Are Relationship Triples also Nodes?

49 views
Skip to first unread message

Frank Guerino

unread,
Apr 21, 2016, 12:15:01 PM4/21/16
to Ontolog Forum
Hello All,

I’d like the community's opinions on the following:

Given a Graph/Network that is composed of Nodes, Edges (a.k.a. Links) and Relationships….
Defining a Node as an entity that has descriptive traits and state.
Defining an Edge as the descriptive connection between any two Nodes (e.g. a Predicate)…
Defining a Relationship as consisting of a Subject Node that is linked through an Edge to an Object Node (a.k.a. a Triple)…

Do you believe that a Relationship Triple can also be treated as a Node and why?

If we decompose the Relationship as a set of 3 components, each component can point to one or more things.  For example, if we have a Relationship…

{
    “Subject Node”: “Tom Smith”,
    “Predicate”: “Product Owner”,
    “Object Node”: “Product ABC”
}

1. The Subject Node points to all data about "Tom Smith”.
2. The Object Node points to all relevant data about "Product ABC”.
3. The Predicate points to all Relationships that contain the Predicate “Product Owner”.

(NOTE: Each of the above elements (Nodes & Predicates) can also point to other things like Voids or Holes, even though the above example does not show it.)

(NOTE: We’ve already proven all of these pointer types in our research so we know they do, in fact, exist and work as described above.)

While the Relationship can exist as an independent entity, on its own, if it is treated as a Node what else do you think that set would/could point to?  In other words, what can that independent Triple, as a whole and singular entity, also point to?

Thanks, in advance, for any thoughts you offer.

My Best,

Frank
--
Frank Guerino, Chairman
The International Foundation for Information Technology (IF4IT)
http://www.if4it.com
1.908.294.5191 (M)

Jack Park

unread,
Apr 21, 2016, 12:19:51 PM4/21/16
to ontolo...@googlegroups.com
One thought comes to mind.

A trivial example:

"Most scientists believe that X causes Y."

That's two triples one nested inside the other. Unless the nested triple {X, causes, Y} is an addressable node, I'm not sure how you would implement the nest.

Cheers
Jack

--
All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontolog Forum content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
---
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 post to this group, send email to ontolo...@googlegroups.com.
Visit this group at https://groups.google.com/group/ontolog-forum.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/D33E74BD.145BF9%25Frank.Guerino%40if4it.com.
For more options, visit https://groups.google.com/d/optout.

David Price

unread,
Apr 21, 2016, 12:52:51 PM4/21/16
to ontolo...@googlegroups.com
Hi Frank,

This is a well-worn discussion. Search for “RDF statement reification” for example. The upper ontologies (e.g. ISO 15926) typically take a reification view of Relationships. However, what happens is that it makes the simple cases overly complex so “normal” people look for shorthand approaches to using simple non-reified relations too. The decision about whether to reify or not is usually domain- or even app-specific.

Cheers,

Kingsley Idehen

unread,
Apr 21, 2016, 1:44:55 PM4/21/16
to ontolo...@googlegroups.com
These matters have been thrashed out over the years (note comment from David Price about Reification etc.).

What I want to add to this conversation is a tool [1] for constructing sentences/statements that describe entities and entity relationships [2][3]. It is my hope that tools like this clear out some of the inertia swirling around this matter i.e., the ability to move the power of language (systematic use of signs, syntax, and semantics to encode and decode information [data in some context] ).

Just click on the buttons in the UI to get your preferred presentation of the entity relationships loaded from the DBpedia example.

