[ontolog-forum] The Great Debate
bruces...@cox.net
I’ve been away from Ontolog for a while, so it was fun today to be on the Bluejeans conference call with John Sowa and Barry Smith. Everybody was muted, with video off, so the drama was bit confined, but I had to laugh at the passion. If I understood it right, John promised to map his “best ontology I have ever seen” (John as quoted by Barry) into the structure of BFO, to show that his preferred model actually contains BFO as a special-case subset, and have it done by tomorrow. Good luck with that. 😊
My ears pick up when I hear mention of the continuum and continuous variation. For me, that’s reality: without distinction. Categories are abstract cognitive constructions composed of distinctions, and imposed on “reality” so we can understand it. And I like hearing from the brave souls willing to talk about “universal ontology” in a serious way. But as the discussion progressed today, I found myself driven back into the same corner – though the positions discussed vary widely in many important ways – in one critical way they are bound together with what I would describe as a common and no doubt professionally responsible methodological assumption: they both presume an inherently specific empirical/immediate context – populated by objects which we must describe in systematic and non-confusing/unambiguous ways. The medical industry is rolling along and people’s lives and big money are a stake. We don’t have time to start exploring universal abstractions that maybe someday we can systematically parse into reliable models of working DNA, or human body parts, or the chemical behaviour of pharmaceuticals. This demanding professional context forces the ontology industry into an endless series of special case models.
I like the distinction between scruffy and neat – but from my point of view, any methodology that starts with empirical observations is inherently scruffy. Looking around, the ontologist says “here’s the stuff I have to categorize”, and he/she starts drawing diagrams and labelling these objects. This method is heuristic – and inevitably leads to what people are calling micotheories and logical subsets. It works great in a specific local context, but try to expand it and the logic strangles. It’s inherently “scruffy”, never “neat”. It’s always “relativistic” (defined with respect to local context) and never “absolute” (defined with respect to the whole).
One obvious problem is that in actual “natural” usage, word meaning is highly context-specific – and by “context” I mean some specific conversation where purposes and definitions can be honed by participants until the meaning of terms becomes sufficiently accurate and unambiguous for purposes of that conversation. In the real world, when people are unsure of meaning, they inquire – in a dialogue process that amounts to “drill-down” – from vaguer generalities to increasingly precise specifics.
I first got the idea of “parsing the continuum under the influence of specific purpose” from John. I think that fundamental principle is absolutely right. That’s how it works. So, I want to see approaches to universal ontology grounded in the undifferentiated continuum, rather than in some highly evolved industry-standard vocabulary – which is professionally responsible in an immediate sense, but lacks the fluency to become universal.
And it’s perfect for me that Jon Awbrey is talking about G. Spencer Brown’s “Laws of Form”, because Brown’s essential idea is the core of my own understanding of how to best model extreme complexity in an absolutely precise zero-ambiguity “parsing of reality” (because definitions defined this way can approach near-continuity to any desired degree). Brown is talking about “distinctions”, and for me, distinction is the primary cognitive act. I like what it says on the Wikipedia page for Laws of Form, down at the bottom: “The biologists and cognitive scientists Humberto Maturana and his student Francisco Varela both discuss LoF in their writings, which identify "distinction" as the fundamental cognitive act.” https://en.wikipedia.org/wiki/Laws_of_Form
So this goes to what I suppose is probably the most fertile approach to a universal ontology: build a compositional semantics where every meaning is defined as a composite of distinctions. Every word is a label for some such composite.
What is a distinction?
“a difference or contrast between similar things or people.
"there is a sharp distinction between domestic politics and international politics"
synonyms: | difference, contrast, dissimilarity, variance, variation; division, differentiation, dividing line, gulf, gap "class distinctions" |
In taxonomy, things in a common taxon are “similar”, but distinguished by measurable/detectible distinctions (they have measurable facets in common, but also have measurable differences). For me, a taxon is a good general illustration of this similarity/difference relationship – both similarity and difference can be measured in numeric dimensions. Objects in a taxon can be linearly ordered by their values in some common dimension.
For me – all of this can be elegantly and “perfectly” grounded by defining distinction in the context of the continuum. The continuum is like the real number line – which in engineering we parse into rational numbers in some specific number of decimal points. It’s a universal “unit interval” with no defined distinctions – until we create one, serving some immediately local purpose. Everything in reality can be described to any degree of precision by this process of stipulation and dimensional measurement.
So the key to developing a universal ontology – this line of reasoning suggests – is to generalize the process of composing and defining abstract objects (the objects of our ontology) in the common language of the continuum, making it possible to define and construct all abstractions in a systematic consistent way – as a nested composite of implicit distinctions with a name/label – which we may have to make explicit when interpretation becomes ambiguous.
Thanks! And if Barry’s slides become available, I’d love to see them.
From: ontolo...@googlegroups.com <ontolo...@googlegroups.com> On Behalf Of bruces...@cox.net
Sent: Tuesday, May 1, 2018 9:47 AM
To: ontolo...@googlegroups.com
Subject: RE: [ontolog-forum] Fwd: pdf
Will Barry Smith’s slides be available for download?
Bruce Schuman
Santa Barbara CA USA
http://charterforcocreation.net
From: ontolo...@googlegroups.com <ontolo...@googlegroups.com> On Behalf Of Ken Baclawski
Sent: Tuesday, May 1, 2018 7:38 AM
To: ontolog-forum <ontolo...@googlegroups.com>
Subject: [ontolog-forum] Fwd: pdf
Here are John Sowa's slides for the Great Debate that is about to start.
