from this group and stop receiving
> emails from
> >>>> it, send an email to
> >>>> ontolog-foru...@googlegroups.com
> >>>> <mailto:ontolog-forum%2Bunsu...@googlegroups.com
> <mailto:ontolog-forum%252Buns...@googlegroups.com>>.
> >>>>
> https://groups.google.com/d/msgid/ontolog-forum/c57dce10-1bc4-cc7b-3843-95ac0611c0fc%40bestweb.net.
> >>>> For more options,
> visithttps://groups.google.com/d/optout
> >>>> --
> >>>> All contributions to this forum are covered by an
> open-source license.
> >>>> For information about the wiki, the license, and how to
> subscribe or
> >>>> unsubscribe to the forum,
> >>>> You received this message because you are subscribed to
> the Google
> >>>> Groups "ontolog-forum" group.
> >>>> To unsubscribe from this group and stop receiving emails
> from it,
> >>>> send an email
> <mailto:toontolog-forum%2Bunsu...@googlegroups.com>
> >>>> <mailto:ontolog-foru...@googlegroups.com
> >>>> To view this discussion on the web visit
> >>>>
> >>>>
> <https://groups.google.com/d/msgid/ontolog-forum/CAGCbhFm7WHj-k4ALZiXPVi0JqbgWtKRsks7xEFYug37fxHVPJw%40mail.gmail.com?utm_medium=email&utm_source=footer>.
> >>>> For more options, visithttps://groups.google.com/d/optout
> > <mailto:ontolog-foru...@googlegroups.com
> > To view this discussion on the web visit
> >
>
> >
> <https://groups.google.com/d/msgid/ontolog-forum/5aed9d7f-388b-a505-c393-95713f8b01e7%40openlinksw.com?utm_medium=email&utm_source=footer>.
> > For more options, visit https://groups.google.com/d/optout.
>
> --
> --------
> Adam Pease
> http://www.ontologyportal.org
> http://www.adampease.org
> @apease_ontology on Twitter
>
> --
> All contributions to this forum are covered by an open-source license.
> For information about the wiki, the license, and how to subscribe or
> unsubscribe to the forum, see http://ontologforum.org/info/
> ---
> You received this message because you are subscribed to the Google
> Groups "ontolog-forum" group.
> To unsubscribe from this group and stop receiving emails from it,
> For more options, visit https://groups.google.com/d/optout.
>
Richard H. McCullough
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/d975a2e1-41cf-23a3-756b-3764ec07170e%40articulatesoftware.com.
Kingsley Idehen
Patrick
I disagree that a single unified ontology is impossible.
In fact, I believe it is inevitable. It is not impossible
because SUMO exists.
SUMO is fine for your applications. But suppose you're working
on a project that has to communicate with other systems that
use Cyc, DOLCE, BFO, ISO 15936, or no ontology at all. How
can all the components communicate if they're using different
ontologies?
Hi John,
It is possible to cross-reference terms across ontologies and then use reasoning and inference for harmonization (or disparate data virtualization) [1].
[1] https://community.openlinksw.com/t/applying-reasoning-inference-to-equivalent-class-relationship-type-semantics/907
-- equivalent class reasoning across FOAF and Schema.org
Ontologies (this is a very simple example; so much more is
possible across ontologies using the same approach).
Azamat Abdoullaev
--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/e549915c-f7cf-bbb4-d085-f8584572f54f%40openlinksw.com.
bruces...@cox.net
I am encouraged by this conversation and by the new voices that are showing up.
I have been inclined to suggest a kind of “ontology summit” that is intended to explore the issues involved in collaboratively developing a “single unified ontology”.
Yes, this is exploratory, yes this is challenging, yes there are feasibility and complexity issues.
As I have settled into this thought over the past week or so, responding to comments here, and attending the very interesting ontology summit seminars – and keeping an eye on what is going on right now in US politics – I have continued to develop what might (?) be a feasible approach to this subject.
Could a network emerge interested in exploration and development, if approached in the right way?
This morning, I am writing a detailed response in dialog format to the Wikipedia article on upper ontology – which discusses the issue in general, and then presents two arguments – one against the feasibility of a unified upper ontology, and another argument defending the feasibility.
My article might be a 15-page word.docx document, but it would be interesting for me to get some comments and response. I might have it done by noon.
I like the Wikipedia article. I think it is well-written, responsible, expert, and fertile.
I think it makes sense for anyone interested in upper ontology – pro or con – to be acquainted with its arguments. Take a look if you have ten minutes.
https://en.wikipedia.org/wiki/Upper_ontology
In its list of working upper ontologies, the first one listed is John Sowa’s (1999). The second is CYC. BFO, Sumo, and WordNet are also discussed, as is the “Process Specification Language” (PSL) created by the National Institute for Standards and Technology (NIST).
From: ontolo...@googlegroups.com <ontolo...@googlegroups.com> On Behalf Of Azamat Abdoullaev
Sent: Thursday, May 9, 2019 7:05 AM
To: ontolog-forum <ontolo...@googlegroups.com>
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAKK1bf_m6hbN3g2H-LezVhOeEMf4VthvYvYo_6JnjyuRDwu1%2BQ%40mail.gmail.com.
Adam Pease
Glad you've taken a look at the project. We've actually had the same
approach of multiple composable files in SigmaKEE and with separate
domain ontologies as part of SUMO since roughly the year 2002. It's a
straightforward approach that allows us to treat different and possibly
conflicting sets of knowledge in a formal inference framework much the
same way one could query a set of human experts. Each shares a large
amount of knowledge, but what they disagree on is held separately, and
may be queried and result in different conclusions, with different
proofs that describe their justifications for those conclusions.
all the best,
Adam
On 5/9/19 3:17 PM, Richard H. McCullough wrote:
> The SUMO knowledge base has made
> significant improvements since 2014.