Links:

[1] http://osde.openlinksw.com -- Structured Data Editor

[2] http://linkeddata.uriburner.com/rdf-editor/#/editor?uri=http:%2F%2Fdbpedia.org%2Fresource%2FParis&view=statements -- EAV view

[3] http://linkeddata.uriburner.com/rdf-editor/#/editor?uri=http:%2F%2Fdbpedia.org%2Fresource%2FParis&view=triples -- RDF View

-- 
Regards,

Kingsley Idehen	      
Founder & CEO 
OpenLink Software     
Company Web: http://www.openlinksw.com
Personal Weblog 1: http://kidehen.blogspot.com
Personal Weblog 2: http://www.openlinksw.com/blog/~kidehen
Twitter Profile: https://twitter.com/kidehen
Google+ Profile: https://plus.google.com/+KingsleyIdehen/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen
Personal WebID: http://kingsley.idehen.net/dataspace/person/kidehen#this

Matthew West

unread,
Apr 21, 2016, 4:50:29 PM4/21/16
to ontolo...@googlegroups.com

Dear Frank,
If you take a nodes and edges view of the world then in principal any edge can be replaced by a node and two more edges. The problem is where to stop? The thing to note is that you can say something about nodes (more edges to other nodes) but not about edges. In my analysis my conclusion was that there was no more information to record after you had identified the state playing the role in a relationship. So no point going beyond that. That is how ISO15926 works. You can always have convenient short cuts across that base layer. In SO 15926 these are templates.
Regards
Matthew West


--
All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontolog Forum content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
---
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 post to this group, send email to ontolo...@googlegroups.com.
Visit this group at https://groups.google.com/group/ontolog-forum.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/57191193.9070100%40openlinksw.com.

Joao Paulo Almeida

unread,
Apr 21, 2016, 5:11:35 PM4/21/16
to ontolo...@googlegroups.com
On the topic of reification of relationships, I'd suggest a recent paper by Giancarlo Guizzardi and Nicola Guarino. 

João Paulo 
--
All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontolog Forum content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
---
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 post to this group, send email to ontolo...@googlegroups.com.
Visit this group at https://groups.google.com/group/ontolog-forum.

Frank Guerino

unread,
Apr 22, 2016, 9:06:48 AM4/22/16
to Ontolog Forum
Hi Joao Paulo,

I read through this paper and, while I agree with much of what the authors present, I do have some foundational disagreements with the them on some specific topics.  For example, they view concepts such as Marriages, Enrollments, Purchases, Mandates, Employments, etc. as types of intrinsic relations.  We’ve found through our work that it is not easy to model the above as relations but that it is much easier, instead, to model them as instances of Nouns or Classes that are composed of elements that have relationships to other Nouns and/or Relationships.

For example, “a Marriage” is an instance of bounded data that contains relationships to help support it…

{
    “Marriage Transaction ID”: “1234”,
    “Marriage Date”: “1/1/1999”,
    “Spouses”: ["Mary Smith", "John Doe”],
    “Marriage Location”: “Facility XYZ”
}

This means that I can now have many specific Marriages that can be represented just like the above and which can exist as individual Nodes in a Graph.

Also, there are many relationships that support the above marriage transaction instance (i.e. “a specific Marriage”)…

Person Mary Smith is related as a Spouse to Person John Doe
Person John Doe is related as a Spouse to Person Mary Smith
Location Facility XYZ is related as a Marriage Location to Marriage Transaction ID 1234
Location Facility XYZ is related as a Marriage Location for Person Mary Smith
Location Facility XYZ is related as a Marriage Location for Person John Doe
Date “1/1/1999” is related as a Marriage Date to Marriage Transaction ID 1234
Date “1/1/1999” is related as a Marriage Date for Person Mary Smith
Date “1/1/1999” is related as a Marriage Date for Person John Doe
etc. (too many too list)

This means that the Marriage transaction instance, above, exists as a singular Noun Instance (i.e. a Node in a Graph), not as a Relationship, and has many Relationships that help support it and that help categorize it as a Marriage.

I’ve found (through modeling these concepts in our own semantic platform called NOUNZ) that this reasoning holds true for all of the Noun Types / Class Types presented above (Marriages, Enrollments, Purchases, etc.).

While I agree with the authors on the concept that Relationships can be treated as full endurants, I didn’t see anything in their article that proved Relationships (as Triples) can point to anything.  For example, if we have the Relationship:

Person Mary Smith is related as a Spouse to Person John Doe

It is the predicate “Spouse” and not the Relationship as a whole that points to the data type Marriage. And, it is the Person’s in the Relationship (and not the Relationship as a whole) that are related to a specific Marriage transaction/event.