Ken Baclawski
--
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.
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.
For more options, visit https://groups.google.com/d/optout.
Matthew West
Dear Bruce,
I’m sorry I was not able to attend the Ontology Summit. My regular timetable just precludes me attending at the usual times. So thank you for your reflections. See some responses from me below.
Regards
Matthew West
I’ve been away from Ontolog for a while, so it was fun today to be on the Bluejeans conference call with John Sowa and Barry Smith. Everybody was muted, with video off, so the drama was bit confined, but I had to laugh at the passion. If I understood it right, John promised to map his “best ontology I have ever seen” (John as quoted by Barry) into the structure of BFO, to show that his preferred model actually contains BFO as a special-case subset, and have it done by tomorrow. Good luck with that. 😊
[MW>] I wonder what John is promising here. I’ll offer three possibilities:
- BFO (and by implication any other top level ontology) is a subset of John’s preferred top level ontology.
This at least is easy to dismiss, because top level ontologies make choices about how to view the world about us. For example, are particulars spatio-temporal extents – extended in time and space, or endurants – wholly present at each point in time they exist, and passing through time, but not extended in time (BFO of course famously has some particulars of each). Clearly, something cannot be both of these, so if you include allowing something to be both of these bit only one of them, then you have an inconsistency. So this won’t work. - BFO is mapped to John’s preferred top level ontology.
It is equally easy to accept that this is likely possible. Since a top level ontology is supposed to cover “life the universe and everything” at least to a first approximation, then if there is some thing that can be categorized by one top level ontology, it should be possible to categorize it in another, and thereby derive a mapping between the top level ontologies. - BFO is a subset of some net of theories.
One of John’s pet ideas is the net of theories. This is a catalogue of all the possible theories you could have, from which you can make various selections to hopefully come up with a subset that are coherent and consistent. Of course adding BFO to such a net of theories and getting it out again should be possible.
My ears pick up when I hear mention of the continuum and continuous variation. For me, that’s reality: without distinction. Categories are abstract cognitive constructions composed of distinctions, and imposed on “reality” so we can understand it. And I like hearing from the brave souls willing to talk about “universal ontology” in a serious way.
[MW>] I’m one of these of course.
But as the discussion progressed today, I found myself driven back into the same corner – though the positions discussed vary widely in many important ways – in one critical way they are bound together with what I would describe as a common and no doubt professionally responsible methodological assumption: they both presume an inherently specific empirical/immediate context – populated by objects which we must describe in systematic and non-confusing/unambiguous ways.
[MW>] The sad thing is that this work has been done. The problem is that people are in too much of a hurry to do stuff to read and understand it. A shame really, because it is quicker and cheaper to read and understand it, and then use it to develop ontologies that are much more stable and reusable, than it is to dive in and “do stuff”, but usually end up in a mess, or at best in a place that is not extensible or reusable in other contexts.
The medical industry is rolling along and people’s lives and big money are a stake. We don’t have time to start exploring universal abstractions that maybe someday we can systematically parse into reliable models of working DNA, or human body parts, or the chemical behaviour of pharmaceuticals. This demanding professional context forces the ontology industry into an endless series of special case models.
[MW>] Exactly.
I like the distinction between scruffy and neat – but from my point of view, any methodology that starts with empirical observations is inherently scruffy. Looking around, the ontologist says “here’s the stuff I have to categorize”, and he/she starts drawing diagrams and labelling these objects. This method is heuristic – and inevitably leads to what people are calling micotheories and logical subsets.
[MW>] Well that is how I started, but I did not go down that route, so at least it is not inevitable.
It works great in a specific local context, but try to expand it and the logic strangles. It’s inherently “scruffy”, never “neat”. It’s always “relativistic” (defined with respect to local context) and never “absolute” (defined with respect to the whole).
[MW>] Quite. I spotted that quite quickly, so started looking for the ways (I was even instructed to look for ways) to avoid the pitfalls. A top level ontology and a methodology for incorporating additional content into it at lower levels so it was reusable was the result.
One obvious problem is that in actual “natural” usage, word meaning is highly context-specific – and by “context” I mean some specific conversation where purposes and definitions can be honed by participants until the meaning of terms becomes sufficiently accurate and unambiguous for purposes of that conversation. In the real world, when people are unsure of meaning, they inquire – in a dialogue process that amounts to “drill-down” – from vaguer generalities to increasingly precise specifics.
[MW>] Yes. A key part of that methodology is analyzing the context, and disambiguating. As a simple example, I’ve just been having a discussion about what a pump is. The question has been around where the boundary is. Clearly it includes the bit that moves the liquid, but does it include the driver as well? From the point of view of the designer/procurer, it is just the bit that moves liquid the “bare pump” because he buys that separately from the drive. However, plant operators always talk about the bare pump and the drive together as the pump, because nothing gets done without both. So it looks like we need two terms, say “bare pump” and “pump assembly” to distinguish these two things, and someone just asking for a pump needs to be asked which of these they mean.
I first got the idea of “parsing the continuum under the influence of specific purpose” from John. I think that fundamental principle is absolutely right. That’s how it works. So, I want to see approaches to universal ontology grounded in the undifferentiated continuum, rather than in some highly evolved industry-standard vocabulary – which is professionally responsible in an immediate sense, but lacks the fluency to become universal.