>
> A quick scan of the SUMO portal first page
> > >>>> You received this message because you are
> subscribed
> > to the
> > >>>> Google Groups "ontolog-forum" group.
> > >>>> To unsubscribe from this group and stop receiving
> > emails from
> > >>>> it, send an email to
> > >>>> ontolog-foru...@googlegroups.com
> <mailto:ontolog-forum%2Bunsu...@googlegroups.com>
> > <mailto:ontolog-forum%2Bunsu...@googlegroups.com
> <mailto:ontolog-forum%252Buns...@googlegroups.com>>
> > >>>>
> <mailto:ontolog-forum%252Buns...@googlegroups.com>
> > <mailto:ontolog-forum%252Buns...@googlegroups.com
> <mailto:ontolog-forum%25252Bun...@googlegroups.com>>>.
> > >>>>
> >
> https://groups.google.com/d/msgid/ontolog-forum/c57dce10-1bc4-cc7b-3843-95ac0611c0fc%40bestweb.net.
> > >>>> For more options,
> > visithttps://groups.google.com/d/optout
> <http://groups.google.com/d/optout>
> > <http://groups.google.com/d/optout>.
> > >>>>
> > >>>> --
> > >>>> All contributions to this forum are covered by an
> > open-source license.
> > >>>> For information about the wiki, the license, and
> how to
> > subscribe or
> > >>>> unsubscribe to the forum,
> > seehttp://ontologforum.org/info/
> > >>>> You received this message because you are
> subscribed to
> > the Google
> > >>>> Groups "ontolog-forum" group.
> > >>>> To unsubscribe from this group and stop receiving
> emails
> > from it,
> > >>>> send an email
> > toontolog-for...@googlegroups.com
> <mailto:toontolog-forum%2Bunsu...@googlegroups.com>
> <mailto:toontolog-forum%252Buns...@googlegroups.com>>
> > >>>> <mailto:ontolog-foru...@googlegroups.com
> > <mailto:ontolog-forum%2Bunsu...@googlegroups.com
> <mailto:ontolog-forum%252Buns...@googlegroups.com>>>.
> > >>>> To view this discussion on the web visit
> > >>>>
> >
> <mailto:ontolog-forum%252Buns...@googlegroups.com>>
> > > <mailto:ontolog-foru...@googlegroups.com
> > <mailto:ontolog-forum%2Bunsu...@googlegroups.com
> <mailto:ontolog-forum%252Buns...@googlegroups.com>>>.
> > > To view this discussion on the web visit
> > >
> >
> >
> > >
> >
> <https://groups.google.com/d/msgid/ontolog-forum/5aed9d7f-388b-a505-c393-95713f8b01e7%40openlinksw.com?utm_medium=email&utm_source=footer>.
> > > For more options, visit https://groups.google.com/d/optout.
> >
> > --
> > --------
> > Adam Pease
> > http://www.ontologyportal.org
> > http://www.adampease.org
> > @apease_ontology on Twitter
> >
> > --
> > All contributions to this forum are covered by an open-source
> license.
> > For information about the wiki, the license, and how to
> subscribe or
> > unsubscribe to the forum, see http://ontologforum.org/info/
> > ---
> > You received this message because you are subscribed to the
> > Groups "ontolog-forum" group.
> > To unsubscribe from this group and stop receiving emails from it,
> > send an email to ontolog-foru...@googlegroups.com
> <mailto:ontolog-forum%2Bunsu...@googlegroups.com>
> <mailto:ontolog-forum%252Buns...@googlegroups.com>>.
> > To view this discussion on the web visit
> >
> > For more options, visit https://groups.google.com/d/optout.
> >
>
> --
> --------
> Adam Pease
> http://www.ontologyportal.org
> http://www.adampease.org
> @apease_ontology on Twitter
>
> --
> All contributions to this forum are covered by an open-source license.
> For information about the wiki, the license, and how to subscribe or
> unsubscribe to the forum, see http://ontologforum.org/info/
> ---
> You received this message because you are subscribed to the Google
> Groups "ontolog-forum" group.
> To unsubscribe from this group and stop receiving emails from it,
> For more options, visit https://groups.google.com/d/optout.
>
> All contributions to this forum are covered by an open-source license.
> For information about the wiki, the license, and how to subscribe or
> unsubscribe to the forum, see http://ontologforum.org/info/
> ---
> You received this message because you are subscribed to the Google
> Groups "ontolog-forum" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to ontolog-foru...@googlegroups.com
> <https://groups.google.com/d/msgid/ontolog-forum/CAGCbhFnqqnGOwt_r0MBRv0yJHGKp7-d5cU3n1Hw46CZ-rP_k%3Dw%40mail.gmail.com?utm_medium=email&utm_source=footer>.
Matthew West
Dear Bruce,
You won’t achieve a single Top Level Ontology, because different ones make different ontological choices that are inconsistent with each other. What might be achievable is collecting the top level ontologies into families that make the same ontological choices, and then map between those families.
Regards
Matthew West
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/006701d50688%24a32b4ae0%24e981e0a0%24%40cox.net.
bruces...@cox.net
Thank you, Matthew. That subject comes up in the Wikipedia article, and the author does make some suggestions regarding this issue.
I think one possible source of confusion has to do with what might be happening at the top level – which the Wikipedia author and I am suggesting should be approached through negotiation and very high-level generalizations – probably taken from cognitive science and concept theory or semantics, or maybe very basic mathematics (“what is a distinction in the continuum, and how is a label or value assigned to it?”)
In many ways I am new to this subject – and might be using the term “ontology” in a slightly different way than most professional working semantic ontologists. I agree, absolutely, that what is at the top is pretty mysterious, and has to be very profound and universal to be workable under all circumstances. So – this idea might be significantly different than what is at the top level of most working special-case ontologies.