I believe, that Relationships can be treated as Nodes in cases where, for example, they point to histories of themselves (implying they have temporal traits such as timestamps and state).  However, it appears that what Relationships can point to is very limited, which is why I asked the original question… to see if others can help prove this belief right or wrong.

Thanks, again, for the article,

Frank
--
Frank Guerino, Chairman
The International Foundation for Information Technology (IF4IT)
http://www.if4it.com
1.908.294.5191 (M)




Frank Guerino

unread,
Apr 22, 2016, 9:18:51 AM4/22/16
to Ontolog Forum
Hi Jack,

In your example "Most scientists believe that X causes Y.”, I see many nested Triples…

There are the obvious ones…

X causes Y
Y is caused by X

And there are many more complex ones…

“Most scientists” implies that in a set of all scientists (all Scientist Nodes in a Graph) there exists a subset of specific scientists (“most but not all” that represent some of those Scientist Nodes) that believe the two cause & effect relationships above.  This implies that most Scientist Nodes have existing Relationships between them and each of the above cause & effect Relationships, such that the number of relationships between those that believe is a higher count/percentage than those that do not believe (or don’t care).

I believe your example helps highlight that Nodes can point to Relationships but I do not see anything in it that it that helps depict whether or not Relationships, as holistic entities, can point to other Nodes or Relationships.  Maybe I’m wrong.

Thanks for example,

Frank
--
Frank Guerino, Chairman
The International Foundation for Information Technology (IF4IT)
http://www.if4it.com
1.908.294.5191 (M)


From: Jack Park <jack...@gmail.com>
Reply-To: Ontolog Forum <ontolo...@googlegroups.com>
Date: Thursday, April 21, 2016 at 12:19 PM
To: Ontolog Forum <ontolo...@googlegroups.com>
Subject: Re: [ontolog-forum] Are Relationship Triples also Nodes?

Frank Guerino

unread,
Apr 22, 2016, 9:26:57 AM4/22/16
to Ontolog Forum
Hi David,

Thanks for the suggestion.  I haven’t found anything in my reification searches that clearly spells out whether or not Relationships as whole entitles can point to other Nodes or Relationships in a Graph.  I still keep looking, though.

My Best,

Frank
--
Frank Guerino, Chairman
The International Foundation for Information Technology (IF4IT)
http://www.if4it.com
1.908.294.5191 (M)

From: David Price <dpr...@topquadrant.com>
Reply-To: Ontolog Forum <ontolo...@googlegroups.com>
Date: Thursday, April 21, 2016 at 12:52 PM
To: Ontolog Forum <ontolo...@googlegroups.com>
Subject: Re: [ontolog-forum] Are Relationship Triples also Nodes?

Frank Guerino

unread,
Apr 22, 2016, 10:16:34 AM4/22/16
to Ontolog Forum
Hi Kingsley,

Thanks for the pointer to OSDE.  We use a different approach.  We have developed a data compiler (called NOUNZ) that, among many other things, harvests Semantic Relationships from Nodes, which exist as real data records within a specific context or domain.  For example, in the sample links I provided, the domain is “an Enterprise” and the Nodes are the data types and records found in common enterprises, where the Relationships are those found between such context-specific records.

I took a look at OSDI and found it very interesting.  My limited assessment leads me to believe that difference between OSDI and NOUNZ is the approach used to arrive at an ontology specific model that is loaded with real data.

  • In OSDE, it appears that we create named graphs from individual documents right in and from the editor, which leads to the ontology specific model as a living artifact.  This seems to imply a paradigm where we work “one document at a time.”  I found the OSDS OSDE integration video, which used the single web page “CIO Summit Survey” as an example, to be quite telling.  It shows the extraction and creation of relationships being done manually, one at a time, which could take forever when dealing with many web pages.
  • In the case of NOUNZ, we use large quantities of structured data from multiple sources (e.g. Systems/Applications) as inputs, parse that data, map it to ontological elements, and then yield an output that contains many named graphs (with many overlaps between them), all which are part of an ontology specific model that represents a domain.  This allows us to deal with millions of different data records of different data types, which may have many tens of millions of relationships, and reverse engineer the results directly from the data.  Also, the goal is to ensure that the outputs are in human and computer consumable forms because most people, for example, don’t speak RDF, WOL, SKOS, Turtle, etc.