[MW>] Not sure what I have been doing is either of these.
And it’s perfect for me that Jon Awbrey is talking about G. Spencer Brown’s “Laws of Form”, because Brown’s essential idea is the core of my own understanding of how to best model extreme complexity in an absolutely precise zero-ambiguity “parsing of reality” (because definitions defined this way can approach near-continuity to any desired degree). Brown is talking about “distinctions”, and for me, distinction is the primary cognitive act. I like what it says on the Wikipedia page for Laws of Form, down at the bottom: “The biologists and cognitive scientists Humberto Maturana and his student Francisco Varela both discuss LoF in their writings, which identify "distinction" as the fundamental cognitive act.” https://en.wikipedia.org/wiki/Laws_of_Form
[MW>] Well, it seems to me that any ontology is principally about categories of things that exist, and how to determine if things are members of those categories or not. So distinction is inherent in ontology. I can’t say I feel there is something missing that I need from this work though.
So this goes to what I suppose is probably the most fertile approach to a universal ontology: build a compositional semantics where every meaning is defined as a composite of distinctions. Every word is a label for some such composite.
[MW>] Could you suggest an alternative approach to this? Sorry, this just seems a statement of the blindingly obvious. The question is which categories that you could choose to construct from distinctions, do you choose, and why? The mistake is to think that there is only one set of choices which will naturally emerge with a bit of analysis.
Regards
Matthew West
John F Sowa
I returned home yesterday evening and made some minor revisions to
the slides I presented on May 1. They're in my temp directory, since
they're still under development. (I'm sorry, Matthew, but I typed
"tamp" in my earlier note): http://jfsowa.com/temp/ontostan.pdf
There are many issues to discuss, but I'll start with comments
about this version. First of all, I thank Barry for raising some
good points, which helped me clarify some of the issues.
In particular, I revised the labels on my diagram of the PW TLO
(slide 13) for the PWWWB TLO (slide 14). The abbreviated labels are
more readable, but the full names show that it is an IS-A hierarchy.
I expanded the acronym PW (Peirce-Wilkins) to PWWWB to give credit
to Whitehead, Wittgenstein, and Bertalanffy for critical issues that
are essential for a general-purpose ontology.
Bruce
> had to laugh at the passion...
Yes, The debate was fun, except for a few rough spots.
Bruce
> I have ever seen” (John as quoted by Barry) into the structure of BFO,
> to show that his preferred model actually contains BFO as a special-case
> subset, and have it done by tomorrow.
describe the mapping today (May 3).
Bruce
> of view, any methodology that starts with empirical observations is
> inherently scruffy.
denounce the logicians, whom he called neat. I was one of the neats,
but I have a strong sympathy for the scruffies.
In particular, I strongly agree that *every* empirical science is
inherently scruffy. The science I know best is physics, which has
a few neat theories (which are inconsistent with one another) and
a hodge-podge of thousands of inconsistent approximations that
are used by experimental physicists and engineers. Chemistry is
scruffiness built on scruffiness, biology is more so, and social
scientists who attempt to be neat suffer from "physics envy".
For some comments about those issues, see slides 25, 29, and 32
of ontostan.pdf and Chapter 7 of my Conceptual Structures book:
http://jfsowa.com/pubs/cs7.pdf .
I wrote that book in 1983, but the first page of cs7.pdf has links
to recent articles that go into more detail about related issues.
Matthew (commenting on the idea of mapping PWWWB to BFO)
>
> This at least is easy to dismiss, because top level ontologies make
> choices about how to view the world about us. For example, are
> particulars spatio-temporal extents – extended in time and space, or
> endurants – wholly present at each point in time they exist, and passing
> through time, but not extended in time (BFO of course famously has some
I agree. My recommendation is based on a comment I've made several
times on Ontolog Forum and in discussions with Barry. BFO should be
considered a special-purpose microtheory, not a general-purpose (GP)
TLO. This corresponds to your option 3.
Matthew, option 2.
> 2. BFO is mapped to John’s preferred top level ontology...
> and everything” at least to a first approximation, then if there
> is some thing that can be categorized by one top level ontology,
> it should be possible to categorize it in another, and thereby
> derive a mapping between the top level ontologies.
is widely used to align the categories of independently developed
ontologies. But its success depends on the fact that it's extremely
underspecified. Question: Can we achieve better success with
something that is somewhat more structured than WordNet?
Matthew, option 3.
> 3. One of John’s pet ideas is the net [lattice] of theories. This
> which you can make various selections to hopefully come up with
> a subset that are coherent and consistent.
and in several articles and talks since then. For a brief summary,
see slides 27 and 28 of ontostan.pdf. This is part of my proposal
to place BFO under the PWWWB top level.
The primary reason why BFO cannot serve as a GP TLO is that it's
built around three assumptions that are inconsistent with modern
science and engineering: (1) a sharp dichotomy of continuants and
occurrents; (2) a theory of mereology that has infinitely thin
boundaries; (3) absence of any theory of systems with dynamic
connections that have much more structure than part/whole.
Furthermore, there is no evidence that any of the formal definitions
of the BFO hierarchy are relevant to any inferences in any of the
modules that are supposedly defined in terms of BFO.
