Views on when to use n-ary relations greater than binary - knowledge representation

150 views
Skip to first unread message

Robert Rovetto

unread,
Sep 10, 2021, 2:28:50 AM9/10/21
to ontolog-forum
All,

(without being restricted to any particular knowledge representation and reasoning language, because some are restricted only to binary, and some have design patters under that circumstance)

What's your view on:
- when to create a greater-than-binary relation rather than a binary relation?

Consider: you want to represent some information, statement, or knowledge, without necessarily being forced to limit to binary relations. A common example is when wanting to reference time. And 'between' is greater than binary.
What are other pieces of knowledge that you'd want assert a ternary, or greater than binary relation to capture it accurately?

Do you have any rules of thumb for knowing when to assert n-ary relations greater than binary?

Robert
--
Actively open to work and PhD study opportunities, worldwide.
-
IAOA Education Committee
Contact

Azamat Abdoullaev

unread,
Sep 10, 2021, 5:43:21 AM9/10/21
to ontolog-forum
Whenever you go for reality and nonlinearity, you need n-relations, as a POSET, chains, tree-like ordering, cycles and networks. A binary relation is just a poor approximation. It was discussed a decade ago. http://ontolog.cim3.net/forum/ontolog-forum/2011-11/msg00085.html

--
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/823f45d3-3a43-4415-9dbe-d69677fd4ad0n%40googlegroups.com.

Igor Toujilov

unread,
Sep 10, 2021, 12:38:36 PM9/10/21
to ontolo...@googlegroups.com

Hi Robert,

Depending on practical use cases people use binary or n-ary relations. They even use unary relations almost in all cases. Some people do not realise they use unary relations: for example, asserting an instance of a class.

The rule of thumb for using n-ary relations where n>2 is pretty trivial: use it if you cannot use just unary and/or binary ones. However, I think you know that any n-ary relation where n>2 can be represented as a combination of unary and/or binary relations.


Cheers,

Igor



--

Michael DeBellis

unread,
Sep 10, 2021, 2:48:01 PM9/10/21
to ontolog-forum
I agree with Igor. I use n-ary relations all the time just modeled as binary relations with additional classes.  One thing I don't understand from some of the arguments for things like RDF* is the idea that n-ary relations are a deficiency in standard RDF and OWL when the semantics are the same whether you implement them as n-ary properties or binary properties with additional classes. This isn't new with OWL either, it is one of the fundamental patterns of  OOP. 

Michael

Michael DeBellis

unread,
Sep 10, 2021, 4:42:38 PM9/10/21
to ontolog-forum
There is a good recent overview paper on Knowledge Graphs: https://arxiv.org/pdf/2003.02320.pdf  This has a good discussion of property graphs (direct implementation of n-ary relations) and RDF/OWL (implementing n-ary relations as binary) in section 2.1

Alex Shkotin

unread,
Sep 11, 2021, 3:59:42 AM9/11/21
to ontolog-forum
Robert,

One way to get a lot of n-ary relations is to look at any relational DB - there is a lot there - any table is an n-ary relation. If we map a row to a sentence we get a proposition. And visa versa any propositional sentence gives n-ary relation as in Executable English, as far as I remember:-)

Alex

пт, 10 сент. 2021 г. в 09:28, 'Robert Rovetto' via ontolog-forum <ontolo...@googlegroups.com>:
--

sowa @bestweb.net

unread,
Sep 11, 2021, 4:37:36 PM9/11/21
to ontolog-forum
Michael,
 
That 136-page article can be boiled down to one sentence of useful information:
 
"If you map your pet notation to and from Common Logic, the tools for the DOL standard will map it to any other format you might need."
 
See the following slides:  http://jfsowa.com/talks/eswc.pdf
 
These slides are an extended version of the slides that got the Best Presentation award at the 2020 KGC (Knowledge Graph Conference hosted by Columbia U.).  That was in March 2020.  In June 2020, I presented an extended version in a keynote talk at the European Semantic Web Conference.  In August 2021, I included a subset of them in my Ontolog Forum talk on Universal Query Language.  Since then, I added a new Section 6:  http://jfsowa.com/talks/uql.pdf
 
With the DOL standard, each group that has any special notation just provides a mapping to and from a subset of Common Logic.  Then anybody who prefers any other notation just uses the DOL tools to share data with you.
 
You can recycle those 136 useless pages.
 
John
 
 
 

From: "Michael DeBellis" <mdebe...@gmail.com>
Sent: Friday, September 10, 2021 4:43 PM

Sheth, Amit

unread,
Sep 11, 2021, 7:08:24 PM9/11/21
to ontolo...@googlegroups.com
I thought I would share the state of the Semantic Web tech exactly 20 years ago. On 09/11/2001 I gave this keynote in Erfurt, Germany
https://lnkd.in/dNzkTY35 (I was running Taalee/Voquette/Semagix founded in 1999 then).
Starting slide 35, I described our commercial #semanticsearch engine that used a large, 25 domain #knowledgegraph (slides 22, 53; also called WorldModel/ontology in the patent) which also supported the concept of KG enhanced Semantic Enrichment we called Rich Media Object (slides 46-49) - the concept was seen again in 2013 as Infobox in Google Semantic Search. Details at: https://lnkd.in/eKX2maV

ps: The rest of the day was unexpectedly eventful. I decided to leave early instead of staying for the conference, took the fast train to reach Frankfurt airport arriving 15-20 minutes before the departure of my flight, bypassed the counter - because I was a platinum premier, they opened the door that had just been closed. Wish I had missed the flight. Just west of Ireland's west coast over the Atlantic, the plane dumped fuel and landed at Dublin along with many other flights. You can image the rest.

doug foxvog

unread,
Sep 13, 2021, 1:59:08 AM9/13/21
to ontolo...@googlegroups.com
Sure, one can always make an n-ary relation more complex by reifying the
situation of the n-ary relation being true as an object and then relating
each argument place to that object with a specific binary relation. This
can be valuable if one wants to provide more information about that
situation -- which would be otherwise considered meta-data about the n-ary
relation assertion.

Needless to say, adding this complexity takes more memory and more
processing time.

Of course, one could do the same with existing binary relations -- reify
an assertion using them as an object and using 2 different binary
relations to relate the first and second argument to the reified object.
The utility of both seem the same to me.

E.g., (motherOf Juan Maria) => Exists (X; (AND (motherhoodSituation X)
(motherInSituation X Maria) (childInSituation X Juan)). If X is reified,
then additional information such as starting time, who knows X, who
reported X, the record # for X in a database Y, etc.

Yes, the above complicates expressing (motherOf Juan Maria), just as
breaking an n-ary (n > 2) relation into binary relations does.

I note that most databases use more than two columns. and that the
multiple columns often do not express binary relations. A shopping
database might have columns for store, date, product type, number of items
of type sold at full price, number of items of type sold at discount,
discount percentage, number of items returned, department of store, and
other columns. It is easy to express this as multiple n-ary relations
which a computer program can process to generate statistics and make
logistical decisions over hundreds of thousands of lines in the DB.
Reifying multiple objects for each line (millions total) to do the
processing would be an added complexity. -- especially since such objects
would have no utility after the desired calculations are made.

The only advantage in expressing n-ary relation as a set of binary
relations seems to me that it allows one to use a deficient language to
express them. I try to avoid such for the same reasons that i do not code
using actual Turing Machines. They are the wrong tool for the job, and as
such take more complexity and memory than using a tool that better fits
the problem.

-- doug foxvog

> On Friday, September 10, 2021 at 11:48:01 AM UTC-7 Michael DeBellis
> wrote:
> I agree with Igor. I use n-ary relations all the time just modeled as
> binary relations with additional classes. One thing I don't understand
> from some of the arguments for things like RDF* is the idea that n-ary
> relations are a deficiency in standard RDF and OWL when the semantics are
> the same whether you implement them as n-ary properties or binary
> properties with additional classes. This isn't new with OWL either, it is
> one of the fundamental patterns of OOP.
> Michael
>
> On Friday, September 10, 2021 at 9:38:36 AM UTC-7 Igor Toujilov wrote:
>
> Hi Robert,
>
> Depending on practical use cases people use binary or n-ary relations.
> They even use unary relations almost in all cases. Some people do not
> realise they use unary relations: for example, asserting an instance of a
> class.
>
> The rule of thumb for using n-ary relations where n>2 is pretty
> trivial: use it if you cannot use just unary and/or binary ones.
> However, I think you know that any n-ary relation where n>2 can
> be
> represented as a combination of unary and/or binary relations.
>
>
>
> Cheers,
>
> Igor
>
>
>
> --
> 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/88b90be93cf0438688106de6923aacfe%40bestweb.net.
>


Alex Shkotin

unread,
Sep 13, 2021, 5:06:06 AM9/13/21
to ontolog-forum
Robert,

Let me underline an important point: first of all, we have found in nature and society one or another relation and ask how many members each example of this relation can have? i.e. arity is a feature of relation itself. 
so,
"when to create a greater-than-binary relation rather than a binary relation?" - when relation itself is n-ary.
"Do you have any rules of thumb for knowing when to assert n-ary relations greater than binary?" - we come here to the logic of relations and its discovery. For me, examples of relations of different arity from one or another domain would be great.

What about [1]?

Alex



пт, 10 сент. 2021 г. в 09:28, 'Robert Rovetto' via ontolog-forum <ontolo...@googlegroups.com>:
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.

Jon Awbrey

unread,
Sep 13, 2021, 5:24:25 AM9/13/21
to ontolo...@googlegroups.com
Well, at least now I know why the Web is making people stupider day by day …

Regards,

Alex Shkotin

unread,
Sep 13, 2021, 7:25:38 AM9/13/21
to ontolog-forum
:-)

пн, 13 сент. 2021 г. в 12:24, Jon Awbrey <jaw...@att.net>:

Jon Awbrey

unread,
Sep 13, 2021, 8:48:46 AM9/13/21
to Ontolog Forum
Re: https://www.w3.org/TR/swbp-n-aryRelations/

I know, I tell myself, calm down, Jon, it's just a tutorial doc from 15 years ago ...
but every time I see someone using “reification” that way it just makes we want to
go back on Facebook and argue with Trumplodytes or anything else less dispiriting ...