I have ideas about ways to combine the two paradigms.

My Best,

Frank
--
Frank Guerino, Chairman
The International Foundation for Information Technology (IF4IT)
http://www.if4it.com
1.908.294.5191 (M)


From: Kingsley Idehen <kid...@openlinksw.com>
Reply-To: Ontolog Forum <ontolo...@googlegroups.com>
Date: Thursday, April 21, 2016 at 1:44 PM
To: Ontolog Forum <ontolo...@googlegroups.com>
Subject: Re: [ontolog-forum] Are Relationship Triples also Nodes?

These matters have been thrashed out over the years (note comment from David Price about Reification etc.).

What I want to add to this conversation is a tool [1] for constructing sentences/statements that describe entities and entity relationships [2][3]. It is my hope that tools like this clear out some of the inertia swirling around this matter i.e., the ability to move the power of language (systematic use of signs, syntax, and semantics to encode and decode information [data in some context] ).

Just click on the buttons in the UI to get your preferred presentation of the entity relationships loaded from the DBpedia example.

Links:

[1] http://osde.openlinksw.com -- Structured Data Editor

[2] http://linkeddata.uriburner.com/rdf-editor/#/editor?uri=http:%2F%2Fdbpedia.org%2Fresource%2FParis&view=statements -- EAV view

[3] http://linkeddata.uriburner.com/rdf-editor/#/editor?uri=http:%2F%2Fdbpedia.org%2Fresource%2FParis&view=triples -- RDF View

-- 
Regards,

Kingsley Idehen	      
Founder & CEO 
OpenLink Software     
Company Web: http://www.openlinksw.com
Personal Weblog 1: http://kidehen.blogspot.com
Personal Weblog 2: http://www.openlinksw.com/blog/~kidehen
Twitter Profile: https://twitter.com/kidehen
Google+ Profile: https://plus.google.com/+KingsleyIdehen/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen
Personal WebID: http://kingsley.idehen.net/dataspace/person/kidehen#this

--
All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontolog Forum content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
---
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 post to this group, send email to ontolo...@googlegroups.com.
Visit this group at https://groups.google.com/group/ontolog-forum.

Frank Guerino

unread,
Apr 22, 2016, 10:44:17 AM4/22/16
to Ontolog Forum
Hi Matthew,

You wrote: “If you take a nodes and edges view of the world then in principal any edge can be replaced by a node and two more edges."

I don’t know that I’ve personally found this statement to be true.  Instead, I’ve found that Edges (i.e. Predicates) can point all other similar Edges in the Graph, in turn yielding all Semantic Relationships that contain the same Predicate (e.g. Predicate = “Subject Matter Expert”).

You wrote: “In my analysis my conclusion was that there was no more information to record after you had identified the state playing the role in a relationship. So no point going beyond that.”

Is your analysis public and, if so, is there a way for me to access it?  I’d love to read it.

In my own analysis, I did see things like Relationships being able to point to the historical instances of all similar relationships.  For example,  in the relationship:

Person Jane Doe is related as a Product Owner to Product XYZ

I’ve found that we could have multiple instances of temporal existence.

  • Jane may have been a Product Owner (1/1/1992),
  • then left the company (12/31/2000),
  • then returned to the company in the same role to again be a Product Owner (1/1/2016).

In this example, the above whole Semantic Relationship can point to all of its temporal instances in time to create a historical view of all cases where the Relationship did or did not exist in the broader temporal Graph/Network, yielding a set of historical relationships that imply different versions of each other.  For example:


V1: Person Jane Doe is related as a Product Owner to Product XYZ (started existence 1/1/1992)
V2: Person Jane Doe is related as a Product Owner to Product XYZ (stopped existence 12/31/2000)
V3: Person Jane Doe is related as a Product Owner to Product XYZ (started existence 1/1/2016)

We currently achieve the above through versioned Graph Snapshots that provide views of the whole temporal Graph at specific instances in time.  I’m researching if there are more things that the whole Semantic Relationship can point to.