In fact, I'll make a counter claim: WordNet with its vague English
definitions can support *every* application claimed for BFO at least
as well or better than BFO. A method for replacing BFO with WordNet:
1. Take the leaves (categories) at the bottom of the BFO hierarchy,
and throw away all the BFO nodes, terms, and structure above that
level.
2. Assign a WordNet synset to each bottom-level category of BFO
and replace the BFO definitions with the WordNet definitions.
3. Organize the resulting categories in a diagram that follows
the WordNet structure.
4. Prepare a design manual for ontology developers that is based
on the terminology of point #2 and the diagram of point #3.
I believe that the diagram of point #4 would be sufficient to
support all or nearly all the applications that have been claimed
for BFO. If there is any need in any such application to refer
to anything in the BFO hierarchy above the bottom level, then I
recommend point #5:
5. Find some word and synset in WordNet that corresponds to whatever
is required for that application and add its definition and its
position in the diagram to the design manual.
From what I have seen of BFO, I strongly suspect that a design
manual based on the above 5 points could be used to replace every
trace of BFO in every the application claimed for BFO.
However, some applications claimed for BFO depend upon semiotics.
Those applications would be significantly improved -- both more
general and more flexible -- if they were mapped to the semiotic
branch of PWWWB.
Furthermore, those applications that depend on dynamic systems and
structures, could be much improved by building an ontology based
on General Systems Theory (Bertalanffy & his followers) and using
their definitions instead of a part/whole theory based on mereology.
I can't claim this for certain, since I don't know enough about
the BFO applications. But from what I know of BFO, I haven't seen
anything that WordNet synsets couldn't specify as well or better.
Finally, to answer the question about how I would map the PWWWB TLO
to BFO, I would follow Matthew's option 3: Make BFO a microtheory
under PWWWB and organize a lattice of microtheories to support all
the additions that anyone has made for BFO applications.
In the debate on May 1, we also discussed other issues, which I'll
address in further email notes in the next few days.
John
bruces...@cox.net
Good morning from Santa Barbara.
It feels to me that this "great debate" is moving into an area of absolute ontological primacy -- maybe some kind of primal duality. Yes, it's about individual/personal models and professional histories and tendencies within disciplines -- but I would hazard the notion that all of this is taking place within a larger universal/ontological "container" that lays out the large-scale options in a very basic way. Right now, from my point of view, the issue IS "applied ontology" and not some elegant but disconnected logical model. But that might be because I am looking at larger more broadly defined issues -- not whether "the industry" is likely to overcome the messy and expensive consequences of "the silo effect" -- but whether the entire human community can overcome these tendencies -- these implications of their innate and seemingly unavoidable and instinctive methodology. Reality is empirical, right? Build your understanding bottom-up. Otherwise you are not a "scientist" -- you're some kind of mystic or religionist floating somewhere above reality and offering unreliable opinions based on metaphors and wishful thinking...
I've been spending a lot of time lately with groups passionately interested in "dialogue" -- and how society can overcome political polarization and start coming together to make good decisions about critical issues we must face in common. There are big networks of organizations working in this area, often with a lot of passion. People are trying to figure this out and come up with constructive governance models that can work across fundamental political divisions ("red state, blue state", etc.). It's a tough business and hard to move forward. It seems to involve a kind of religious conversion process for absolutely everybody. https://www.bridgealliance.us/ Come to a common meeting point and work together across all divides. But if we are driven by selfish or narrow motives, who wants to do that?
For me, taking the view from 50,000 feet, all of this, including the scruffy/neat debate, can be logically contained within a single and highly traditional philosophical/mathematical container: the fundamental ontological tension between "the many" and "the one". Maybe this can be described as the tension between bottom-up and top-down. This seems to be a primary debate that has been going on since the beginning of philosophic thought. I'd say the issue is well illustrated by the tension between Plato (mysticism, idealism, the one, the whole, top-down) and Aristotle (empirical science, undeniable facts, the many, the part, bottom-up) -- as represented in the famous "School of Athens" painting. https://goo.gl/aAA6D
How can these two primary schools of thought come together into a single universal container that coordinates the critical relationship between these two perspectives? For me, this is the fundamental issue for "applied universal ontology". In the context of a world under increasingly high stress, we (somebody) has to figure this out.
A couple of examples. Speaking of political dialogue, I like the article from an Australian emeritus professor: "Dialogical Citizenship: Dancing towards Solidarity" -- https://goo.gl/b88sud He centers his entire discussion around the many/one tension. This is probably the strongest intellectual statement I have seen on this general issue of political dialogue fully embracing the task of balancing these forces.
The tension between libertarianism (independent, local, personal freedom) and communitarianism (interdependent, global, personal responsibility) can be well described by this tension. So can the tension between nationalism and globalism.
The emergence of Brexit can be described in these terms. Are we "many" (local, strong borders, independent nation/state and economy) or are we "one" (global, minimal borders, part of EU, conjoined union)? It's a powerful and primal political oscillation. Maybe it's dangerous. I'd say it is.
And this morning, on CNN, there was a story featuring Ian Bremer and his new book "Us Versus Them: The Failure of Globalism" https://goo.gl/a7EtkJ Amazon lists this book as the number-one best seller in the topic of nationalism.
Excerpt from the top review:
The “we/they” division is global in scale and catastrophic in scope. It is already testing our civility, our security, our cultural identity, and our commitment to the ideals of democracy. But you already know that.