At any rate, for what it's worth, as we used to say ...

Here's a first introduction to k-adic or k-ary relations from a mathematical perspective.

Relation Theory
https://oeis.org/wiki/Relation_theory

Here's a few additional resources and assorted discussions with folks around the web.

Survey of Relation Theory
https://inquiryintoinquiry.com/2020/05/15/survey-of-relation-theory-4/

More than anything else it's critical to understand the differences among:

1. The relation itself, which is a mathematical object,
a subset embedded in a cartesian product of several
sets called the “domains” of the relation.

2. The individual k-tuple, sometimes called an elementary relation,
a single element of the relation and hence the cartesian product.

3. The syntactic forms, lexical or graphical or whatever,
used to describe elements and subsets of the relation.

4. The real phenomena and real situations, empirical or hypothetical,
we use mathematical objects such as numbers, sets, functions, groups,
algebras, manifolds, relations, etc. to model, at least approximately
and well enough to cope with their realities in practice.

Regards,

Jon
>>> *Actively open to work and PhD study opportunities, worldwide.*
>>>
>>> *-*
>>> OntologPage <https://ontologforum.org/index.php/RobertRovetto>
>>> IAOA Education Committee
>>> Contact <https://ontospace.wordpress.com/contact/>
>>>

William Frank

unread,
Sep 13, 2021, 10:24:43 AM9/13/21
to ontolog-forum
Alex,

Well said!  

I was going to reply to the question and some of the answers with my semi-yearly rant about the reductionist unexamined belief that if something *can* be expressed using a smaller syntax, then it *really* is the translated expression, rather than what was said.   In my case, all the logic students who seem to have "learned" that there 'really is' only one truth functional connector: 'nor', when we find almost a dozen in human reasoning, and when there is no given reason to prefer one minimal set of these rather than another as more fundamental than any other.    

I might add to your advice:  'if you have n-ary relations in your language, use them for n-ary relations in the langauge your are modelling!   If not, then translate them if you can to what you *do* have in some standard way. 

Wm   

sowa @bestweb.net

unread,
Sep 14, 2021, 12:01:28 AM9/14/21
to ontolo...@googlegroups.com
Doug,
 
I agree with that point:
 
 
DF:  Sure, one can always make an n-ary relation more complex by reifying the
situation of the n-ary relation being true as an object and then relating
each argument place to that object with a specific binary relation. This
can be valuable if one wants to provide more information about that
situation -- which would be otherwise considered meta-data about the n-ary
relation assertion.
 
For example, the verb 'give' has three obligatory arguments:  Agent, Theme, and Recipient.  In linguistics, the links to those arguments may be represented by binary "case relations" or "thematic roles".  Then the syntax for any natural language can be related to those relations.
 
But this is a principled reason for introducing those relations.  In general, people should study the subject matter to determine when, whether, why, and how to add such features.
 
John

Alex Shkotin

unread,
Sep 14, 2021, 6:02:39 AM9/14/21
to ontolog-forum
William,

Math is a world of metamorphosis, polymorphisms, and so on. Just amazing :-)

Alex

пн, 13 сент. 2021 г. в 17:24, William Frank <william...@gmail.com>:

Alex Shkotin

unread,
Sep 14, 2021, 6:30:47 AM9/14/21
to ontolog-forum
Jon,

IMHO we should begin from your p.4 calling the math we need. So let's take case 3 from the tutorial you refer to:"
  1. John buys a "Lenny the Lion" book from books.example.com for $15 as a birthday gift. There is a relation, in which individual John, entity books.example.com and the book Lenny_the_Lion participate. This relation has other components as well such as the purpose (birthday_gift) and the amount ($15).
"
We have here the action done by John with the purpose of action in the future - as a birthday gift. As they pointed out we have individual, entity, book, $15, and birthday in the future. Are they related? Yes! How? Well, goto 1. But there is no relation they mentioned here at all unless they invent it consciously. This is from the RDB paradigm I think:-)

Alex

пн, 13 сент. 2021 г. в 15:48, Jon Awbrey <jaw...@att.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
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.

Giancarlo Guizzardi

unread,
Sep 14, 2021, 7:13:03 AM9/14/21
to ontolo...@googlegroups.com
Dear colleagues,

It has been hard to keep track of all details of the discussions.
In any case, perhaps the following papers can be of use somehow to a discussion about ontological aspects of relations:


The point here is not about arity, but differentiating categories of relations based
on the nature of their truthmakers.

best regards,
Giancarlo Guizzardi



--
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.

Aldo Gangemi

unread,
Sep 14, 2021, 7:35:41 AM9/14/21
to ontolo...@googlegroups.com, Aldo Gangemi
… and that’s a chapter from the "Ontology Engineering with Ontology Design Patterns” book, about alternative representations of multi-varied predicates (n-ary relations), with pros and cons. The chapter addresses both extensional and intensional relations:


Best
Aldo




________________________________

Aldo Gangemi

Full professor
University of Bologna

Director
Institute for Cognitive Sciences and Technology,
National Research Council (ISTC-CNR)




dr.matt...@gmail.com

unread,
Sep 14, 2021, 8:45:45 AM9/14/21
to ontolo...@googlegroups.com

OK. I can’t resist any more. Lots of interesting stuff here, but the basic problem is that most of it conflates relationships/situations/activities with relations. They are not the same thing. Relationships (for example) are what one thing has to do with another (or itself) and the other is a mathematical structure that is often found useful in representing the first (and there are choices about how to do that).

Regards

Matthew West

Giancarlo Guizzardi

unread,
Sep 14, 2021, 9:31:41 AM9/14/21
to ontolo...@googlegroups.com
Matthew,

I agree that relationship is "what one thing has to do with another" but we deny that this thing is (necessarily) a mathematical tuple.
In fact, for the most interesting material relationships, these "things" are concrete full-fledged endurants (e.g., marriages, enrollments, employments, presidential mandates).
The key to the analysis is understanding the nature of this mediating "thing" as the grounding of relations (relations hold, relationships exist)
and truthmakers of relational propositions.

Two other papers that might be useful in that respect are:

In the background of all this, and with a more philosophical guise
is our theory of "weak truthmaking": http://www.inf.ufes.br/~gguizzardi/truthmaking2019.pdf

best,
Giancarlo

PS: Thanks for the very nice talk at OntoCom yesterday

Chris Partridge

unread,
Sep 14, 2021, 1:02:33 PM9/14/21
to ontolo...@googlegroups.com
Hi Giancarlo,

The interesting question is whether these things "(e.g., marriages, enrollments, employments, presidential mandates)" are event objects in their own right - or somehow dependent relations.
Aldo's paper makes the point - noting of course, Davidson (D. Davidson. The Logical Form of Action Sentences. In The Logic of Decision and Action. University of Pittsburgh Press, Pittsburgh, 2nd edition, 1967.)
And also, the issues of multigradeness if treating them as relations - A. Oliver and T. Smiley Multigrade Predicates Mind, 113, 2004.

Chris

Giancarlo Guizzardi

unread,
Sep 14, 2021, 1:23:28 PM9/14/21
to ontolo...@googlegroups.com
Hi Cris,

That is an interesting question but not the only one. As you will see in the paper, our view is that they are not events (we actually explicitly reject) that. As you know, we tridimensionalists and we believe these things can genuinely chance “as a whole” like any other endurant.

Best,
G

To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-forum+unsubscribe@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 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-forum+unsubscribe@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 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-forum+unsubscribe@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 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-forum+unsubscribe@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 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-forum+unsubscribe@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 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-forum+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAMWD8MpChCpx3sUuyvbNOLJ5SboDRrbEMgVpd80uxODSMi%3DHjQ%40mail.gmail.com.

William Frank

unread,
Sep 14, 2021, 2:54:36 PM9/14/21
to ontolog-forum
My views about this may be naive, perhaps because of my predilection to consider thought as expressed in natural languages to be the subject under study, while relations, endurants, occurents, are models of the various ways people have and can experience and think about the same things (in themselves) as phenomena, depending on the semantic conventions in their languages. 

It seems to me this debate is mostly a matter of ways of thinking about things in a given language, rather than what they really are.  For instance, some languages separate masses of a substance, (gold, rice, hair in English) from individuals (beans, grains of rice, hair in French and German).   Hard for me to find cases where something can't be cast as being in one such category or another.  

Aren't there actually two things, governed by laws of duality

Endurants: marriages, enrollments, presidential proclamations, chickens
Occurents: the act of getting married, of enrolling, of proclaiming, the act of hatching, dying, 

So that the marriage endures from the time it occurs to the time of its dissolution, the duality between the two.  Etc. 
And of course since every event takes a length of time, and is not atomic except relative to a given viewpoint, and because some things one might think of most simply as events are treated as endurants - e.g., hurricanes, (after all, we give them names).

Though there are multiple roles in endurants and occurrents, does not prove they are 'really' relations.  It just shows they can be (perhaps only partially) modelled as relations.   What is missing, to me, about the model of an event or an enduant as a relation are the roles of the other things in the conceptual structure.  

 


To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@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 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.

--
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.

--
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.

--
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.

--
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.

--
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/CAGG%2B%3DmXkiTS8n7PdKFHDn4o5Es5FqjrvBZjKq94FKM4iCKGkDA%40mail.gmail.com.

Giancarlo Guizzardi

unread,
Sep 14, 2021, 3:06:13 PM9/14/21
to ontolo...@googlegroups.com
William,

My position (and I sure Chris would agree) is that these are ontological distinctions.

Language is polysemic and often polysemic in a systematic way. So, the term
Marriage can refer both to an endurant and an event, and these two referents would
bear some systematic relation between them.