My Best,

Frank
--
Frank Guerino, Chairman
The International Foundation for Information Technology (IF4IT)
http://www.if4it.com
1.908.294.5191 (M)

From: Matthew West <dr.matt...@gmail.com>
Reply-To: Ontolog Forum <ontolo...@googlegroups.com>
Date: Thursday, April 21, 2016 at 4:50 PM
To: Ontolog Forum <ontolo...@googlegroups.com>
Subject: Re: [ontolog-forum] Are Relationship Triples also Nodes?

Christopher Menzel

unread,
Apr 22, 2016, 11:36:10 AM4/22/16
to ontolo...@googlegroups.com
Trivial? The analysis of belief contexts, and intensional contexts generally, has arguably been the most fraught topic in logic and the philosophy of language for the past 100+ years. And it is pretty much a certain that they are not easily representable by triples. At a minimum, the object of “believes" has to be some sort of first-class individual in its own right (as I think Jack is suggesting).

-chris

Kingsley Idehen

unread,
Apr 22, 2016, 11:51:00 AM4/22/16
to ontolo...@googlegroups.com
On 4/22/16 10:16 AM, Frank Guerino wrote:
Hi Kingsley,

Thanks for the pointer to OSDE.  We use a different approach.  We have developed a data compiler (called NOUNZ) that, among many other things, harvests Semantic Relationships from Nodes, which exist as real data records within a specific context or domain.  For example, in the sample links I provided, the domain is “an Enterprise” and the Nodes are the data types and records found in common enterprises, where the Relationships are those found between such context-specific records.

I took a look at OSDI and found it very interesting.  My limited assessment leads me to believe that difference between OSDI and NOUNZ is the approach used to arrive at an ontology specific model that is loaded with real data.

OSDE (the Editor) simply presents an interface for creating, editing, and saving sentences. The assumption here are:

[1] a sentence represents a datum while a collection of statements (a paragraph) represents data
[2] a sentence is made up of subject, predicate, and object or entity, attribute, value parts.

That's it.


  • In OSDE, it appears that we create named graphs from individual documents right in and from the editor, which leads to the ontology specific model as a living artifact.  This seems to imply a paradigm where we work “one document at a time.”  I found the OSDS OSDE integration video, which used the single web page “CIO Summit Survey” as an example, to be quite telling.  It shows the extraction and creation of relationships being done manually, one at a time, which could take forever when dealing with many web pages.

It assumes that sentences are grouped by pages, as they are in our everyday life, in regards to using language sentences to persist the descriptions of observed phenomena.

As for size of documents, they can be quite large, as per my DBpedia examples [1][2] .


  • In the case of NOUNZ, we use large quantities of structured data from multiple sources (e.g. Systems/Applications) as inputs, parse that data, map it to ontological elements, and then yield an output that contains many named graphs (with many overlaps between them), all which are part of an ontology specific model that represents a domain.  This allows us to deal with millions of different data records of different data types, which may have many tens of millions of relationships, and reverse engineer the results directly from the data.  Also, the goal is to ensure that the outputs are in human and computer consumable forms because most people, for example, don’t speak RDF, WOL, SKOS, Turtle, etc.

Most people speak in sentences which are made up of a subject, predicate, and objects. That's been the case for eons. Unfortunately, we have this thing called "modern computing" that's rife with bizarre marketing-oriented buzzwords that are simply messing up everything and actually undoing eons of progress, exponentially.

I have ideas about ways to combine the two paradigms.


For more options, visit https://groups.google.com/d/optout.

Frank Guerino

unread,
Apr 22, 2016, 12:52:42 PM4/22/16
to Ontolog Forum
Hi Kingsley,

You wrote:  “OSDE (the Editor) simply presents an interface for creating, editing, and saving sentences. The assumption here are… [1]… and [2]…"

Yes, I understand.  Just as an FYI, in the case of [1], I believe the containment assumptions suggest composition of “Data" at all levels:

  • a pixel is data
  • a character composed of multiple structured pixels is data
  • a word composed of multiple structured characters is data
  • a phrase composed of multiple structured words and characters is data
  • a sentence composed of multiple phrases, words and characters is data
  • Etc.
I believe that Data is composed of Data.  Given that all data can be decomposed, I find that it is difficult to point to a singularity or single Datum element because what is one person’s Datum is another person’s Data.  This doesn’t mean I’m right but simply guides the way I work.