MW
> What might be achievable is collecting the top level ontologies into families that make the same ontological choices, and then map between those families.
I like your idea and suggestion. It is an unpretentious and workable starting point. What would happen if we explored negotiation between representatives of these ontologies? I do have some ideas on that I might share tomorrow. “Start slow, ask questions, one step at a time, what do people want to do?”
Richard H. McCullough
Godfrey Rust
Except that the top level hierarchy you have drawn directly contradicts the two conceptual definitions you have written, which imply that “relation” is a subclass of a top level element “entity”, and neither of them is a subclass of “universe” (whatever it is you mean by that).
Godfrey
From: ontolo...@googlegroups.com <ontolo...@googlegroups.com>
On Behalf Of Richard H. McCullough
Sent: 10 May 2019 06:17
To: ontolo...@googlegroups.com
Subject: Re: [ontolog-forum] Re: Why a single unified ontology is impossible
To all,
--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to
ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/BYAPR06MB62157FA4B18D2785A9D4A2ADCB0C0%40BYAPR06MB6215.namprd06.prod.outlook.com.
Azamat Abdoullaev
--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/BYAPR06MB62157FA4B18D2785A9D4A2ADCB0C0%40BYAPR06MB6215.namprd06.prod.outlook.com.
Matthew West
Dear Bruce,
Thank you, Matthew. That subject comes up in the Wikipedia article, and the author does make some suggestions regarding this issue.
I think one possible source of confusion has to do with what might be happening at the top level – which the Wikipedia author and I am suggesting should be approached through negotiation and very high-level generalizations – probably taken from cognitive science and concept theory or semantics, or maybe very basic mathematics (“what is a distinction in the continuum, and how is a label or value assigned to it?”)
[MW>] I don’t think negotiation is going to work. There is not room for compromise, the choices are distinct and non-scalar. Here are some examples:
- Nominalist: There are no universals
- Conceptualist: It’s all in the mind
- Realist: There is an external reality
- Presentist: All that exists is what exists now
- Three Dimensionalist: There are continuants and occurrents where continuants are wholly present at each point in time they persist (so they do not have temporal parts) whilst occurrents occur over time (so have temporal parts)
- Four Dimensionalist: Objects extend in time as well as space (so they do have temporal parts and there are no continuants)
- Possibilist: Possible as well as actual objects exist
- Actualist: recognizes only actually existing entities (at the level of both instances and universals)
Generally you have to pick one from each group. As an example of the different choices here are the choices made by BFO and ISO 15926.
Basic Formal Ontology
- Realist
- Three Dimensionalist
- Continuants and occurrents
- Temporal relationships
- Actualist
- Both sense and reference
- Intensional types
- Material entities and occurrents are extensional
ISO 15926-2/HQDM/IDEAS
- Realist
- Four Dimensionalist
- Spatio-temporal extents
- States and simple relationships
- Possibilist
- No sense without reference
- Extensional sets/classes
- Spatio-temporal extents are extensional
These choices affect how such ordinary things as plans, physical objects and information are handled.
In many ways I am new to this subject – and might be using the term “ontology” in a slightly different way than most professional working semantic ontologists. I agree, absolutely, that what is at the top is pretty mysterious, and has to be very profound and universal to be workable under all circumstances. So – this idea might be significantly different than what is at the top level of most working special-case ontologies.
[MW>] It certainly is.
MW
> What might be achievable is collecting the top level ontologies into families that make the same ontological choices, and then map between those families.
I like your idea and suggestion. It is an unpretentious and workable starting point. What would happen if we explored negotiation between representatives of these ontologies? I do have some ideas on that I might share tomorrow. “Start slow, ask questions, one step at a time, what do people want to do?”
[MW>] Yes, you can then hope to compare how they address the same things, and how well they actually manage to be universal (hint, many of them are more limited than they would have you believe).
ISO 21838-1 sets out a standard for documenting Top Level Ontologies. BFO already has a draft of conformant documentation, and I notice early work on DOLCE and I think PSL has recently been published. It will be interesting to see how this progresses as it shines a light on what we have at present.
Regards
Matthew West
--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/004801d506b1%24b0a2b950%2411e82bf0%24%40cox.net.
bruces...@cox.net
I am continuing to explore this issue, and very much appreciate Matthew’s interesting comments and suggestions. I want to review those issues and “division points”, and consider what might (?) be a possible path towards reconciliation. I note that an announcement posted on Ontolog immediate following Matthew’s message mentioned an upcoming conference workshop where these issues will be discussed. I thought that was fascinating and very encouraging timing.
Just for the moment, I want to send my experimental top-level hypothesis in a graphic format. Yes, this is very brief and raises many complex and controversial issues. And I am not proposing it – merely suggesting it as an object for creative contemplation. But this point of view does suggest a top-level integration where all (supposedly all conceivable) facets of diversity in conceptual structure can be integrated and mapped into a common form.
***
***
Azamat Abdoullaev
Note the nominal level is the lowest measurement level used from a statistical point of view.
Again, in statistics, all these levels of measurement are aimed to process, classify, rank, or specify DATA, while Ontology is to process, classify, rank, or specify ENTITY IN THE WORLD, as basjc kinds, classes, or clusters of things.
What might be really innovative is Statistical Ontology representing the World in terms Data Sets or Data Structures using Entity Categorization, scales of measure and statistical modeling and graphical techniques.
The statistical domain ontologies tend to be populations and samples and use cases, variables and values, data sets and big data, as well as complex applied ontologies, as artificial neural networks, common frameworks for machine learning algorithms.