What makes a marriage (qua-endurant) relational is the fact that it is
existentially dependent on a multitude of individuals. For example, the specific marriage
of John and Mary is existentially dependent on John and Mary, thus, binding them.
That is why it can be the truthmaker of the relational proposition "John and Mary are married".
If we look inside such an entity, this Marriage (again, qua-endurant) is a bundle of relational
qualities (tropes, modes) that are relational in nature (e.g., John's commitments and claims towards Mary,
and Mary's commitments and claims towards John, etc).

Now, what makes something like a marriage a true endurant is that it has essence and accidents
(like any other endurant). John&Mary's marriage is essentially a marriage but only contingently
a happily marriage, a marriage with full separation of assets, etc.

This is the gist of the idea

best,
Giancarlo


William Frank

unread,
Sep 14, 2021, 4:39:38 PM9/14/21
to ontolog-forum
Thanks, Giancarlo 

Your simple responses let me understand and see the wisdom of what you guys are saying.

I found this very helpful: 

Now, what makes something like a marriage a true endurant is that it has essence and accidents
(like any other endurant). John&Mary's marriage is essentially a marriage but only contingently
a happily marriage, a marriage with full separation of assets, etc.Thanks, Giancarlo

And this is part of what I was trying to say: 

Language is polysemic and often polysemic in a systematic way. So, the term
Marriage can refer both to an endurant and an event, and these two referents would
bear some systematic relation between them.

And this seems to me to reconcile and clarify what I was claiming about roles being more that 'just' what shows up in an n-tuple.

If we look inside such an entity, this Marriage (again, qua-endurant) is a bundle of relational
qualities (tropes, modes) that are relational in nature (e.g., John's commitments and claims towards Mary,
and Mary's commitments and claims towards John, etc).

I am still at a loss to know what this means: 

My position (and I sure Chris would agree) is that these are ontological distinctions.
Unless an ontological distinction is between different ways that people can see the same experience (for example, as a substance (rice and English hair) or as an individual (a bean, or French hairs). 

I would not be surprised, I must say, if you clarified this, too. 


On Tue, Sep 14, 2021 at 3:06 PM Giancarlo Guizzardi <gguiz...@gmail.com> wrote:
William,

My position (and I sure Chris would agree) is that these are ontological distinctions.

Language is polysemic and often polysemic in a systematic way. So, the term
Marriage can refer both to an endurant and an event, and these two referents would
bear some systematic relation between them.

What makes a marriage (qua-endurant) relational is the fact that it is
existentially dependent on a multitude of individuals. For example, the specific marriage
of John and Mary is existentially dependent on John and Mary, thus, binding them.
That is why it can be the truthmaker of the relational proposition "John and Mary are married".
If we look inside such an entity, this Marriage (again, qua-endurant) is a bundle of relational
qualities (tropes, modes) that are relational in nature (e.g., John's commitments and claims towards Mary,
and Mary's commitments and claims towards John, etc).


Roberto Rovetto

unread,
Sep 14, 2021, 6:02:16 PM9/14/21
to ontolog-forum
Returning to my original post/questions, please see that (there seem to have been tangents). To supplement and perhaps refocus:

Sometimes rich or complex expressions, sentences, beliefs or pieces of knowledge seem to express what in KRR and associated formal and computational activities would formalize in a greater-than-binary n-ary relation. However, in computational ontology some would model them--as others have already pointed out--using binary relations or patterns. Sometimes this decision is unfortunately forced upon them because of the expressive limitations of the KRR language or formalism (e.g., in OWL).

So I asked what are views when faced with expressions, sentences et al, that seem to express or need creating n-ary relations?
When would you create a greater-than-binary n-ary relation in a computational ontology rather than binary?

In other words...

  • Think of some greater-than-binary relationships. When would you formalize them--create logical or computable relations--as such rather than some other approach (binary or otherwise)? And why?

  • Think of a statement, fact, situation or piece of knowledge that you want to model or represent either in a KRR/ontological manner (i.e. computational focus) OR more abstractly in a purely symbolic logic.
    How would you, personally, determine whether n-ary is needed?
    (An obvious answer, already provided by others, is if there are more than 2 individuals or argument involved. Perhaps it's as simple as that, but i'd like to hear if there is a variety of views and approaches)

  • How would you, personally, decide if to assert binary or greater relations?
    Would it be based on convenience? Or trying to match as accurately or truthfully as possible what is expressed in the sentence or piece of knowledge? etc...?
  • If what you're trying to model seems to involve need of asserting an n-ary relation, or if it explicitly involves phrases that call for the assertion of an n-ary relation, what conditions would incline you to formalize them as such rather than another way?

Robert

P.S. William, this is obviously not the intended topic of my post, but I'll address some of what you asked and said. You are correct in various ways so do not be quick to discount or change your views. For example you wrote: "It seems to me this debate is mostly a matter of ways of thinking about things in a given language, rather than what they really are". Yes--there is much of that. The interplay between our ideas, thoughts, language, and inquiry to understand the world, etc. is complex with many unanswered questions and the literature from philosophy (ancient to contemporary), linguistics, KRR and AI (including ontology engineering), social sciences, etc. shows that. As others on the forum have expressed in one way or another over the years.

The topic and questions associated with endurants/perdurants (and the like) are highly abstract metaphysical/philosophical-ontology concepts and accounts of some phenomena (whether mind-dependent or otherwise)...certainly that is the historical sense that continues in the philosophical discipline to date. That very high abstractness--just like in natural langauge with generic words--means they are often open to interpretation, and certainly have a plethora of metaphysical accounts, and they reflect unanswered (perhaps unanswerable) philosophical question (e.g., for endurants/perdurants it is about time and persistence). Indeed, there are no answers (at least not universally accepted) to the associated questions from which these concepts and accounts are spawned. Some, assume a threedimensional account, others fourdimensional, and others make no explcit assumptions, or different ones. Likewise in this community of a computational focus, some (but certainly not all) ontologies or their developers assume things or make variations.

When these metaphysical concepts are injected--in some form--in this community of a computational focus, there does not need to be (nor should there be as it often muddies things) any metaphysical claims about the world or whatever target subjects being modeled, but rather the focus on computational goals, and things like internal consistency, etc. But moreover, they need not be injected/used in contemporary ontology engineering in any case. And, also given the great generality, new and hybrid concepts and accounts can be made.

Again, there's many such metaphysical accounts, and some computational ontologies (sometimes the most generic ones), as Giancarlo pointed out, use some concepts from these accounts or modify them or assert new concepts for their ontologies. But in the present context (e.g., this forum), putting forth metaphysical accounts is not the goal; metaphysicians and philosophical ontologists do that full-time as it were (and therefore with much greater thoroughness without the expressive limitations of computational formalisms, requirements, etc.). There is a computational focus, so internal consistency, rather than metaphysical claims is primary.

E.g., A marriage, and other things, can be modeled in different ways, and different metaphysical accounts will do so. Likewise, distinct computational ontologies will model them differently. And more generally, distinct persons or groups can model it uniquely.

On Friday, September 10, 2021 at 2:28:50 AM UTC-4 Roberto Rovetto wrote:
All,

(without being restricted to any particular knowledge representation and reasoning language, because some are restricted only to binary, and some have design patters under that circumstance)

What's your view on:
- when to create a greater-than-binary relation rather than a binary relation?

Consider: you want to represent some information, statement, or knowledge, without necessarily being forced to limit to binary relations. A common example is when wanting to reference time. And 'between' is greater than binary.
What are other pieces of knowledge that you'd want assert a ternary, or greater than binary relation to capture it accurately?

Do you have any rules of thumb for knowing when to assert n-ary relations greater than binary?

Robert
--

dr.matt...@gmail.com

unread,
Sep 15, 2021, 2:47:01 AM9/15/21
to ontolo...@googlegroups.com

Dear Giancarlo,

Yes, I also looked at this some time ago:

West, Matthew Information Modelling: An analysis of the uses and meanings of associations PDT Europe 2002, PDF

And in my book

West, Matthew Developing High Quality Data Models Morgan Kaufmann 2011

Where Chapter 11 has a section on associations (aka situations).

Regards

Matthew West

Giancarlo Guizzardi

unread,
Sep 15, 2021, 4:18:10 AM9/15/21
to ontolo...@googlegroups.com
Hi William,

What I meant to say with that passage is the following.

1) yes, there are multiple ontological cuts that can be made to reality. For example, a fourdimensionalist such as Chris would
believe that not only John&Mary's marriage but also John and Mary are events. Someone holding my views, in contrast, will
see John, Mary, and John&Mary's Marriage as endurants to which there are events associated that will constitute John's Life, Mary's Life,
John&Mary's Marriage's Life (actually, to each of these endurants, there are associated multiple possible lives but that is a different story -
see ).

2) What I mentioned that both Chris and I would hold is the view that things like John, Mary and John&Mary's Marriage are things
that really exist in the world. So, the piece of language "John&Mary's Marriage" picks up something really out there. Again, for Chris,
this would be an event; for me, this can polysemically refer to both the endurant binding them and one of its possible lives (an event).

3) Here is what I think you mean: we can have different conceptualizations
to the same portion of reality. These conceptualizations would influence perspective, granularity, etc. Given this, we can see a phenomenon
as a portion of blood (an amount of matter) or a collective of individual blood cells. To that I would agree.

Hope I managed to clarify my point

best,
Giancarlo


Giancarlo Guizzardi

unread,
Sep 15, 2021, 4:19:01 AM9/15/21
to ontolo...@googlegroups.com
Thanks for the pointers, Matthew.
I will definitely look into them

best,
Giancarlo

Giancarlo Guizzardi

unread,
Sep 15, 2021, 4:21:44 AM9/15/21
to ontolo...@googlegroups.com
Sorry, forgot to past a link in this passage:



On Wed, Sep 15, 2021 at 10:17 AM Giancarlo Guizzardi <gguiz...@gmail.com> wrote:

1) yes, there are multiple ontological cuts that can be made to reality. For example, a fourdimensionalist such as Chris would
believe that not only John&Mary's marriage but also John and Mary are events. Someone holding my views, in contrast, will
see John, Mary, and John&Mary's Marriage as endurants to which there are events associated that will constitute John's Life, Mary's Life,
John&Mary's Marriage's Life (actually, to each of these endurants, there are associated multiple possible lives but that is a different story -
see ).