You wrote: “Most people speak in sentences which are made up of a subject, predicate, and objects. That's been the case for eons. Unfortunately, we have this thing called "modern computing" that's rife with bizarre marketing-oriented buzzwords that are simply messing up everything and actually undoing eons of progress, exponentially. "

I agree.  In my own work, I always think in terms of human digestible knowledge constructs and work backwards, later retrofitting for computers.  I always, however, try to pre-apply constraints that computers expect as a means of how to structure and represent those knowledge constructs so that it becomes easier to translate for computers, later.

You wrote: “Paradigms aren't the point of difference here per se., the tools in question serve different purposes :) “ and provided links…

Maybe I’m wrong but I think the tools attempt to serve very similar purposes, which is to deconstruct data into one or more Graphs and then use said Graphs to facilitate better knowledge management.  What paths each tool has taken to achieve such results and their respective maturity levels are certainly different and can result in some very interesting conversations.

The last link, in particular, is very cool.  The views of data lead me to a question: What are you doing to make the outputs Human-Readable/Digestible so that Jane Doe, who sits in and is brilliant at Finance but who has no technical experience can traverse, explore, digest, translate and get to understanding from the outputs, easily?  The outputs are not exactly user friendly.

Kingsley Idehen

unread,
Apr 22, 2016, 2:36:30 PM4/22/16
to ontolo...@googlegroups.com
On 4/22/16 12:52 PM, Frank Guerino wrote:
Hi Kingsley,

You wrote:  “OSDE (the Editor) simply presents an interface for creating, editing, and saving sentences. The assumption here are… [1]… and [2]…"

Yes, I understand.  Just as an FYI, in the case of [1], I believe the containment assumptions suggest composition of “Data" at all levels:

  • a pixel is data
  • a character composed of multiple structured pixels is data
  • a word composed of multiple structured characters is data
  • a phrase composed of multiple structured words and characters is data
  • a sentence composed of multiple phrases, words and characters is data
  • Etc.
I believe that Data is composed of Data.  Given that all data can be decomposed, I find that it is difficult to point to a singularity or single Datum element because what is one person’s Datum is another person’s Data.  This doesn’t mean I’m right but simply guides the way I work.

I have a presentation titled "Understanding Data" one of its goals is to get away from tautology like "Data is composed of Data" [1]. Fundamentally, there is a tendency to forget the importance of language (system of signs, syntax, and semantics for encoding and decoding information [data in some context]) in regards to data definition, manipulation, integration, flow etc...



You wrote: “Most people speak in sentences which are made up of a subject, predicate, and objects. That's been the case for eons. Unfortunately, we have this thing called "modern computing" that's rife with bizarre marketing-oriented buzzwords that are simply messing up everything and actually undoing eons of progress, exponentially. "

I agree.  In my own work, I always think in terms of human digestible knowledge constructs and work backwards, later retrofitting for computers.  I always, however, try to pre-apply constraints that computers expect as a means of how to structure and represent those knowledge constructs so that it becomes easier to translate for computers, later.

You wrote: “Paradigms aren't the point of difference here per se., the tools in question serve different purposes :) “ and provided links…

Maybe I’m wrong but I think the tools attempt to serve very similar purposes, which is to deconstruct data into one or more Graphs and then use said Graphs to facilitate better knowledge management. 

OSDE isn't deconstructing data, it is presenting data if loaded form structured data source. Otherwise, its enabling you express observations in reusable form via entity relationship.
What paths each tool has taken to achieve such results and their respective maturity levels are certainly different and can result in some very interesting conversations.

The last link, in particular, is very cool.  The views of data lead me to a question: What are you doing to make the outputs Human-Readable/Digestible so that Jane Doe, who sits in and is brilliant at Finance but who has no technical experience can traverse, explore, digest, translate and get to understanding from the outputs, easily?  The outputs are not exactly user friendly.