Statistical Ontology is to underpin not only the mathematical science of statistics, but its many future applications, as Statistical AI, Computational statistics, or Statistical Computing, Statistical learning theory, etc, all to progressively represent/measure the world and its content as data, information, knowledge and wisdom.
--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/012401d50827%24fffcec10%24fff6c430%24%40cox.net.
bruces...@cox.net
BS
> I am fascinated by this notion of “universal container”.
MW
> The real question is not the container, it is what you put in it.
JS
> I agree. But I would say that it's irrelevant what you call it.
Some basic questions:
What's in it? How did it get that way? Why? How can we describe it?
What does it mean for us? What can we do about it? What should we do about it?
What are the implications of the various options? Why should we care?
BS
- What’s in it?
- A catalog of differing perspectives on fundamental ontological issues, plus proposals for reconciling those differences into an integrated comprehensive framework
- How did it get that way?
- We got our heads together and we built it – using everything we know or can discover, and everything everybody brought to the table from any point of view
- Why?
- Why we built it: because it is obviously needed in the world, plus its fun and very interesting
- Why it got that way: because we been around the block and we think it’s cool, plus we think the alternatives are a bore. Why else did we get born?
- How can we describe it?
- We can list the elements
- We can describe how they are connected, and why they are connected, and what they can do when they are connected, and why that is a step up for everybody concerned
- We can list the participants
- We can list their sources and inspirations and teachers and precedents
- What does it mean for us?
- It means we took one of the greatest and most interesting challenges in philosophic and scientific history and turned it into a working actual-problem-solving technology for the benefit of everybody
- What can we do about it?
- Spread the word, get more participation, solve more particular problems, become involved with the greatest initiatives moving across the planet today
- What should we do about it?
- Unfold and explore the fullest range of its implications, into all aspects of civilization and individual human life. Keep babies happy and healthy, support mothers, give fathers a day off now and then
- What are the implications of the various options?
- We meet objections on their own terms, we don’t try to convert them to another viewpoint, we listen, we take notes, we understand them, we get it. We build an interface. We listen, we talk nice, we integrate and incorporate, we hook A and B together, and both A and B like it and it works for both of them. The tide and the boats rise.
- Why do we care?
- Because we are decent human beings and that is what we do. Plus it is interesting, and it looks good on the resume.
And oh yes, from the back cover of Applied Ontology: An Introduction, by Barry Smith and Katherine Munn, http://ontology.buffalo.edu/AppliedOntology.pdf
Ontology is the philosophical discipline which aims to understand how things in the world are divided into categories and how these categories are related together. This is exactly what information scientists aim for in creating structured, automated representations, called 'ontologies,' for managing information in fields such as science, government, industry, and healthcare. Currently, these systems are designed in a variety of different ways, so they cannot share data with one another. They are often idiosyncratically structured, accessible only to those who created them, and unable to serve as inputs for automated reasoning. This volume shows, in a nontechnical way and using examples from medicine and biology, how the rigorous application of theories and insights from philosophical ontology can improve the ontologies upon which information management depends.
*******
I've been wanting to reply to Matthew's suggestions for a few days, as I build out some way to approach his concerns and incorporate his suggestions.
For me, this notion of "container" makes a lot of sense. So, if we are pursuing a unified ontology, what would it contain?
Let's say we start with a few broad definitions -- what is ontology, what is "semantic ontology", what is an "upper" (or "foundational") ontology, what are its components?
I'm happy to start with Wikipedia definitions. These pages sketch out the subject area and some of the common ground, which I cite below in descending order of generality:
https://en.wikipedia.org/wiki/Ontology -- broadest "container" – this highlighted sentence generally points in our specific direction
Ontology is the philosophical study of being. More broadly, it studies concepts that directly relate to being, in particular becoming, existence, reality, as well as the basic categories of being and their relations. Traditionally listed as a part of the major branch of philosophy known as metaphysics, ontology often deals with questions concerning what entities exist or may be said to exist and how such entities may be grouped, related within a hierarchy, and subdivided according to similarities and differences.
https://en.wikipedia.org/wiki/Ontology_(information_science) – Our general topic in more specific terms
In computer science and information science, an ontology encompasses a representation, formal naming and definition of the categories, properties and relations between the concepts, data and entities that substantiate one, many or all domains.
https://en.wikipedia.org/wiki/Ontology_components -- a list of particular objects and elements in some specific context or domain
Contemporary ontologies share many structural similarities, regardless of the language in which they are expressed. Most ontologies describe individuals (instances), classes (concepts), attributes, and relations.
Matthew has cited a list of possible positions on underlying questions and methods, which he sees as incommensurate deal-breakers (without following the helpful approach he suggests).
What I am inclined to do – is review these issues – and perhaps develop a detailed catalog of these concerns and positions – nominalism, realism, conceptualism, etc.
So as predicate, I have to say I don’t see these various positions as unreconcilable – because I see them all as facets of (or perspectives within) a single underlying unity. Our job as I see it is to define this underlying unity – and I personally see this as a reasonable and entirely rational undertaking. We have many sources of insights, ranging from across the entire spectrum of scientific and philosophical history. The times they are a-changing. The failures from last week are an instructive part of the learning curve. This is do-able.
MW> What might be achievable is collecting the top level ontologies into families that make the same ontological choices, and then map between those families.
I need to see this in more specific and concrete terms – but yes, this looks like the right way to start. But let’s also ask along the way how if at all do any of these existing frameworks limit their power by making limiting choices? Why are we compelled to choose from those first three or four choices? Do these choices really make sense in the context of modern science?
Wikipedia redirects a search for “nominalism, realism, conceptualism” to this page:
https://en.wikipedia.org/wiki/Problem_of_universals
There is a long list of related subjects on this page, which might a very good start on a “catalog of philosophic positions” on these questions.