This is the latest in a growing list of books that seeks to understand why the we/they divide exists without, to its credit, falling into the trap of using the data to simply fan the fires of partisan division. Bremmer has a political agenda (we all do), and he’s no fan of Trump, the person. He does, however, go out of his way to note, “Donald Trump didn’t create us vs. them. Us vs. them created Donald Trump, and those who dismiss his supporters are damaging the United States.” Whether you agree with that or not, he is one of a handful of analysts willing to try and rise above the personal vilification that defines so much of our current political debate.
The author reviews the “we/they” division around the world and his analysis of current events in places like Nigeria and Venezuela is revealing and informative. I must admit, however, that for a time I found the analysis to be just a bit repetitive and a little superficial. There are lots of facts and figures but not a lot of insight into the why behind the what.
I do believe, however, that Bremmer essentially closes the “why” loop in the last section of the book when he takes up the obvious question of a way forward. In short he believes that we must do no less than redefine the social contract between the government and the governed.
And it is here that he once again opens his thinking in a way that few other authors have. All too often any discussion about the social contract devolves into a largely PC debate about freedom of the press, representative democracy, and the legal protection of marginalized people. We talk about authoritarianism and fascism, but what most citizens want, in the end, is a government that is fair, trustworthy, and competent, treats them with respect, and, most importantly, has their collective interests at heart.
And that, Bremmer points out, is a social contract we can find common ground on. We are never going to agree on every aspect of what a good government should or should not do. If we can agree on the framework of a social contract that acknowledges the inequities created by globalism, the challenges presented by the mass migration of people, the need for lifelong education in a technologically advancing world (without ignoring the continued importance of the traditional liberal arts), and the global desire for personal security, we can make a start.
The huge question the world is facing – the “Big We” that must deal with truly global issues like climate change or global economy – is how are we going to create a balance between these two natural and defensible but opposing tendencies in the body politick?
My case is – this is not essentially a political issue. It’s essentially an ontological issue that has existed since the time of Plato and Aristotle and will continue to trouble humanity until the fabled Don Quixote Prometheus Icarus task of clearly “solving” this issue comes into the world in sound scientific format with immediate practical applications.
A fool’s errand? A kamikaze mission? Maybe. But my instinct is – if evolution is going to succeed on this planet, and we are going to get through this vast systemic transition into globalization, the analytic thinkers and scientists and philosophical visionaries of this world must come together to resolve this tension and open an entirely unprecedented and renaissance-like level of co-creative intersector social power and constructive collaboration.
I appreciate this “great debate” on ontolog. It’s been a valuable part of my life for five years. Thank you.
Obrst, Leo J.
Aldo Gangemi, Nicola Guarino, Claudio Masolo, and Alessandro Oltramari. 2003. Sweetening WORDNET with DOLCE. AI Mag. 24, 3 (September 2003), 13-24.
Nicola Guarino
Alessandro Oltramari
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-forum+unsubscribe@googlegroups.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 email, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontology-summit" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontology-summit+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Research Scientist and Project Leader
2555 Smallman St., Suite #301, Pittsburgh, PA 15222
Tel. +1-412-689-1514
John F Sowa
Leo
> A. Gangemi, N. Guarino, C. Masolo, & A. Oltramari (2003)
> Sweetening WORDNET with DOLCE. AI Mag. 24:3. pp 13-24.
https://pdfs.semanticscholar.org/94f3/f0573cb8fc5dd1f7bdfc0964c4b3bc2e294e.pdf
Alessandro
> was that Christiane Fellbaum and George Miller decided to introduce
> "instance of" relation in the semantic network, because some of
> WordNet synsets, indeed, were instances.
But my recommendation doesn't involve any of those issues.
Recommendation: For the Gene Ontology and other BFO applications:
(1) take a small set of synsets from WordNet (none problematical);
(2) replace all BFO terms that are used in applications with the
names of the WN synsets that correspond to the BFO names; and
(3) ignore the BFO hierarchy and the jargon that explains it.
I claim that this process, which removes all traces of BFO, would
make no significant change in any applications claimed for BFO.
I would love to see any example to the contrary.
Nicola
> DOLCE: So Far, So Close… Cosmos + Taxis 4(4), 2017, pp. 10-18
is well taken. I would strengthen it further: that distinction
contradicts both classical Newtonian physics and all 20th & 21st c
physics. It cannot be part of any foundational ontology. That
implies BFO can never be more than a special-purpose microtheory.
Re your review of my KR book (2000): I admit that Chapter 2 is an
attempt to cover far too much material in far too short a space,
especially for an introduction. I also admit that the way I related
Peirce, Whitehead, and others required much more clarification.
For a more recent version, see the slides for a five-day short course,
which I presented in 2013: http://jfsowa.com/talks/patolog1.pdf
I took some slides from patolog4.pdf for the debate on May 1.
The major difference between the patolog lectures and my KR book is
that I now treat semiotics as a metalevel theory. That method is
simpler, clearer, and closer to what Peirce had intended. It's also
the basis for the way I relabeled the tree by Wilkins in slide 14 of
http://jfsowa.com/temp/ontostan.pdf .
Re foundational ontology: I maintain that any top level must be
philosophically neutral. That implies that any philosophical jargon
about which philosophers have not reached a consensus (ie, almost
everything) shall be relegated to microtheories, not the top level.