The UI is a work in progress. Our goal is to leverage the fact that Jane Doe already understands that things exist, in relation to other things, in a variety of ways via interfaces that look like "property / data sheets".  Ultimately, Jane Doe would really just like to write sentences (no specifically English) that computers understand en route to being more productive :)

Links:

[1] http://www.slideshare.net/kidehen/understanding-29894555 -- Understanding Data .

Kingsley
--
All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontolog Forum content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
---
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 post to this group, send email to ontolo...@googlegroups.com.
Visit this group at https://groups.google.com/group/ontolog-forum.

For more options, visit https://groups.google.com/d/optout.

Ravi Sharma

unread,
Apr 22, 2016, 4:50:26 PM4/22/16
to ontolog-forum
Valuable examples, thanks

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/571A6F26.2020609%40openlinksw.com.

For more options, visit https://groups.google.com/d/optout.



--
Thanks.
Ravi
(Dr. Ravi Sharma)

Frank Guerino

unread,
Apr 22, 2016, 6:49:59 PM4/22/16
to Ontolog Forum
Hi Kingsley,

You wrote: “I have a presentation titled "Understanding Data" one of its goals is to get away from tautology like "Data is composed of Data" [1].”

If you’re up for sharing the presentation, I’d be happy to read it.

You wrote: “Fundamentally, there is a tendency to forget the importance of language (system of signs, syntax, and semantics for encoding and decoding information [data in some context]) in regards to data definition, manipulation, integration, flow etc…”

I don’t know if I agree with this statement only because I’ve found that looking at recursive decomposition keeps us exploring questions and concerns related to the human or machine value of such decompositions.  Furthermore, I believe that context dictates whether or not data is looked at as datum or not.  For example, an entire visual picture is datum to the human eye because the picture stands as a snapshot (composed of many subsets of data).  However, to visualization testing software, it is a massive set of color encoded pixels that act as a reference set or an actual generated set, used for comparison and validation of visualization algorithms.

You wrote: “OSDE isn't deconstructing data, it is presenting data if loaded form structured data source. Otherwise, its enabling you express observations in reusable form via entity relationship.”

Maybe we’re seeing things differently but I do see OSDE as a deconstruction of data.  You’re doing it manually with the editor but you are, in fact, breaking data down into objects, relationships, types, etc.

You wrote: “The UI is a work in progress. Our goal is to leverage the fact that Jane Doe already understands that things exist, in relation to other things, in a variety of ways via interfaces that look like "property / data sheets".  Ultimately, Jane Doe would really just like to write sentences (no specifically English) that computers understand en route to being more productive :) ”

I suggest you spend some time with the Jane Does of the world.  I believe you’ll find they want to do far more than just write sentences.  For example, she may want to explore and learn… a case where she does not know that things exist in relation to each other but learns from her exploration. Also, among many other examples, there is the case where she wants to apply rules in order to get to answers for her questions.

Question: How can I check how many committers there are for the OSDE project on GitHub?

Pat Hayes

unread,
Apr 23, 2016, 1:41:10 AM4/23/16
to ontolog-forum, Christopher Menzel
+1, Chris. 

And by the way, y'all need to be careful about the distinction between "X causes Y" and "that X causes Y". The first is (the form of) a statement, which would have a truth-value. The second is (the form of) a way to refer to a proposition. Trying to reduce this to 'triples' is way too crude and simplistic. 

Pat Hayes

--
All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontolog Forum content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
---
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 post to this group, send email to ontolo...@googlegroups.com.
Visit this group at https://groups.google.com/group/ontolog-forum.

For more options, visit https://groups.google.com/d/optout.

------------------------------------------------------------
IHMC                                     (850)434 8903 home
40 South Alcaniz St.            (850)202 4416   office
Pensacola                            (850)202 4440   fax
FL 32502                              (850)291 0667   mobile (preferred)





Melvin Carvalho

unread,
Apr 23, 2016, 10:04:05 AM4/23/16
to ontolo...@googlegroups.com
This thread may also be of interest.  "Why do we name nodes and not edges".

https://lists.w3.org/Archives/Public/semantic-web/2012Jul/0156.html

In a graph there is a distinction between a node and an edge.  A derived node from an edge can be used in a new graph.  You would need a technique to do that.  I think it would be useful for some (particularly economic) use cases.
 