But in my genteel opinion, as application-oriented scientists and engineers, we can cut through this in a hurry. We don’t need to debate whether “properties are real” as per nominalism. From an applied practical point of view, that discussion is a waste of time. Call them “mere heuristics” if that makes you happy. To say that “this actual automobile is real but there are no such things as automobiles” is nonsense for somebody in the automobile business.
HOW TO FIX THIS
Years ago, I came to the conclusion that this entire discussion can be avoided and dismissed by the adoption of a single methodology or observation: What we need to recognize is that this entire business involves two (three) stages: 1) the study of the world, and 2) the representation of facets or qualities or aspects of the world in abstract symbolic models, and (if we are scientists) 3) the testing of those models by empiricism. As scientists and theorists and engineers, we are moving back and forth constantly between the “real” object and our symbolic representation of it, testing and confirming and validating our model. In the end, as computer scientists or ontologists, we must (should) recognize that we are studying the properties of abstract symbolic models, and then comparing those models to reality by testing and “empiricism”. We can come up with unified and non-contentious/non-fragmented ways to do this.
If I really want to stop traffic on this subject, I am inclined to cite the $400 1200-page Springer Handbook of Model Based Science, which gathers top world authorities and authors to review these questions. https://www.amazon.com/Springer-Handbook-Model-Based-Science-Handbooks/dp/3319305255/
The handbook offers the first comprehensive reference guide to the interdisciplinary field of model-based reasoning. It highlights the role of models as mediators between theory and experimentation, and as educational devices, as well as their relevance in testing hypotheses and explanatory functions. The Springer Handbook merges philosophical, cognitive and epistemological perspectives on models with the more practical needs related to the application of this tool across various disciplines and practices. The result is a unique, reliable source of information that guides readers toward an understanding of different aspects of model-based science, such as the theoretical and cognitive nature of models, as well as their practical and logical aspects. The inferential role of models in hypothetical reasoning, abduction and creativity once they are constructed, adopted, and manipulated for different scientific and technological purposes is also discussed. Written by a group of internationally renowned experts in philosophy, the history of science, general epistemology, mathematics, cognitive and computer science, physics and life sciences, as well as engineering, architecture, and economics, this Handbook uses numerous diagrams, schemes and other visual representations to promote a better understanding of the concepts. This also makes it highly accessible to an audience of scholars and students with different scientific backgrounds. All in all, the Springer Handbook of Model-Based Science represents the definitive application-oriented reference guide to the interdisciplinary field of model-based reasoning.
****
Plus, there are emerging conference workshops on these themes at the upcoming JOWO 2019 conference.
I looked at these and find them fascinating and exciting and highly relevant:
The JOWO workshops address a wide spectrum of topics related to ontology research, ranging from Cognitive Science to Knowledge Representation, Natural Language Processing, Artificial Intelligence, Logic, Philosophy, and Linguistics. JOWO is especially suitable for interdisciplinary and innovative formats.
JOWO 2019 website: https://www.iaoa.org/jowo/2019/
The following workshops are being organized:
* 2nd International Workshop on Bad or Good Ontology (BOG).
Chairs: Torsten Hahmann, Rafael Peñaloza, Stefan Schulz, Giancarlo Guizzardi, Oliver Kutz and Nicolas Troquard
* Cognition And OntologieS (CAOS IV).
Chairs: Oliver Kutz, Maria M. Hedblom, Guendalina Righetti, Danielle Porello and Claudio Masolo
* Workshop on Foundational Ontology (FOUST).
Chairs: Antony Galton, Stefano Borgo, Oliver Kutz, Frank Loebe and Fabian Neuhaus
***************************************
[MW>] What might be achievable is collecting the top level ontologies into families that make the same ontological choices, and then map between those families.
I like your idea and suggestion. It is an unpretentious and workable starting point. What would happen if we explored negotiation between representatives of these ontologies? I do have some ideas on that I might share tomorrow. “Start slow, ask questions, one step at a time, what do people want to do?”
[MW>] Yes, you can then hope to compare how they address the same things, and how well they actually manage to be universal (hint, many of them are more limited than they would have you believe).
ISO 21838-1 sets out a standard for documenting Top Level Ontologies. BFO already has a draft of conformant documentation, and I notice early work on DOLCE and I think PSL has recently been published. It will be interesting to see how this progresses as it shines a light on what we have at present.
Regards
Matthew West
Bruce Schuman
Santa Barbara CA USA, 805-705-9174
Weavingunity.net
-----Original Message-----
From: ontolo...@googlegroups.com <ontolo...@googlegroups.com> On Behalf Of John F Sowa
Sent: Monday, May 13, 2019 11:50 AM
To: ontolo...@googlegroups.com
Subject: Re: [ontolog-forum] Re: Yin Yang Symbol
On 5/13/2019 2:03 PM, Matthew West wrote:
> I am fascinated by this notion of “universal container”.
>
> [MW] There is a universal container. It’s called a dustbin (trash
> can). The real question is not the container, it is what you put in it.
I agree. But I would say that it's irrelevant what you call it.
Some basic questions:
What's in it? How did it get that way? Why? How can we describe it?
What does it mean for us? What can we do about it? What should we do about it? What are the implications of the various options?
Why should we care?
John Bottoms
This is one of those "butts in the seat" definitions. And those butts are attached to large brains that contain extensive contexts.
This approach will work with AI if we include concise summaries. Humans also like pyramid structured writing.
This may be why Britannica is now expressed in three levels + Index (Micropædia, Propedia and Micropedia, Index).
-John Bottoms
Concord, MA
--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/012001d509e5%24c3e197c0%244ba4c740%24%40cox.net.
John F Sowa
The methods you suggest are called science. They have produced
an enormous amount of useful information with valuable applications.
But as time progresses, it keeps getting more diversified, not unified.
> # How can we describe it?