I mean that the interested reader in what I meant by an "endurant being associated with multiple possible lives"
could find useful to take a look at http://www.inf.ufes.br/~gguizzardi/BPM2016.pdf

best,
G

William Frank

unread,
Sep 15, 2021, 9:42:25 AM9/15/21
to ontolog-forum

Jon Awbrey

unread,
Sep 16, 2021, 11:00:28 AM9/16/21
to Ontolog Forum
Cf: Relations & Their Relatives • Discussion 22
https://inquiryintoinquiry.com/2021/09/16/relations-their-relatives-discussion-22/

Re: Ontolog Forum
https://groups.google.com/g/ontolog-forum/c/0pDK8IJiFDc
::: Roberto Rovetto
https://groups.google.com/g/ontolog-forum/c/0pDK8IJiFDc/m/jwg3FbzfBgAJ
https://groups.google.com/g/ontolog-forum/c/0pDK8IJiFDc/m/HrdY9T1QAAAJ

<QUOTE RR:>
What's your view on:

When to create a greater-than-binary relation rather than a binary relation?

Consider: You want to represent some information, statement, or knowledge,
without necessarily being forced to limit to binary relations. A common
example is when wanting to reference time. And “between” is greater than
binary. What are other pieces of knowledge that you'd want assert a ternary,
or greater than binary relation to capture it accurately?

Do you have any rules of thumb for knowing when
to assert n-ary relations greater than binary?
</QUOTE>

Dear Roberto,

Let me return to your original question and give it better attention.

You have probably noticed you got a wide variety of answers coming from
a diversity of conceptual frameworks and philosophical paradigms. It
gradually dawned on me some years ago these differences are most likely
matters of taste about which all dispute is futile, however much we go
ahead and do it anyway. So I'll just say what I've found works best in
my particular applications of interest, namely, applying relational logic
to mathematics and research sciences.

To avoid the kinds of culture clashes I remember from the
Standard Upper Ontology Lists and other ancestors of this
Forum at the turn of the millennium, I'll develop the rest
of this line of inquiry on the Relations & Their Relatives
thread ( https://groups.google.com/g/ontolog-forum/c/cL22KqWr8PI )
and, as I usually do, post better-formatted copy on my blog
( https://inquiryintoinquiry.com/ ).

Regards,

Jon

Michael DeBellis

unread,
Sep 16, 2021, 11:22:00 AM9/16/21
to ontolo...@googlegroups.com
You have probably noticed you got a wide variety of answers coming from
a diversity of conceptual frameworks and philosophical paradigms.  It
gradually dawned on me some years ago these differences are most likely
matters of taste about which all dispute is futile, however much we go
ahead and do it anyway. 

Exactly. One thing I learned the hard way is that there is no perfect model,  you can always model a domain in multiple ways and thrashing around looking for the model that is the "real" model leads to analysis paralysis.  Every methodology that is any good always emphasizes this. Also, I think most of these discussions are pointless in another more important way: when you do actual software development they almost never come up and if they do a good team leader shuts them down really quickly.  You will be constrained on any real project to use some sort of modeling paradigm: UML, OWL, etc. That paradigm (and methodologies associated with it) will typically define what kind of formalisms you have to work with and when to use what.  

Michael

--
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 a topic in the Google Groups "ontolog-forum" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/ontolog-forum/0pDK8IJiFDc/unsubscribe.
To unsubscribe from this group and all its topics, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/0bbbcf11-6b01-96ea-d855-bcbbbd76c3d5%40att.net.

doug foxvog

unread,
Sep 16, 2021, 11:00:09 PM9/16/21
to ontolo...@googlegroups.com
On Tue, September 14, 2021 18:02, 'Roberto Rovetto' via ontolog-forum wrote:
> Returning to my original post/questions, please see that (there seem to
> have been tangents). To supplement and perhaps refocus:
>
> Sometimes rich or complex expressions, sentences, beliefs or pieces of
> knowledge seem to express what in KRR and associated formal and
> computational activities would formalize in a greater-than-binary n-ary
> relation. However, in computational ontology some would model them--as
> others have already pointed out--using binary relations or patterns.
> Sometimes this decision is unfortunately forced upon them because of the
> expressive limitations of the KRR language or formalism (e.g., in OWL).
>
> So I asked what are views when faced with expressions, sentences et al,
> that seem to express or need creating n-ary relations?
> When would you create a greater-than-binary n-ary relation in a
> computational ontology rather than binary?
>
> In other words...
>
>
> - Think of some greater-than-binary relationships. When would you
> formalize them--create logical or computable relations--as such rather
> than
> some other approach (binary or otherwise)? And why?

I would use an n-ary relation when the knowledge i wish to model has more
than two arguments and there seems no need to reify a permanent object to
model that knowledge. If the binary relations themselves had little
intrinsic meaning i would avoid creating them.

If i had a simple database with columns:
STORE DEPT PROD# AmountSold SalesAmount Week
for all the stores in a chain for all the weeks in a quarter, i would want
relations:
* weeklySalesByProductAndStore (4-ary)
* weeklySalesAmtByProductAndStore (4-ary)
* storeProductDepartment (3-ary)
Using them, i would want to calculate other n-ary relations:
* quarterlySalesByProductAndStore (4-ary)
* quarterlySalesAmtByProductAndStore (4-ary)
* weeklySalesByProduct (3-ary)
* weeklySalesAmtByProduct (3-ary)
* quarterlySalesByProduct (3-ary)
* quarterlySalesAmtByProduct (3-ary)
* averagePriceOfProductAtStoreByDate (4-ary)
I would allow most of these to use more general product type at many
levels. E.g., 6 oz. peach Organiq (tm) yogurt => peach Organiq yogurt, 6
oz. peach yogurt, => flavored Organiq yogurt, 6 oz flavored yogurt, peach
yogurt => Organiq yogurt, 6 oz. yogurt, flavored yogurt => yogurt => dairy
product, etc.

If i did not have n-ary relations, i would have to reify an object for
each of the hundreds of thousands (or millions) of uses of these relations
just to express binary relations about them. And the binary relations by
themselves would have little meaning.

-- doug foxvog

> - Think of a statement, fact, situation or piece of knowledge that you
> want to model or represent either in a KRR/ontological manner (i.e.
> computational focus) OR more abstractly in a purely symbolic logic.
> How would you, personally, determine whether n-ary is needed?
> (An obvious answer, already provided by others, is if there are more
> than 2 individuals or argument involved. Perhaps it's as simple as
> that,
> but i'd like to hear if there is a variety of views and approaches)
>
> - How would you, personally, decide if to assert binary or greater
> relations?
> Would it be based on convenience? Or trying to match as accurately or
> truthfully as possible what is expressed in the sentence or piece of
> knowledge? etc...?
>
>
> - If what you're trying to model seems to involve need of asserting an
>> *Actively open to work and PhD study opportunities, worldwide.*
>>
>> *-*
>> OntologPage <https://ontologforum.org/index.php/RobertRovetto>
>> IAOA Education Committee
>> Contact <https://ontospace.wordpress.com/contact/>
>>
>
> --
> 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/b63fbd24-5f3f-4d0e-a751-dc2c7fef2a9fn%40googlegroups.com.
>


doug foxvog

unread,
Sep 16, 2021, 11:24:08 PM9/16/21
to ontolo...@googlegroups.com
On Wed, September 15, 2021 04:17, Giancarlo Guizzardi wrote:
> Hi William,
>
> What I meant to say with that passage is the following.
>
> 1) yes, there are multiple ontological cuts that can be made to reality.
> For example, a four-dimensionalist such as Chris would
> believe that not only John&Mary's marriage but also John and Mary are
> events. Someone holding my views, in contrast, will
> see John, Mary, and John&Mary's Marriage as endurants to which there are
> events associated that will constitute John's Life, Mary's Life,
> John&Mary's Marriage's Life (actually, to each of these endurants, there
> are associated multiple possible lives but that is a different story -
> see ).

Note that if one can reify both events and endurants, then both of you can
reify the same objects: John, Mary, John&Mary'sMarriage, John'sLife,
Mary'sLife, John&Mary'sMarriage'sLife.

One then defines these objects as instances of the same classes: Person,
Marriage, Person'sLife, Marriage'sLife. These can be agreed to be
subclasses of the same superclasses: Animal, Agreement, Animal'sLife,
Agreement'sLife. At some higher level, the ontologies would separate into
whether things are events or endurants (if desired).

> 2) What I mentioned that both Chris and I would hold is the view that
> things like John, Mary and John&Mary's Marriage are things
> that really exist in the world. So, the piece of language "John&Mary's
> Marriage" picks up something really out there. Again, for Chris,
> this would be an event; for me, this can polysemically refer to both the
> endurant binding them and one of its possible lives (an event).

This is no problem. It does not have to force itself at the level of the
individual or the immediate classes that each individual is an instance
of.

> 3) Here is what I think you mean: we can have different conceptualizations
> to the same portion of reality. These conceptualizations would influence
> perspective, granularity, etc. Given this, we can see a phenomenon
> as a portion of blood (an amount of matter) or a collective of individual
> blood cells. To that I would agree.

Agreed.