Please look at the PWWWD tree in slide 14 of ontostan.pdf. Note that
there are only two distinctions, type/token and substance/accident,
that could be called "philosophical jargon". David Armstrong (1989)
said there is "only one distinction that practically all contemporary
philosophers accept... It is the distinction between token and type
by Peirce." (see p. 2 of http://jfsowa.com/pubs/signs.pdf )
For the substance/accident distinction, I'd be happy to replace it
if and when I could find a more common pair of terms. If anyone
objects to the left bottom distinction quality/reaction/mediation,
I would replace it with monadic/dyadic/triadic sign relations.
Azamat
> John is again right. I agree with most of his points.
Thanks for the support. I'll return the compliment by endorsing
your book as an important collection of issues and aspects that
any general-purpose ontology must be able to represent and reason
about. But I consider the book as a set of requirements for any
formal ontology, rather than a formal ontology itself.
Bruce
> and not some elegant but disconnected logical model.
immense territory that any general-purpose ontology must address.
WordNet is another map that has been more successful for a wide
range of applications. By collecting and organizing all the words
in common use, it has become a map of everything that anybody
thinks, says, or does.
Bruce
I certainly believe that empirical methods (experiment & testing)
are important for science, engineering, and even politics. But
understanding requires a "big picture" that uses a top-down map
to put the bottom-up views in perspective.
Bruce
> the scruffy/neat debate, can be logically contained within a single
I agree, but only if that container is philosophically *neutral*.
No philosophical jargon about ongoing disputes among philosophers
may be used to specify or explain any foundational ontology. For
that issue, see Armstrong (1989). For citations and discussion,
see http://jfsowa.com/pubs/signs.pdf
John
bruces...@cox.net
John, I went back to your http://jfsowa.com/pubs/cs7.pdf Chapter 7 of Conceptual Structures, “Limits of Conceptualization”, and translated it into word.docx. This chapter in particular is so familiar to me, and was a kind of revelation to me when I read it in 1984. At the time, this was the best analytical/philosophical overview of semantics I had ever seen. It was clear. It made so many critical points. It emphasized this essential issue of intention and the many alternative particular ways "the same reality" can be parsed. My view was -- and is -- if you don't get this point, your semantics is blind. This is not a minor point -- it's the essence. This is why the noble effort of numeric taxonomy by Snokal and Sneath doesn't really work (https://goo.gl/8cscH4 ). I'd say it's why Wittgenstein was confused to the end. People think they can build rigid top-down linear models that work in every case. Hey gang, that's a mythology that's been disproven -- get over it. As John emphasized, every taxonomy or cascade of categories involves some "logically arbitrary but motivated" set of specific choices, and there is no one way to define these choices, and can't be -- for this reason. In the end, a stable taxonomy is an "industry (or social) agreement" -- not because it has solved the primary ontological problem of word meaning, but because human beings have to understand each other under real-world working conditions, so we'll settle for this empirical model, and it's close enough for us.
I did look at Nicola Guarino's review of your 2000 book on knowledge representation. I haven't read that book, but I understand it to be an extension of your foundational work, and the continuing effort to map the bottom-up methods of conceptual graphs into a top-level and integral/holistic framework. When you put terms like "Logos" or "Tao" or “God” at the top of the cascade, I am 100% with you. Your approach to this question is in many ways very similar to mine -- that started out with an analysis of concepts in deep psychology and "mystical geometry" -- concepts like the "wheels" of Ramon Lull. Taxonomies defined in layers by concentric rings -- that for me put the top level at the center (a point of convergence defined by absolute similarity – “what everything has in common”) and empirical particulars and parts at the periphery – then tried to perfectly map the connection between those levels as a 100% fluent linear cascade starting at the absolute level and guided in its flexible specifics by context and particular intention.
I can't cite particular reasons for agreeing with Nicola Guarino's critique -- but I think I understand it, and think I understand the problems he sees. Guarino writes:
My feeling is that this book is meant for adults only. I mean that although Sowa’s first book was suitable for a large audience (and I myself still recommend it to knowledge engineering newcomers), this one is less systematic and more problematic, reflecting years of passionate inquiry into the deepest foundations of conceptual analysis and knowledge representation. What emerges is a vivid picture of the author’s peculiar view of history of logic and philosophy, which is deeply intertwined with the analysis of the big problems of knowledge representation and knowledge engineering.
What I am saying here is just a guess – but my feeling about this issue is that your vision has involved this passionate quest to map the bottom-up methods of conceptual graphs, to which you were deeply committed, into the transcendental top levels of traditional ancient/modern mysticism. I’d say your philosophical instincts led you to the “God at the top” way to think – but your commitment to bottom-up empiricism and conceptual graphs as the starting point created some very difficult tangles for that mapping process.
Here’s my problem with the bottom up methods: they seem to involve this massive and inherently bottomless effort to fully catalogue every possible empirical instance, by creating some utterly (impossibly) vast library of particulars. Catching my breath – I think this is a fundamental methodological error – driven by the “industry commitment” to solid empirical models – essential for the industry, impossible for the philosophers and conceptual analysts. You found yourself trying to do both – link to the absolute top, yet map through specific and very bushy and probably ambiguous particular models. That task looks to me either impossible, or exceedingly confusing.