>
> * We can list the elements
> * We can describe how they are connected, and why they are connected,
> up for everybody concerned
> * We can list their sources and inspirations and teachers and precedents
All that information has been generated by several millennia of people
working in every branch of science, engineering, business and the arts.
Librarians have been working for about two millennia in gathering and
organizing that information. Most of their work has now been put on
the WWW. And it's being organized and cross referenced in all sorts
of ways. But that hasn't produced a universal ontology.
> # What does it mean for us?
>
> * It means we took one of the greatest and most interesting challenges
> actual-problem-solving technology for the benefit of everybody
plan, for which DARPA was the biggest source of funds. Funding was
renewed in 1989, 1994, and 1999. In 2004, there was a 20-year review.
By that time, there was one thousand person-years of work: twenty
years times 50 people per year. The result: DARPA cut the funding
and the Cyclers were told that they should get more money from
applications. After 15 more years, they have been doing more
practical work. But they haven't yet reached the goals that
they had hoped to achieve by 1994
I keep citing the work that has been done since the 1970s on databases,
knowledge bases, the Semantic Web, and may related project. See
http://jfsowa.com/ikl .
Net results: Ontologies have been useful for many applications.
But the biggest systems, such as Cyc and SUMO, are frameworks with
multiple modules (AKA microtheories) that are specialized for various
purposes. At one point Cyc had 6,000 microtheories. I don't know
how many they have today. SUMO also has a large number of
microtheories.
Those things are closer to families of independent applications with
common interfaces and conventions that enable them to share data.
They are coalitions or federations. They are not unified systems.
And nobody has a clue about how or whether they could ever be
unified.
Some people keep talking about a Manhattan Project for ontology.
In less than four years, the Manhattan Project produced some
big bangs. After 35 years, Cyc has yet to produce a bang.
John
Azamat Abdoullaev
Ontology is the philosophical discipline which aims to understand how things in the world are divided into categories and how these categories are related together. This is exactly what information scientists aim for in creating structured, automated representations, called 'ontologies,' for managing information in fields such as science, government, industry, and healthcare. Currently, these systems are designed in a variety of different ways, so they cannot share data with one another. They are often idiosyncratically structured, accessible only to those who created them, and unable to serve as inputs for automated reasoning. This volume shows, in a nontechnical way and using examples from medicine and biology, how the rigorous application of theories and insights from philosophical ontology can improve the ontologies upon which information management depends.
Applied Ontology: An Introduction, by Barry Smith and Katherine Munn, http://ontology.buffalo.edu/AppliedOntology.pdf
--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/ee1893e2-e650-873f-14c4-008e360a567f%40bestweb.net.
Richard H. McCullough
|
|||
John F Sowa
This thread has gone about as far as it can, and these posts
summarize the range of issues.
Azamat
> physics is the best progressing science due to its never-ending
> unification of fundamental forces and their models and theories.
> This I certify as having PhD in theoretical physics. And its
> ultimate goal is the ToE as presented as a "single, all-encompassing,
> coherent theoretical framework of physics that fully explains and
> links together all physical aspects of the universe".
I agree. A Theory of Everything would be a thing of beauty for
physicists, and everybody else could admire it with awe. But it
would have zero effect on 99.99% of any useful ontology.
String theory, for example, is a strong candidate for a ToE, but
so far, nobody has found a single testable prediction. Even if
physicists could find some weird phenomena somewhere in the
universe, that would not be useful for any applications on earth.
In any case, applied physics (in the many branches of engineering)
is a hodge-podge of inconsistent approximations for every imaginable
application. Even a single major application will require ^many*
such approximations for the design of different parts.
Doug
> An Ontology of Everything would be infinite.
>
> What is desired and possible is a top-level ontology that defines
> sufficient classes, relations, and types of context so that any
> desired types (i.e. "anything") can be modeled in terms of the
> entities in the existing ontology. Such a model would not be
> a complete definition, but would be consistent with the prior
> ontology (in some predefined or newly created context) and the
> thing being modeled.
>
> Such an ontology needs to be able to have sets of contexts that
> are mutually inconsistent.
Yes. This is a good summary of the design that CYC pioneered.
It's also the best known method for organizing and managing any
large ontology that needs to relate the ever-growing diversity
(AKA hodge podge) of technology, business, politics, and life.
The CYC term for a context is 'microtheory', and any ontology
that does not support microtheories or contexts is a toy.
Azamat
> In my humble view, [the book by Maria Keet] has little to do with
> true ontology. I've never read such an idiosyncratic definition...
Unfortunately, that book does not mention CYC, microtheories, or
contexts of any kind. Any student who takes a course based on it
would use OWL to design toys. That is also true of the BFO book.
Since CMK discusses BFO in some detail, I suspect that she was
misled by both the DL dogma of decidabiliity and the BFO dogma
of minimalism (AKA triviality). The CYC developers learned the
hard way: years of painful trial and error by working with a
large number of clients in a wide range of diverse applications.
Some of the CYC lessons: Decidability is irrelevant, a wide range
of inference methods are important for different purposes, and
the top level is important for providing common terminology, but
the most important inferences are performed in the microtheories
In the last count I'm aware of, they had 6,000 microtheories.
Mike
> On a semi-related issue, I would love to one day look at Searle's
> ontology of social constructs and trace back...
Social constructs are fundamental, and they must be accommodated
at the top level. You can trace them back to colonies of bacteria.
The origin of life coincides with the origin of systems that can
interpret and generate signs.
Jack
> An ontology of everything will be possible when we all have
> identical brains.
Any single human brain can accommodate a wide range of inconsistent
contexts. People are very good at compartmentalizing different
aspects of their lives. That enables them to deal with the open-
ended variety of contexts or microtheories in an open-ended variety
of activities.