-- doug foxvog
>>>>>>>> *From:* 'Aldo Gangemi' via ontolog-forum <
>>>>>>>> ontolo...@googlegroups.com>
>>>>>>>> *Sent:* 14 September 2021 12:36
>>>>>>>> *To:* ontolo...@googlegroups.com
>>>>>>>> *Cc:* Aldo Gangemi <gan...@mac.com>
>>>>>>>> *Subject:* Re: [ontolog-forum] Views on when to use n-ary
>>>>>>>> relations
>>>>>>>> greaterthanbinary - knowledge representation
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> … and that’s a chapter from the "Ontology Engineering with
>>>>>>>> Ontology
>>>>>>>> Design Patterns” book, about alternative representations of
>>>>>>>> multi-varied
>>>>>>>> predicates (n-ary relations), with pros and cons. The chapter
>>>>>>>> addresses
>>>>>>>> both extensional and intensional relations:
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> https://www.dropbox.com/s/y73e7qk0nja9zxw/multilayered-nary-patterns-odpbook.pdf?dl=0
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> Best
>>>>>>>>
>>>>>>>> Aldo
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> ________________________________
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> Aldo Gangemi
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> Full professor
>>>>>>>> University of Bologna
>>>>>>>>
>>>>>>>> Via Zamboni 32 40126, Bologna, Italy
>>>>>>>> <https://www.google.com/maps/search/Via+Zamboni+32%C2%A040126,+Bologna,+Italy?entry=gmail&source=g>
>>>>>>>> <https://groups.google.com/d/msgid/ontolog-forum/5de3b00f56c941d0b54a46131284977c%40bestweb.net?utm_medium=email&utm_source=footer>
>>>>>>>> .
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> --
>>>>>>>> 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/CAGG%2B%3DmVmW7O%3DXxjPDyN_jK9sAr_2Ht8Zbv_TH1Gh29SWNpgxww%40mail.gmail.com
>>>>>>>> <https://groups.google.com/d/msgid/ontolog-forum/CAGG%2B%3DmVmW7O%3DXxjPDyN_jK9sAr_2Ht8Zbv_TH1Gh29SWNpgxww%40mail.gmail.com?utm_medium=email&utm_source=footer>
>>>>>>>> .
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> --
>>>>>>>> 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/3B0F7534-B146-43A4-AB3B-0E32113621BF%40mac.com
>>>>>>>> <https://groups.google.com/d/msgid/ontolog-forum/3B0F7534-B146-43A4-AB3B-0E32113621BF%40mac.com?utm_medium=email&utm_source=footer>
>>>>>>>> .
>>>>>>>>
>>>>>>>> --
>>>>>>>> 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/043b01d7a966%246c7d2c60%2445778520%24%40gmail.com
>>>>>>>> <https://groups.google.com/d/msgid/ontolog-forum/043b01d7a966%246c7d2c60%2445778520%24%40gmail.com?utm_medium=email&utm_source=footer>
>>>>>>>> .
>>>>>>>>
>>>>>>> --
>>>>>>> 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/CAGG%2B%3DmUWJkFzejgCy2K7qgCuMnxzUhfJwAnrCmr2rhda3VCu4w%40mail.gmail.com
>>>>>>> <https://groups.google.com/d/msgid/ontolog-forum/CAGG%2B%3DmUWJkFzejgCy2K7qgCuMnxzUhfJwAnrCmr2rhda3VCu4w%40mail.gmail.com?utm_medium=email&utm_source=footer>
>>>>>>> .
>>>>>>>
>>>>>> --
>>>>>> 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/CAMWD8MpChCpx3sUuyvbNOLJ5SboDRrbEMgVpd80uxODSMi%3DHjQ%40mail.gmail.com
>>>>>> <https://groups.google.com/d/msgid/ontolog-forum/CAMWD8MpChCpx3sUuyvbNOLJ5SboDRrbEMgVpd80uxODSMi%3DHjQ%40mail.gmail.com?utm_medium=email&utm_source=footer>
>>>>>> .
>>>>>>
>>>>> --
>>>>> 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/CAGG%2B%3DmXkiTS8n7PdKFHDn4o5Es5FqjrvBZjKq94FKM4iCKGkDA%40mail.gmail.com
>>>>> <https://groups.google.com/d/msgid/ontolog-forum/CAGG%2B%3DmXkiTS8n7PdKFHDn4o5Es5FqjrvBZjKq94FKM4iCKGkDA%40mail.gmail.com?utm_medium=email&utm_source=footer>
>>>>> .
>>>>>
>>>> --
>>>> 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/CALuUwtB2AQUm1YGM9pg%2B90qJ%2BDtTCeQfejf%3DJSY10hh--v_jmg%40mail.gmail.com
>>>> <https://groups.google.com/d/msgid/ontolog-forum/CALuUwtB2AQUm1YGM9pg%2B90qJ%2BDtTCeQfejf%3DJSY10hh--v_jmg%40mail.gmail.com?utm_medium=email&utm_source=footer>
>>>> .
>>>>
>>> --
>>> 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/CAGG%2B%3DmWLbSK4OE2dgXF3VHpVzXpNSHtX356ucqhi8oQ2cSNUjA%40mail.gmail.com
>>> <https://groups.google.com/d/msgid/ontolog-forum/CAGG%2B%3DmWLbSK4OE2dgXF3VHpVzXpNSHtX356ucqhi8oQ2cSNUjA%40mail.gmail.com?utm_medium=email&utm_source=footer>
>>> .
>>>
>> --
>> 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/CALuUwtBbJ4cXWvV2vtn7NL2BQat-uf-%3DPXb8DKfgQVS4NMvOOQ%40mail.gmail.com
>> <https://groups.google.com/d/msgid/ontolog-forum/CALuUwtBbJ4cXWvV2vtn7NL2BQat-uf-%3DPXb8DKfgQVS4NMvOOQ%40mail.gmail.com?utm_medium=email&utm_source=footer>
>> .
>>
>
> --
> 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/CAGG%2B%3DmVWCWywTYq%2BE_%2BHSNaE157kYR_dnF4SzfYyVcaHgSmkNw%40mail.gmail.com.
>


doug foxvog

unread,
Sep 17, 2021, 12:43:28 AM9/17/21
to ontolo...@googlegroups.com
On Tue, September 14, 2021 16:38, William Frank wrote:
> Thanks, Giancarlo
>
> I found this very helpful:
>
>> Now, what makes something like a marriage a true endurant is that it has
>> essence and accidents
>> (like any other endurant). John&Mary's marriage is essentially a marriage
>> but only contingently
>> a happily marriage, a marriage with full separation of assets, etc.
>> Thanks, Giancarlo
>
> And this is part of what I was trying to say:
>...
> Marriage can refer both to an endurant and an event, and these two
> referents would bear some systematic relation between them.
> ...

Events are rarely instantaneous. They normally have a starting and ending
time, and thus endure for some period of time.

For me, i normally use the word "situation" for relatively static states
of affairs, and restrict my use of the word "event" to a subclass of
"situation" in which change is important throughout its duration.

But i understand the term "event" is used above for what i would call a
"situation".

A marriage is a situation that is useful to reify, since there is a lot to
say about it. The marriage starts with a subevent called a wedding and
ends with an event such as a death or divorce. Complex rules can be
created dealing with marriages

(implies
(and
(isa ?LAW BigamyLaw)
(isa ?MAR1 Marriage)
(isa ?MAR2 Marriage)
(partyTo ?MAR1 ?PERSON)
(partyTo ?MAR2 ?PERSON)
(partyTo ?MAR1 ?P2)
(partyTo ?MAR2 ?P3)
(different ?PERSON ?P2 ?P3) ; all 3 are different
(temporalExtent ?MAR1 ?PERIOD1)
(temporalExtent ?MAR2 ?PERIOD2)
(temporalExtent ?LAW ?PERIOD3)
(temporalOverlap ?PERIOD1 ?PERIOD2 ?PERIOD3 ?OVERLAP)
; ?OVERLAP is the temporal overlap of the 3 preceding time periods
(jurisdictionRegion ?LAW ?REGION)
(inRegionSometimeDuring ?PERSON ?REGION ?OVERLAP))
(inViolation ?PERSON ?LAW))
This says that a person is in violation of a bigamy law if they are in the
jurisdiction covered by that law at a time the law is in effect and they
are party to two marriages (or a marriage with at least 3 parties).

It would be more complex to write such a rule if marriages were not
reified or if only binary relations could be used.

-- doug f

doug foxvog

unread,
Sep 17, 2021, 12:54:07 AM9/17/21
to ontolo...@googlegroups.com
On 14 September 2021 14:31, Giancarlo Guizzardi wrote:

> I agree that relationship is "what one thing has to do with another" but
> we deny that this thing is (necessarily) a mathematical tuple.

This is in fact in agreement with Matthew West, who said (below):
>> the basic problem is that most of [referenced discussion] conflates
>> relationships ... with relations. ...
>> [A relationship] is a mathematical structure that is often
>> found useful in representing the [relation]

> In fact, for the most interesting material relationships, these "things"
> are concrete full-fledged endurants (e.g., marriages, enrollments,
> employments, presidential mandates).

It appears the three of us agree.

> The key to the analysis is understanding the nature of this mediating
> "thing" as the grounding of relations (relations hold, relationships
> exist)

> and truthmakers of relational propositions.