Here’s the error in a slightly snarky nutshell: “You guys have not generalized the principles of empiricism. You venerate it but you don’t understand it – and haven’t taken the time to do so. Your so-called “primitives” are far from primitive: in general, they are highly composite logical units with a deeply nested and important and significant but unspecified implicit decomposition. THAT is absolutely a formula for failure. Plugging those assumptions into the bottom of your logical cascade creates endless confusion, ambiguity in the foundational definitions, and obliterates any possibility for wide-spread agreement. It’s a lost cause and a dumb thing to do.” There ARE general principles by which bottom-level items can become absolutely well-categorized in a fluent linear cascade that extends across all levels to whatever we want to call the top. But “for some reason” we’re not doing that. Instead we are depending on overwhelming statistical power (this Quora explanation of how Google translation works is pretty interesting: https://goo.gl/foQUnv ).
I think those general principles involve measurement, and “scales” and levels and classes of variables – and from my point of view, what is needed is a generalized way to construct these logical objects out of unquestionably and indisputably rock-bottom and “atomic” conceptual units. For me, that atomic unit is the concept of distinction – which I define in terms of linear measurement as the smallest discernible (or significant) difference in some dimension – perhaps defined as a boundary value range like the google map definition of longitude and latitude that locates a point. It’s accurate – as is the case with all empiricism – to within a bounded range – to within “acceptable error tolerance”, to within some particular number of decimal points. Agreeing on an acceptable error tolerance is the foundation of precise agreement on anything.
Bottom-level items in any ontological/taxonomic cascade should be defined this way – without attempting what looks to me like the hysterical task of trying to explicitly list every possibility. Just create those categories ad hoc on the fly, according to the best systematic model of empiricism, in some local motivated context. Then you don’t have to enumerate them all – just list them when you need them. That’s how human beings do it.
Once building from that foundation – then the next task is to define the relationship between “quantity and quality” on that same basis. “Qualities” can be stipulatively defined in terms of numeric dimensionality – though in common human practice we don’t do that – because it is too complex and overwhelms the realities of practical psychological economy (abstractions are shorthand for complexities we don’t have the time or bandwidth to explain – but when we interpret them differently, we fight).
It’s this common inability to map this connection between quantity (“one dimensional”) and quality (‘multi-dimensional”, with an implicit decomposition) that creates the split in the structure of knowledge – from “the sciences” to “the humanities”. That split is the essence of “The Great Debate” – the inability to build a smooth and continuous map from any set of highly detailed empirical particulars to a single logical container. So we fight and fuss and fume like fools and the world limps along…
bruces...@cox.net
https://www.academia.edu/33548143/BFO_and_DOLCE_So_Far_So_Close.
I seem to do a lot better with "words" than I do with mathematical symbolism. The article goes over many themes I have considered important -- like precise description of qualitative attributes -- and describes them in terms that make a lot of intuitive sense to me. So I translated the complex web layout into a straight word.docx format, and if a careful reading of the article doesn't persuade me that what I am saying is nonsense, I might try to write a response from the point of view of "Synthetic Dimensionality" -- the concept that emerged for about that same time -- the early 1980's -- under the influence of many of the same books and themes.
My vintage 1980s bibliography: http://originresearch.com/sd/biblio.cfm
But I want to comment on John's response to me, which I appreciate.
Bruce
> and not some elegant but disconnected logical model.
> immense territory that any general-purpose ontology must address.
> WordNet is another map that has been more successful for a wide range
> of applications. By collecting and organizing all the words in common use,
> it has become a map of everything that anybody thinks, says, or does.
I am not too familiar with WordNet -- but I have always been a big fan of George Miller. His "Magical Number 7" tends to influence a lot of how I see the world -- including one reason why the world is so divided politically: the dimensionality and bandwidth of the issues we face together is totally overwhelming, so people gravitate towards comprehensible subsets -- like "parts of the elephant" -- and then fight at the border lines. We need a really strong method for "elephant construction". https://goo.gl/R7MGU
Regarding a "map of everything that anybody thinks, says, or does" -- my instinct is to look for a general purpose method that works in every case, and is driven by local specifics and context, rather than build a universal catalog. Insofar as possible, I do want "perfect elegance" -- because I think logical aesthetics is a reliable guide to good analysis -- but I don't want to create "logical jewelry" -- pretty and complex mathematical objects that nobody can do anything with. This has all got to be driven by down-to-earth engineering motivation. Fix real ("actual") problems. There's no shortage....
Bruce
John
> for science, engineering, and even politics. But understanding requires
> a "big picture" that uses a top-down map to put the bottom-up views in perspective.
Bruce
> scruffy/neat debate, can be logically contained within a single and
John
I spent a lot of time staring at the work of Douglas Hofstadter, and maybe some of this derives from his influence. I keep going back to something like a Moebius strip as a way to define this unbounded interval -- with its infinite extension in any one direction, and its two-sided properties like the Janus-faced holon. Is that thing a "primal distinction" if we understand it right? Does that twist somehow define a dimension that is "making a cut in itself?" Can every possible logic distinction cascade from it in some reliable and "pure" ("philosophically neutral") way? Is this the top level container of "absolutely everything" that you have been talking about?
We are cutting up the absolutely undefined continuum -- into "pieces" -- under the influence of immediate purpose. "Is it Animal, Plant, or Mineral?" I played that child's game with my sister, and I think Linnaeus started that way.
John
> may be used to specify or explain any foundational ontology.