That ability is essential. But it can also lead to Nazis and
Mafia Dons who love their families while killing others. Ontology
must support ethics and related value judgments at the top level.
Eric,
> But "statistical" behaviors can be more predictable for results
> than "results of individual behaviors" !
I agree. And this is another reason for having a multiplicity
of contexts that may be inconsistent with one another. It's
essential to support different methods of reasoning, different
logics, or even no explicit logic in some of the contexts or
microtheories.
George
> a general ontology of algorithms... would be quite useful,
> especially in building dynamic architectures for the IoT...
> The best generalization I can find is the Wikipedia taxonomy
> of algorithms which is at least a beginning.
I strongly agree. An ontology of ontologies with a systematic
method for relating different ontologies (or microtheories) would
be extremely valuable. But that would be a *federation*, not a
unification.
Bruce
> We should not be talking about the attempt to catalog an infinite
> number of sub-explanations then throwing our hands in the air when
> it doesn't work. We should be generalizing the process of
> explanation itself.
Yes. An ontology of ontologies must have some criteria for
evaluating a federation and the connections among its members.
But that's a topic for a new thread.
John
John F Sowa
I agree with your points. But I was trying to say that even
a single individual has many different views. You can pick
any term you like -- compartmentalized, conflicting, competing,
complementary, supplementary, or just "broad".
> any ontology is likely to convey different meanings to different
> humans unless those humans think alike.
Yes, but...
1. No single individual at different instants will think
in exactly the same way.
2. Even young children (preschool age) can detect their
parents' intentions very quickly. Even cats know
how to control their humans' behavior.
3. The fact that people can laugh at jokes means that they
can recognize and appreciate conflicts and ambiguities.
People from very different cultures laugh at the same
kinds of things. (Hollywood banks on that point.)
4. People are very flexible in adjusting to each other's
preferences, aversions, and idiosyncrasies -- even with
people who speak different languages or come from
different cultures. Computers can't begin to do that.
You can enable a computer system to detect many ambiguities
just by feeding it a big dictionary with all the typical
word senses. But the ability to detect unstated intentions
is critical.
A child or even a cat can do that. You can avoid the word 'vet',
and you can pretend that yearly exam day is just like any other
day. But the cat will be hiding under the bed.
(My niece Leslie is a veterinarian who specializes in cats.
And they're constantly frustrated by customers who miss or come
late to their appointments because they can't catch their cat.)
And -- this is the hardest task -- any system that can understand
language will have to laugh at jokes. When it comes to jokes,
Alexa, Siri, and their friends are clueless. You could train
them to respond to the old joke books, but they couldn't detect
any new ones -- or the ones where "you gotta be there".
> However ‘ethics’ won’t get us there because confirming mutuality
> of ‘what’s GOOD” is not as achievable as confirming mutuality
> of ‘what’s harmful.'
I used the word 'ethics' because it's the shortest word to type.
The broader terms 'value judgments' or 'normative sciences'
cover a wide range. They include the positive (Beauty, Goodness,
Truth, Comfort, Love) and the negative (Ugliness, Pain, Fear,
Hatred, Suffering, Disgust, Repulsion, Revulsion, ...).
Value judgments determine everything we do, say, or think.
But very few ontologies include them. Where are they in BFO?
Do the textbooks about OWL tell us how to represent them?
Or what to do about them?
No theory about space-time or intergalactic strings will
enable our ontologies to handle these issues.
John
bruces...@cox.net
I wonder if some of our differences in this ontolog discussion have to do with different objectives -- or concept of "what ontology is" -- or what "an ontology is" --
"what are we trying to do"?
What I'd like to see emerge is some big-picture conventions about modeling and symbolic representation -- e.g., what is meant by a term like "real"
Where "reality" is located, and what it means when we represent an aspect of reality in symbols -- what is the correlation between that abstract symbolic representation and "the reality it represents" --
Maybe focus particularly on "representation in a computer" -- since that can probably be defined with high precision -- whereas "how people do it" is probably a bottomless blur of variations and alternatives
My sense is -- it might be possible to draw up a shared body of strict conventions about these broad generalities --
If this was all defined in a kind of very simple "general-purpose computer logic" -- a simple body of essentially mathematical definitions or operating system design -- from that body of fairly simple core definitions, it might be possible to branch out in many related and more specific directions: how does this relate to specific domain ontologies, how does this relate to cognitive science and human behaviour, how does this relate to or help illuminate issues in philosophy or epistemology, what does this say about "what concepts are", or "what abstraction is" -- etc. Can we locate various disciplines in this framework? What is "holism"? What is "empiricism"? What is "science"? Could we possibly see all these questions through one framework?
Every time I come back to this subject, and take a deep breath and start with a clean piece of paper -- this is the way it always looks:
There is this big round shiny empty circle called "ontology" -- which we populate with everything human beings are doing -- or talking about, or arguing about -- or killing each other about --
And because we don't see the broad general outlines very well, we just continue to crash blindly into one another, wondering why everything is incommensurate with everything else -- and sometimes persuading ourselves that the problem is baked in and unavoidable...
***
For me, the foundations for a common unity can be derived as an interpretation of the continuum. Everything arises within the continuum -- every idea, every distinction -- and everything is grounded in the continuum
Yes, you could say that is "Logos" -- or if you want, "Tao".
Or some other concepts -- like "One"
If we really wanted to tie the big picture together -- it would probably be a good idea to clarify why this makes sense.
It's not that hard.
And it would be hugely tectonic....
***
Just to note -- for me, all this stuff about "it's all derived from the continuum" actually started with stuff I learned from JS long ago, and have had on the internet for years.
Maybe no author who has evolved their thinking wants their ancient work cited -- but these quotes remain foundational for me. I'd say "Everybody on ontolog should read this." Let's work from there.
http://originresearch.com/sd/sd4.cfm
And oh yes, and importantly -- maybe in this broad context there emerges a significant and inherent foundational principle for ethics....