-- doug foxvog
> aldo.g...@unibo.it <mailto:aldo.g...@unibo.it>
> https://www.unibo.it/sitoweb/aldo.gangemi
>
>
>
> Director
>
> Institute for Cognitive Sciences and Technology,
>
> National Research Council (ISTC-CNR)
>
>
>
> http://scholar.google.com/citations?user=-iVGcoAAAAAJ
>
> skype aldogangemi
>
> twitter: @aldogangemi
>
> orcid: http://orcid.org/0000-0001-5568-2684
>
>
>
>
>
>
>
> On 14 Sep 2021, at 13:12, Giancarlo Guizzardi <gguiz...@gmail.com
> <mailto:ontolog-foru...@googlegroups.com> .
> <https://groups.google.com/d/msgid/ontolog-forum/5de3b00f56c941d0b54a46131284977c%40bestweb.net?utm_medium=email&utm_source=footer>
> .
>
>
>
> --
> 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/CAGG%2B%3DmVmW7O%3DXxjPDyN_jK9sAr_2Ht8Zbv_TH1Gh29SWNpgxww%40mail.gmail.com
> <https://groups.google.com/d/msgid/ontolog-forum/CAGG%2B%3DmVmW7O%3DXxjPDyN_jK9sAr_2Ht8Zbv_TH1Gh29SWNpgxww%40mail.gmail.com?utm_medium=email&utm_source=footer>
> .
>
>
>
> --
> 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/3B0F7534-B146-43A4-AB3B-0E32113621BF%40mac.com
> <https://groups.google.com/d/msgid/ontolog-forum/3B0F7534-B146-43A4-AB3B-0E32113621BF%40mac.com?utm_medium=email&utm_source=footer>
> .
>
> --
> 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/043b01d7a966%246c7d2c60%2445778520%24%40gmail.com
> <https://groups.google.com/d/msgid/ontolog-forum/043b01d7a966%246c7d2c60%2445778520%24%40gmail.com?utm_medium=email&utm_source=footer>
> .
>
> --
> 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/CAGG%2B%3DmUWJkFzejgCy2K7qgCuMnxzUhfJwAnrCmr2rhda3VCu4w%40mail.gmail.com
> <https://groups.google.com/d/msgid/ontolog-forum/CAGG%2B%3DmUWJkFzejgCy2K7qgCuMnxzUhfJwAnrCmr2rhda3VCu4w%40mail.gmail.com?utm_medium=email&utm_source=footer>
> .
>
> --
> 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/005501d7a9fd%24799fe120%246cdfa360%24%40gmail.com.
>


Giancarlo Guizzardi

unread,
Sep 17, 2021, 5:14:13 AM9/17/21
to ontolo...@googlegroups.com
On Fri, Sep 17, 2021 at 5:24 AM doug foxvog <do...@foxvog.org> wrote:
On Wed, September 15, 2021 04:17, Giancarlo Guizzardi wrote:
> Hi William,
>
> What I meant to say with that passage is the following.
>
> 1) yes, there are multiple ontological cuts that can be made to reality.
> For example, a four-dimensionalist such as Chris would
> believe that not only John&Mary's marriage but also John and Mary are
> events. Someone holding my views, in contrast, will
> see John, Mary, and John&Mary's Marriage as endurants to which there are
> events associated that will constitute John's Life, Mary's Life,
> John&Mary's Marriage's Life (actually, to each of these endurants, there
> are associated multiple possible lives but that is a different story -
> see ).

Note that if one can reify both events and endurants, then both of you can
reify the same objects: John, Mary, John&Mary'sMarriage, John'sLife,
Mary'sLife, John&Mary'sMarriage'sLife.

The point is that a fourdimensionalist would have that
John = John's life, Mary = Mary's life and John&Mary'sMarriage = John&Mary'sMarriage'sLife
 

One then defines these objects as instances of the same classes: Person,
Marriage, Person'sLife, Marriage'sLife.  These can be agreed to be
subclasses of the same superclasses: Animal, Agreement, Animal'sLife,
Agreement'sLife.  At some higher level, the ontologies would separate into
whether things are events or endurants (if desired).

> 2) What I mentioned that both Chris and I would hold is the view that
> things like John, Mary and John&Mary's Marriage are things
> that really exist in the world. So, the piece of language "John&Mary's
> Marriage" picks up something really out there. Again, for Chris,
> this would be an event; for me, this can polysemically refer to both the
> endurant binding them and one of its possible lives (an event).

This is no problem.  It does not have to force itself at the level of the
individual or the immediate classes that each individual is an instance
of.


If I understand correctly what you mean then all of these reinforce your main point, i.e.,
that one can define these classification and identity relations
a posteriori. Yes. However, in my view, Endurants and Perdurants (events, occurrences)
and very different in nature, and from a modeling perspective,
knowing that beforehand is central to the approach

best,
Giancarlo
 

Giancarlo Guizzardi

unread,
Sep 17, 2021, 5:38:15 AM9/17/21
to ontolo...@googlegroups.com
On Fri, Sep 17, 2021 at 6:43 AM doug foxvog <do...@foxvog.org> wrote:
On Tue, September 14, 2021 16:38, William Frank wrote:
> Thanks, Giancarlo
>
> I found this very helpful:
>
>> Now, what makes something like a marriage a true endurant is that it has
>> essence and accidents
>> (like any other endurant). John&Mary's marriage is essentially a marriage
>> but only contingently
>> a happily marriage, a marriage with full separation of assets, etc.
>> Thanks, Giancarlo
>
> And this is part of what I was trying to say:
>...
> Marriage can refer both to an endurant and an event, and these two
> referents would bear some systematic relation between them.
> ...

Events are rarely instantaneous.  They normally have a starting and ending
time, and thus endure for some period of time.

Absolutely but endurants exist as a whole (with all its parts together)
for the period of time that they endure; events unfold in time accumulating temporal parts.
So, when an event is occurring is (typically) only some of its proper parts that are there occurring
 

For me, i normally use the word "situation" for relatively static states
of affairs, and restrict my use of the word "event" to a subclass of
"situation" in which change is important throughout its duration.

But i understand the term "event" is used above for what i would call a
"situation".

In my view, events cannot genuinely change but I understand what you mean here:
change in the sense of variation, i.e., having proper parts that have incompatible properties
(e.g., the game was boring in the first half and exciting in the second one).
I understand your use of situation as a synonym to stasis, more or less homeomerous
occurrences.
 

A marriage is a situation that is useful to reify, since there is a lot to
say about it.  The marriage starts with a subevent called a wedding and
ends with an event such as a death or divorce. 

In my view, there is the event (wedding) that creates the (particular) Marriage as an endurant,
which in turn is constituted by relational modes (technically, a variable embodiment of relational modes - commitments, claims, intentions etc)
That wedding event then also kicks off the life of this entity. The life of this entity is btw the manifestation of exactly the relational modes.
I understand that it is useful to build a situation (perhaps a situoid) that encompasses the wedding event, the marriage as an event,
the marriage as an endurant, all events constituting that process, etc.

Again, a NL term "John&Mary's marriage" might at times refer to the marriage event, to the marriage endurant, to the whole situation, etc.
My main point is that we need the marriage as an endurant in that picture as well.

A final point: a situation such as this complex "marriage situation" would not really be homeomerous
and could not really be conceived as fully-determined entities (as traditional situations normally are).
In this paper, we sketch a view of what these situations could be:

best,
Giancarlo
Sure

best,
Giancarlo


-- doug f


--
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.

sowa @bestweb.net

unread,
Sep 17, 2021, 11:16:41 PM9/17/21
to ontolo...@googlegroups.com, ontolog...@googlegroups.com
Michael,
 
MDB:  One thing I learned the hard way is that there is no perfect model,  you can
always model a domain in multiple ways and thrashing around looking for the model
that is the "real" model leads to analysis paralysis.
 
I have been saying that for years.  The only perfect top-level ontology is the one that
has exactly one node, and exactly one axiom.  I'll call that node Entity, and the axiom
is very simple:  For every x, Entity(x).  In English this says "Everything that exists is
an entity." 
 
Underneath Entity, you need an infinite lattice of all possible types for all possible
options for classifying anything and everything that may exist.  Since the lattice
is infinite, you can't implement the whole thing.  What you do is implement as much
as you need whenever you need  it.
 
I wrote a paper about that in 2009:  "A dynamic theory of ontology",
 
And best of all, we implemented it in our old VivoMind company, and it worked
beautifully.  We had a clean, simple general-purpose ontology,which was
automagically extended in any way that might be needed.
 
The magic part of automagic was based on Cognitive Memory.  For an overview
of a wide range of applications, see http://jfsowa.com/talks/cogmem.pdf
 
Every one of those applications was paid for by a customer that specified the
problem to be solved and provided the data to be used and the specifications
of the software it had to be connected to.  All of them were legacy systems of
one kind or another, and we could not assume that any ontology we had could
be used without major extensions.
 
Our new company, called GraphStax, goes many steps farther.  The old
Cognitive Memory was based on conceptual graphs.  The new GraphStax
is based on stacks of graphs.  They include everything Cognitive Memory
could do, but with completely new algorithms that are bigger, faster, and more
general than the old ones.
 
That's what I had in mind when I presented the talk about a Universal Query
Language in August.   Here are the updated slides:  http://jfsowa.com/talks/uql.pdf
 
Section 6 of those slides talks about the need to support a dialogue with the
computer system.  The people converse with the computer in any language or
notation they prefer, and the computer system asks for help when it doesn't
understand.  The computer is NOT superintellligent, but it's smart enough to
know when it needs to ask for help.
 
And as I said, I don't worry about superintelligent computer systems during the
21st century.  But I do believe that we can develop good ,flexible systems that
know when they need to ask for help.
 
John

Azamat Abdoullaev

unread,
Sep 20, 2021, 2:54:18 PM9/20/21
to ontolog-forum
John wrote:
The only perfect top-level ontology is the one that
has exactly one node, and exactly one axiom.  I'll call that node Entity, and the axiom
is very simple:  For every x, Entity(x).  In English this says "Everything that exists is
an entity." 
The top node is rather the total whole of Everything, The World, or Reality. The first axiom is "Everything exists".
"And as I said, I don't worry about superintelligent computer systems during the
21st century".
There are a few ways to create a superintelligence — AI, whole brain emulation, brain implants, biological enhancement, networks or human-machine superminds.
 To create a superintelligence, one could link together existing natural and narrow artificial intelligences, as Rankbrain by Google / Google Search, Siri by Apple, Alexa by Amazon, Cortana by Microsoft and other virtual assistants, IBM's Watson. Image / facial recognition software, self-driving cars, etc.
I believe in the hybrid machine-human superintelligence (cyber-human superminds), which is now emerging in the context of the internet, Narrow ML/DL/AI applications, the internet of things and human brains. 
All what you need, it is a sort of psychometric g intelligence/master algorithm underlying your special "superhuman" intelligences. 
It is not as hard as generally presented.

--
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/3ef695463a0e4222b973afdc20e46346%40bestweb.net.

sowa @bestweb.net

unread,
Sep 20, 2021, 10:57:06 PM9/20/21
to ontolog-forum
Azamat,
 
All major breakthroughs in science and engineering have come from radically new innovations.  They don't come from a rehash of fragments that come out of a spare parts bin and are hooked up with scotch tape and bailing wire.
 