> For that issue, see Armstrong (1989). For citations and discussion,
> see http://jfsowa.com/pubs/signs.pdf
All of this points to what seems like the Holy Grail of cognitive science -- and I was literally defining it in those terms as I followed this deeply holistic and perhaps inherently impossible (like an Escher drawing) vision as a 22 yr old college student, reading things like Ernst Cassirer's Phenomenology of Symbolic Forms and going from there to engineering books....
***
Hmm. Just a PS. In some other notes, I seem to be protesting "bottom up" in general -- but what I am really protesting is defining the bottom elements of the universal taxonomic cascade as broadly conceived holistic categories with implicit (and thereby ambiguous) decomposition. Actually, what I want to see is perfect (potentially continuous) logical fluidity in both directions -- up and down the hierarchy -- with the entire framework constructed from rock-bottom "atomic" information structures, so absolutely every discernible micro-iota of the cascade can be defined explicitly, with zero potential for ambiguity or potential alternative interpretations. Everything is intentional, and everything is specified explicitly to the micro-degree. Yes, I love metaphor -- but not under conditions of war...
Joao Paulo Almeida
From: Giancarlo Guizzardi <gguiz...@gmail.com>
Date: Sat, 5 May 2018 at 17:58
Subject: Fwd: [Ontology Summit] Re: [ontolog-forum] The Great Debate
To: Joao Paulo Andrade Almeida <jpandrad...@gmail.com>
---------- Forwarded message ----------
From: <gguiz...@inf.ufes.br>
Date: Saturday, May 5, 2018
Subject: [Ontology Summit] Re: [ontolog-forum] The Great Debate
To: ontolog...@googlegroups.com
Cc: ontolo...@googlegroups.com, "ar...@kyndi.com" <ar...@kyndi.com>, Benjamin Grosof <benjami...@gmail.com>, Andre LeClerc <andre....@kyndi.com>
Free University of Bozen-Bolzano, ItalyProf. Giancarlo GuizzardiGiancarloBest,it would be important to have this issue clarified.(which, unfortunately, I was not there to follow) but I thoughtI know this is a small detail in your whole discussionwould be mistaken.and (First-Order) UNIVERSAL would be INSTANCES of higher-order UNIVERSAL,as instances of a higher-order universal MODADIC PREDICATE.committing to having Universal and ParticularObviously, also in these cases, one is notParticular(x) or Universal(x) as predicates.instances of a (higher-order universal) UNIVERSAL.commitment to UNIVERSAL and PARTICULAR beingand PARTICULAR are represented by classes holds nothat (in these diagrams only) the categories of UNIVERSALto illustrate UFO's categories. However, from the fact@Barry: sometimes we use a class diagramUFO there is completely mistaken.reacting to at this point is pointing out that Barry's analysis ofbefore voicing a more general opinion. The only thing I cannot resistSince I missed most of the discussion so far, I have to catch up with itHowever, by chance, I ran into Barry's response to John here:I hope this was recorded so that the rest of us can benefit from it later.Dear all,I am very sorry that I missed what I am sure was a very interesting and lively discussion.
https://www.youtube.com/watch?v=TDJk9xTvdqY
This is analogous to when in the formalization of BFO/UFO/GFO, one usesNotice that we only use the class diagram for illustratingUFO's categories to make it easier to the UML/Conceptual Modeling/Information Systemscommunity to see the homomorphism to the OntoUML metamodel.(I guess your confusion was triggered by what peopled in the metamodelcommunity calls the confusion between ontological and linguistic instantiation -I hate the names but that's their jargon)In any case, these diagrams should not be taken as replacementsfor UFO's formalization.of a higher-order universal UNIVERSAL, in that case, PARTICULARFinally, notice that even if we were committing to the existence
not specializations of it. So, even if this were true, your analysis
==============================
https://www.inf.unibz.it/krdb/core/
your book as an important collection of issues and aspects that
any general-purpose ontology must be able to represent and reason
about. But I consider the book as a set of requirements for any
formal ontology, rather than a formal ontology itself.
BruceRight now, from my point of view, the issue IS "applied ontology" and not some elegant but disconnected logical model.
I agree, but note my reply to Azamat. His book is a map of the
immense territory that any general-purpose ontology must address.
WordNet is another map that has been more successful for a wide
range of applications. By collecting and organizing all the words
in common use, it has become a map of everything that anybody
thinks, says, or does.
BruceReality is empirical, right? Build your understanding bottom-up.
Otherwise you are not a "scientist" -- you're some kind of mystic...
I certainly believe that empirical methods (experiment & testing)
are important for science, engineering, and even politics. But
understanding requires a "big picture" that uses a top-down map
to put the bottom-up views in perspective.
BruceFor me, taking the view from 50,000 feet, all of this, including
the scruffy/neat debate, can be logically contained within a single
and highly traditional philosophical/mathematical container...
I agree, but only if that container is philosophically *neutral*.
No philosophical jargon about ongoing disputes among philosophers
may be used to specify or explain any foundational ontology. For
that issue, see Armstrong (1989). For citations and discussion,
see http://jfsowa.com/pubs/signs.pdf
John
--
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 email, see http://ontologforum.org/info/
--- You received this message because you are subscribed to the Google Groups "ontology-summit" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontology-summ...@googlegroups.com.
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 email, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontology-summit" group.
João Paulo