Bruce Schuman
Santa Barbara CA USA, 805-705-9174
Weavingunity.net
-----Original Message-----
From: ontolo...@googlegroups.com <ontolo...@googlegroups.com> On Behalf Of John F Sowa
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
John F Sowa
> For me, the foundations for a common unity can be derived as an
> interpretation of the continuum. Everything arises within the
> continuum -- every idea, every distinction -- and everything is
> Just to note -- for me, all this stuff about "it's all derived from
That's from my 1984 book, _Conceptual Structures_. I have intended
to reformat it and put it on my web site, but it would take a lot
of work to convert the diagrams and special characters.
But Chapter 7, Limits of Conceptualization, has no diagrams and
just a few special characters. So I decided to convert it to PDF
and upload it: http://jfsowa.com/csbook/cs7.pdf
This chapter has many of the parts you quoted, and I added links
to six recent articles that go into more detail on related issues:
Signs, processes, and language games (2002); The challenge of
knowledge soup(2006); Worlds, models, and descriptions (2006);
What is the source of fuzziness? (2013), Signs and reality (2015);
The virtual reality of the mind (2017).
Chapter 7 and those additional articles show the many reasons why
we won't see a single unified ontology for a long, long time.
However, these articles outline directions to pursue in order to
develop ontologies that are significantly better than BFO + OWL.
I'll send a note about these issues by Monday.
John
bruces...@cox.net
I'd been looking at Maria's book, and something I think she cites by Michael Gruninger on Foundational Ontologies for Units of Measure, and yesterday I discovered another "big book" on concept theory which was $120.00 until I found a complete version in PDF at https://academiaanalitica.files.wordpress.com/2016/10/cohen-h-lefebvre-c-eds-handbook-of-categorization-in-cognitive-science.pdf -- or http://bit.ly/2JO6C45. Yesterday after responding to John's message about "cats hiding under the bed" I started sketching out some possible soft approach to a collaborative universal ontology. John cites Tao and other such at the end of http://jfsowa.com/csbook/cs7.pdf
Is there some common foundation? A foundation that span all disciplines – and indeed, "all religions" – not to mention "all sciences"? Could there be? What would it looks like? How powerful would it have to be to overcome the seemingly endless controversies – some of which Maria cites (empiricist, universalist), some of which Matthew cited the other day:
*****
I woke up last night about 2:30 am, chanted a ritual prayer or two -- and then suddenly found my head (again) cooking up some conceptual integration -- kind of like those famous stories about the creative process. Should I write these ideas down -- or will I remember them in the morning? I got out of bed and turned on the computer. Get this big-picture master theme defined: "concept of all concepts". It's a container (i.e., a boundary) -- but it has an unambiguous and determinate implicit/internal structure. It's all about "bounded intervals". Boundary values. Every conceivable distinction within any possible spectrum of measurement at any level of precision defined in any combination of dimensions. And it directly maps all the big holistic/composite ideas from ontology -- semantic or otherwise. Abstraction, taxonomy, part/whole, alphabets, words, semantics, arithmetic, logic, Boolean algebra, many other elements -- and it’s all contained within and derived from the continuum. Every idea or conceptual structure or symbolic representation of any sort whatsoever emerges from the continuum and can be mapped in infinite micro-detail and micro-precision to it.
How much “uncertainty” or “ambiguity” in natural language can be explained by boundary values in the definition of abstractions? The answer is: a lot.
I see the world drowning in what looks to me like semantic idiocy. You guys paying any attention to the dialogue between Barr and Mueller and the “rule of law”? That entire issue is a poisoned swamp for reasons that are the province of semantic ontology. Where is the semantic ontology of law? So yes, I can be a bit volcanic about this.
If I had my way, we’d round up a small crew with the time and energy to start patching together the big picture across the entire range of ontology – from the highest levels to the most specific. We should take that catalog of supposedly incommensurate positions listed by Matthew and interpret them as alternative perspectives within a common spectrum. Each of them has its reasons and is motivated, each of them has its meaning and value and context and is important in its own way. Taken all together, they form the general framework of ontological inquiry. Maybe we should add a catalog from the philosophy of mathematics (constructivism, intuitionism, etc.)
Put that statement of general principles from John at the top of the process and work from there. Keep it simple. Clarify those few big points and make them foundational.
*
So yes, there are significant differences within this spectrum. In my case, my entire life has been beneficially influenced by these principles I learned from John – but I am a top-down analyst, and I want comprehensive system-based solutions, not local-only heuristics that are mutually incommensurate in a global context. “Do it right – see it from the right point of view -- and it’s not impossible.”
Should we sit here stupidly while the frogs boil, because we can’t see how to connect these alternative perspectives? Are we waiting for “somebody else to do it” – or are we just waiting around “for God to be nice”?
I’d say there is tremendous co-creative force in the air right now – probably in ways that are “historically unprecedented”.
Ok. Waiting to exhale. Thanks.
-----Original Message-----
From: ontolo...@googlegroups.com <ontolo...@googlegroups.com> On Behalf Of John F Sowa
Sent: Friday, May 31, 2019 8:20 PM
To: ontolo...@googlegroups.com; ontolog...@googlegroups.com
Subject: [ontolog-forum] Why a single unified ontology is impossible
On 5/31/2019 12:16 PM, bruces...@cox.net wrote:
> For me, the foundations for a common unity can be derived as an
> interpretation of the continuum. Everything arises within the
> continuum -- every idea, every distinction -- and everything is
> grounded in the continuum...
>
> Just to note -- for me, all this stuff about "it's all derived from
> the continuum" actually started with stuff I learned from JS long ago...