AA: There are a few ways to create a superintelligence — AI, whole brain emulation, brain implants, biological enhancement, networks or human-machine superminds.
 
That sounds like scotch tape and bailing wire for connecting pieces of humans with left over computer parts.
 
AI researchers have been talking about superintelligent things since the 1950s.  But there is nothing available today that can compete with a three-year-old child.
 
John

sowa @bestweb.net

unread,
Sep 21, 2021, 12:26:31 AM9/21/21
to ontolog...@googlegroups.com, ontolo...@googlegroups.com
Kingsley
 
That's the point of an open-ended lattice:
 
KI:  How about the notion: every entity is related to some other
entity, in a variety of ways.
 
When a mathematician says "infinity", that is short for saying
"as many as you need for any purpose you can imagine".
 
But I would add that there is no limit to the number of different
ways of carving up any experience of any kind for as many different
purposes, requirements, or points of view.
 
Just take a walk in the woods.  There is no clear way to decide
what aspects of what you see should be called entities.  No two
people would agree on how many entities they see.  In fact, no
single individual could find any way to stop counting.
 
You don't even need to walk.  Just stop anywhere.  There is no
definitive way to decide where to put the dividing lines that
determine which aspects of what you see should be considered
entities.
 
That's why I use the word 'infinity.  For any number you choose,
somebody else will find more.  There is no way to define a single
best or standard ontology for what you see anywhere in nature.
 
The built environment, however, usually does have parts that
some builder chose.  But if you give it some time, nature willt
turn that environment  into a natural one.  Nature --or chaos --
always wins.
 
John

Azamat Abdoullaev

unread,
Sep 21, 2021, 10:44:26 AM9/21/21
to ontolog-forum
All major breakthroughs in science and engineering have come from radically new innovations.  
Agree. Today, disruptive discoveries and technologies could be only transdisciplinary solutions, as Real-World AI or NBIC technologies. 
They don't come from a rehash of fragments that come out of a spare parts bin and are hooked up with scotch tape and bailing wire.
Some patching could change many, call it the "neocortex effect". Adding additional layers on the cerebral cortex for several millions years resulted in higher cognitive functioning.
Again, machine learning is just adding a patch to your databases enhancing data processing functionalities.
In all, there are incremental, sustained, revolutionary or radical and disruptive or foundational innovations.
"AI researchers have been talking about superintelligent things since the 1950s.  But there is nothing available today that can compete with a three-year-old child".
Indeed, we need to go for a real-world AI paradigm shift, recalling how it was originally defined, as a science and engineering producing intelligent machines without respect of biological intelligence. 
 

--
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.

Nadin, Mihai

unread,
Sep 21, 2021, 11:05:22 AM9/21/21
to ontolo...@googlegroups.com

Azamat Abdoulaev wrote: Indeed, we need to go for a real-world AI paradigm shift, recalling how it was originally defined

Here it is:

John McCarthy called it “artificial intelligence.” The participants at the 1956 Dartmouth Conference (formally, the Dartmouth Summer Research Project on Artificial Intelligence) agreed: “…every aspect of learning or any other feature of intelligence can be so precisely described that a machine can be made to simulate it.” Machine meant “an agent that manipulates symbols.” The extended workshop at Dartmouth College (where McCarthy, who applied for funds at the Rockefeller Foundation, was teaching) was documented in detail by Ray Solomonoff (1956).

(for more: In folly ripe. In reason rotten. Putting machine theology to rest;

https://arxiv.org/abs/1712.04306

Mihai Nadin

Azamat Abdoullaev

unread,
Sep 21, 2021, 11:18:55 AM9/21/21
to ontolog-forum

What is AI? / Basic Questions

It is the science and engineering of making intelligent machines, especially intelligent computer programs. It is related to the similar task of using computers to understand human intelligence, but AI does not have to confine itself to methods that are biologically observable. 

sowa @bestweb.net

unread,
Sep 21, 2021, 4:01:40 PM9/21/21
to ontolog-forum
Azamat,
 
I went to that website and browsed some links.  I expect more and better applications to come from such developments.  But they have  correctly emphasized collaboration between humans and computers.  The humans are definitely in charge.
 
I would expect better interfaces that facilitate communication between humans and their computer systems.  But that would be at the same level as humans and their horses.  That was an excellent interface for the 19th century, and it would be great to have a similar collaboration with computers for the 21st c.
 
Nothing they are doing would suggest superintelligent systems at any time during the 21st c..
 
John                                                                                                                     

bruces...@cox.net

unread,
Sep 21, 2021, 4:47:36 PM9/21/21
to ontolo...@googlegroups.com

Thanks for this comment, and for highlighting the same idea in a second illustration.

 

Unless some specific force or motive or context drives the selection of specific facets of some situation or object, the generalization of ontology is a lost cause.  There is no doubt this is true.

 

So – any attempt to “universalize ontology” should build this principle into the design.  In this sense, every working ontology or categorization of reality into words and categories is “context specific”.  Ontology must contain a motivational principle that creates a specifying criteria for selection – in this case, “what the customer wants”.

 

> Every one of those applications was paid for by a customer that specified the problem to be solved and provided the data to be used and the specifications of the software it had to be connected to. 

 

It’s not that “a universal ontology is impossible.”  It’s that it must be universally fluent and adaptive to infinite variation in all its dimensionality, so that it can respond to every possible specific context and situation for every possible purpose.

 

Bruce Schuman

Santa Barbara CA USA

bruces...@cox.net / 805-705-9174

www.origin.org / www.integralontology.net / www.newcongress.net

 

 

****

 

John

 

********

hpo...@verizon.net

unread,
Sep 21, 2021, 5:26:00 PM9/21/21
to ontolo...@googlegroups.com

Bruce,

 

I agree, but would like to point out that one can generalize the scope of any given context-specific ontology beyond what a single customer wants. And many ontologies do just that. The issue from my perspective is that we don’t have a generally accepted way for describing what specific range of customers and contexts in which a given ontology might be applicable or useful. The SCOPE model is an extensible attempt to provide such a multi-dimensional scope “yardstick” for certain types of contexts.

 

We don’t have to choose between the “universal” ontology and the context-specific ontology. Rather we can choose ontologies that are useful in many commonly encountered contexts and scopes of applicability – and get buy in from the target group of customers. But we have to be specific about what their scope of applicability is, and ideally, make that information available as part of the representation of that  ontology itself. This description of the applicable scope dimensions doesn’t have to be universally understandable – just by the target user and application set and applicable scope ranges.

 

If we have a growing body of such ontologies, we can then develop publicly accessible mappings between them, along with warnings/caveats about where the mappings won’t work, or work only with some pragmatic work-arounds that have some known flaws/shortcomings. Potential users can then decide whether to use them, develop their own mappings, or develop some new context-specific or domain specific ontology that addresses whatever shortcoming the existing ontologies have for their current purposes (assuming that’s feasible/pragmatic).

 

Hans Polzer

--

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.

bruces...@cox.net

unread,
Sep 23, 2021, 6:20:32 PM9/23/21
to ontolo...@googlegroups.com

Hans –

 

Good points, and thank you.  I did study the SCOPE project a couple of years ago, and put the PDF online here: http://originresearch.com/docs/ScopeInteroperability2008.pdf

 

Right now, I am laughing, thinking about Azamat’s claim that supercomputers are going to figure out how human beings are going to cooperate to solve problems like the United Nations sustainable development goals (SDGs).  I’d like to see a rough sketch of that algorithm (step by step process to make it happen).  How do they fix what John Sowa recently called “lying politicians”?

 

So, my thoughts on “universal ontology” are nested inside this kind of thinking.  It so happens that just today, I sent an idealistic 1200 word essay to a group I work with called The Great Transition Initiative https://greattransition.org/ for publication in their current dialogue about a “World Constitution” and the problems of global governance.  Earlier this morning, I watched a fascinating YouTube video on “the dark side” of life in the Sahara desert in Mauritania, where (among other things) slavery is still practiced.

 

What could ontology do to support some global governance or stability initiative?  I’d say it would need to embody a very strong inherent ethic supporting some of kind of homeostatic/cybernetic model of democracy. 

 

This kind of thinking influences my motivation for a universal or extremely inclusive ontology, that innately fastens all these fundamental parts together. I was just now looking at the “Sowa Diamond” model – and not understanding how that model supports an integral ontology.  http://www.jfsowa.com/ontology/toplevel.htm

 

This is a hint on my motivation. I want to find a way to get beyond inherent “localization of motivation” and the assumption that this is the only way to go.  I incline to the claim that this is an industry-wide myopia (short-sightedness), and we should be pushing for breakthrough, rather than presuming that the subject is impossible and there’s no way forward.

 

In this comment, I want to consider the claim that it might (?) be innate underlying presumptions or tradition that limit the scope of what an ontology can do.  I know this might be  bit reckless – but my idea right now is that the entire project or industry is thrown a little off-balance by the dominance of an empirical (or bottom up) point of view, which functions by looking at the world and trying to organize the objects or events that are visible or apparent.  My argument might simply be that we ought to start with the category system first, and then converge that framework towards a perfect fit with the real-world empiricism or experience. This approach might then hope to argue that every specific subset can be correctly interpreted as part of a global whole, where all the relations are or could be highly ordered.

 

Somebody here more knowledgeable than I might immediately show why this is wrong, but I think what is happening is that professional working ontologists are very naturally looking around their world, and forming lists of the things they see or can identify, and building category systems to organize those objects.  Perhaps the prevailing assumption is that “the objects are (of course) naturally prior to the categories we place them in.”

 

Here’s a picture of a lattice, that looks a lot like the Sowa Diamond.  I would need to study this more to see how this idea supports a working practical ontology, and maybe somebody could show me an example. https://en.wikipedia.org/wiki/Lattice_(order)#/media/File:Lattice_of_partitions_of_an_order_4_set.svg

 

I want to explore ways to supplant the lattice model of ontology with a hierarchical model.  Why can’t a working ontology be 100% linearly hierarchical?  If this approach could work, I think the entire subject of semantic ontology could explode towards generalization.  I might take a little time to consider how all the elements defined in the Sowa Diamond could be rearranged or interpreted within a hierarchy rather than a lattice.

 

****

 

HP> I agree, but would like to point out that one can generalize the scope of any given context-specific ontology beyond what a single customer wants.

 

Bruce> Thank you.  I am thinking about how that could work (explore what customers and situations have in common, etc.)

 

HP> And many ontologies do just that. The issue from my perspective is that we don’t have a generally accepted way for describing what specific range of customers and contexts in which a given ontology might be applicable or useful. The SCOPE model is an extensible attempt to provide such a multi-dimensional scope “yardstick” for certain types of contexts.

 

Bruce> I thought the SCOPE model was a good example of cooperative development.

 

HP> We don’t have to choose between the “universal” ontology and the context-specific ontology. Rather we can choose ontologies that are useful in many commonly encountered contexts and scopes of applicability – and get buy in from the target group of customers. But we have to be specific about what their scope of applicability is, and ideally, make that information available as part of the representation of that  ontology itself. This description of the applicable scope dimensions doesn’t have to be universally understandable – just by the target user and application set and applicable scope ranges.

 

Bruce> Yes. Maybe the buy-in from customers reflects their adaptability to work with assumptions or concepts shared by other customers.

 

I am visualizing these more specific solutions in a very broad context, such as the UN SDGs.  So, I want to see “interdisciplinary interoperability” in a single context, in a single space.  We need cooperative interaction from 1000+ independent agencies exerting influence across many disciplines, and learning from one another.  Maybe this includes private citizens, local governments, etc.   Plus, we want a built-in ethic.

 

HP> If we have a growing body of such ontologies, we can then develop publicly accessible mappings between them, along with warnings/caveats about where the mappings won’t work, or work only with some pragmatic work-arounds that have some known flaws/shortcomings.

 

Bruce> Yes.  And maybe (?) broader general standards could emerge from these interactions (??)

 

HP> Potential users can then decide whether to use them, develop their own mappings, or develop some new context-specific or domain specific ontology that addresses whatever shortcoming the existing ontologies have for their current purposes (assuming that’s feasible/pragmatic).

 

Bruce> Yes.  I think I am presuming a kind of common underlying motivation – that may seldom exist in most commercial contexts.  Yes, I want to sell my product, but I want to do it in a collaborative context.

 

Just one thought that is floating in my head regarding “top level ontology”.  Instead of calling the top level “thing” – how about calling it “everything” – and then partition that space under local motivation.... (?)

 

And just as I post this, I get a news flash on my phone about advances in “reservoir computing”

https://www.sciencealert.com/a-new-neural-network-solved-the-hardest-of-maths-problems-a-million-times-faster

John Bottoms

unread,
Sep 23, 2021, 7:01:41 PM9/23/21
to ontolo...@googlegroups.com

Bruce,

re: Azamaat's Claim:

"If you wish to make an apple pie from scratch, you must first invent the Universe".  - Carl Sagan

Are we doing AI now, not ontology? Or, are you asking how do we make an ontology that can hold the universe?

-John Bottoms
 FirstStar Systems
--
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.

bruces...@cox.net

unread,
Sep 23, 2021, 9:01:53 PM9/23/21
to ontolo...@googlegroups.com

Hi John. I’m not sure I understand your question.

 

But this below quick Google search probably makes my point.

 

If an ontology is “a specification of the meanings of symbols in an information system – a specification of a conceptualization”, I am interested in ways to generalize that process.  So, this involves a theory of semantics and the structure of concepts.

 

People on Ontolog have argued about “what a concept is” – but I personally think this is a silly question.  If you keep it simple, you understand it.  If you think about it too much, you might get confused, and invent some goofy or meaningless projects (“there are no such things as trees”).

 

What I am looking for is common ground – a place where the human community can come together and figure out how we can live together in our shrinking one-world context.

 

Maybe the starting point for me is 0 – the coordinate origin.  Let’s agree on that. To paraphrase Lord of the Rings, “One zero to rule them all”

 

But I tend to suppose we CAN figure out a “concept of all concepts” – a “concept that contains all concepts” – and which is formed from (defined in terms of) a universal symbolic language, probably grounded in  binary, from which every possible abstract symbolic structure can be defined.

 

So let’s define that thing – and maybe get it running in a global network of AI super computers, as their common ground and maybe their operating system, or some level within it.

 

Let’s not do philosophy as “hot air rising”.  Let’s do it as highly motivated real-world systems engineering for a world that works.

 

image001.png
aiontology.png

Michael DeBellis

unread,
Sep 23, 2021, 10:06:40 PM9/23/21
to ontolo...@googlegroups.com
People on Ontolog have argued about “what a concept is” – but I personally think this is a silly question.  If you keep it simple, you understand it.  If you think about it too much, you might get confused, and invent some goofy or meaningless projects (“there are no such things as trees”).

Actually, anthropologist Scott Atran has written some interesting things about this. See his book Folkbiology with Douglas Medin. While all cultures have the concept of "tree" in reality there is no precise definition for a tree in botany. This shows the difference between science and common sense or folk science. It also shows why we can't just "keep it simple" and just hand waving away problems by saying things are "silly questions" is exactly the opposite of good science. In Galileo's time it was a "silly question" to ask why some substances like rocks fell to the Earth and others like steam rose. Every educated person "knew" that it was just the nature of certain substances to seek their natural level. As Chomsky has said: 

"One of the great breakthroughs of modern science, Galileo and the Galilean period, was the ability to be puzzled about things that seemed obvious… that’s a tremendous step forward, for literally thousands of years it had been accepted by important scientists that we have answers to some very simple questions like [cups fall and steam rises because as Aristotle said] they are seeking their natural place.”


Michael

You received this message because you are subscribed to a topic in the Google Groups "ontolog-forum" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/ontolog-forum/0pDK8IJiFDc/unsubscribe.
To unsubscribe from this group and all its topics, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/004001d7b0df%24c12d3540%2443879fc0%24%40cox.net.

sowa @bestweb.net

unread,
Sep 23, 2021, 11:52:29 PM9/23/21
to ontolo...@googlegroups.com, ontolog...@googlegroups.com
Michael,
 
All these issues can be stated, analyzed, and represented by anybody who has taken a solid course in logic
 
MDB: People on Ontolog have argued about “what a concept is” – but I personally think this is a silly question.  If you keep it simple, you understand it.  If you think about it too much, you might get confused, and invent some goofy or meaningless projects (“there are no such things as trees”).
 
MDB:  While all cultures have the concept of "tree" in reality there is no precise definition for a tree in botany. This shows the difference between science and common sense or folk science. It also shows why we can't just "keep it simple"
 
The word 'tree' has a perfectly good scientific definition in botany:  A plant with tall, solid woody stem.  For example, a rose bush and an apple tree are closely related plants, but roses have weak stems and grow as bushes, but apple trees have tough woody stems and grow as trees.  But a fern tree and an oak tree are widely different kinds of plants, but they both have tough woody stems.  This is  one of many reasons why an ontology must be a lattice with criss-crossing links
 
Fundamental principle:  Every widely used word in any language has proved its value for many applications, and every general-purpose ontology must be able to represent it.  Any ontology that cannot represent every word in every language may be useful for some special-purpose applications, but it is not acceptable as a general purpose ontology.
 
John

Azamat Abdoullaev

unread,
Sep 24, 2021, 7:08:00 AM9/24/21
to ontolo...@googlegroups.com
I found it by chance in the Spam folder:

"Right now, I am laughing, thinking about Azamat’s claim that supercomputers are going to figure out how human beings are going to cooperate to solve problems like the United Nations sustainable development goals (SDGs).  I’d like to see a rough sketch of that algorithm (step by step process to make it happen).  How do they fix what John Sowa recently called “lying politicians”?

 So, my thoughts on “universal ontology” are nested inside this kind of thinking.  It so happens that just today, I sent an idealistic 1200 word essay to a group I work with called The Great Transition Initiative https://greattransition.org/ for publication in their current dialogue about a “World Constitution” and the problems of global governance.  Earlier this morning, I watched a fascinating YouTube video on “the dark side” of life in the Sahara desert in Mauritania, where (among other things) slavery is still practiced.

 What could ontology do to support some global governance or stability initiative?  I’d say it would need to embody a very strong inherent ethic supporting some of kind of homeostatic/cybernetic model of democracy".


It looks, the author missed the whole message. The things are exactly opposite, the RSI involves the collective human intelligence enhanced with SUPER-POWERFUL FRIENDLY COMPLEMENTARY machine intelligence:  

As for the RSI algorithm, it is currently classified as a trade secret, as, I believe, the Scope MIC widely practices. As a side note, all their market cap will not be enough to buy out the RSI algorithms. 
Still, some general points for smart minds are publicly accessible in the EU AI Futurium site:  

--
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.

Azamat Abdoullaev

unread,
Sep 24, 2021, 10:57:35 AM9/24/21
to ontolo...@googlegroups.com

Bruce wrote: 

But I tend to suppose we CAN figure out a “concept of all concepts” – a “concept that contains all concepts” – and which is formed from (defined in terms of) a universal symbolic language, probably grounded in  binary, from which every possible abstract symbolic structure can be defined.

So let’s define that thing – and maybe get it running in a global network of AI super computers, as their common ground and maybe their operating system, or some level within it.

A “concept that contains all concepts” is "Everything/World/Realty".

A “concept that contains no concepts or contained in all concepts” is "Nothing".

And "Anything" is between the supremum and infimum in the world' lattice order, like the Hasse diagram:

image.png

This paradigm makes the core of the RSI's Reality model and its Master algorithm.

https://futurium.ec.europa.eu/en/european-ai-alliance/posts/real-superintelligence-why-and-how-we-create-top-human-machine-technology-platform



Reply all
Reply to author
Forward
0 new messages