My Best,

Frank
--
Frank Guerino, Chairman
The International Foundation for Information Technology (IF4IT)
http://www.if4it.com
1.908.294.5191 (M)

--
All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontolog Forum content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
---
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 post to this group, send email to ontolo...@googlegroups.com.
Visit this group at https://groups.google.com/group/ontolog-forum.

Simon Spero

unread,
Apr 23, 2016, 12:11:43 PM4/23/16
to ontolog-forum

On Apr 22, 2016 11:36 AM, "Christopher Menzel" <chris....@gmail.com> wrote:
> [that the analysis of belief contexts is not trivial]

(Assuming  propositional attitudes & propositions) are there any recent developments which suggest plausible representations?

Have you returned to this since Menzel (1993) and IKL (2006)?

Also, ex RDF Named Graphs quodlibet.

Simon

IKL (2006).  IKL Specification Document. Available at http://www.ihmc.us/users/phayes/IKL/SPEC/SPEC.html

Menzel, Christopher (1993). The Proper Treatment of Predication in Fine-grained Intensional Logic, in Philosophical Perspectives, 7, Language and Logic, pp. 61–87, Ridgeview Publishing Company, Atascadero, CA

Christopher Menzel

unread,
Apr 23, 2016, 12:45:09 PM4/23/16
to ontolo...@googlegroups.com
On 23 Apr 2016, at 11:11 AM, Simon Spero <sesu...@gmail.com> wrote:
On Apr 22, 2016 11:36 AM, "Christopher Menzel" <chris....@gmail.com> wrote:

> [that the analysis of belief contexts is not trivial]

(Assuming  propositional attitudes & propositions) are there any recent developments which suggest plausible representations?

Have you returned to this since Menzel (1993) and IKL (2006)?


Not with respect to knowledge engineering. My research interests returned to logic, philosophy of mathematics, and metaphysics around 2006.

-chris

Frank Guerino

unread,
Apr 23, 2016, 1:58:16 PM4/23/16
to Ontolog Forum
Hi Melvin,

You wrote: “This thread may also be of interest.  "Why do we name nodes and not edges”. https://lists.w3.org/Archives/Public/semantic-web/2012Jul/0156.html … In a graph there is a distinction between a node and an edge.  A derived node from an edge can be used in a new graph.  You would need a technique to do that.  I think it would be useful for some (particularly economic) use cases.”

We do, in fact, already name Edges.  Here, for example, is a Catalog of all Named Edges.  (Note: In our case, we call them Predicates or Relationship Descriptors.)

Your statement that “a derived node from an edge can be used in a new graph” is correct.  Naming Edges between Nodes absolutely creates a new form of a graph within a graph that allows you to, from any one Edge, to have direct connections to all other Edges with the same name.  So, if we have the Semantic Relationship:

Person Jane Doe is a “Subject Matter Expert” for Product XYZ

The named edge is “Subject Matter Expert” and it allows us to quickly access all other edges in the super-Graph with the same Name… as in the Use Case: “Find me all Subject Matter Experts, regardless of what they’re experts in”. 

In other words… we treat these Named Edges as if they are their own Resources, just like Named Nodes.

Doing so, however, creates a new phenomena that identifies two distinct types of Edges…

  1. Named Edges that exist between two Named Nodes and
  2. non-Named Edges that exist between any two Named Edges

I hope this makes sense.

My Best,

Frank
--
Frank Guerino, Chairman
The International Foundation for Information Technology (IF4IT)
http://www.if4it.com
1.908.294.5191 (M)



From: Melvin Carvalho <melvinc...@gmail.com>
Reply-To: Ontolog Forum <ontolo...@googlegroups.com>
Date: Saturday, April 23, 2016 at 10:03 AM
To: Ontolog Forum <ontolo...@googlegroups.com>
Subject: Re: [ontolog-forum] Are Relationship Triples also Nodes?

Doug McDavid

unread,
Apr 23, 2016, 2:11:52 PM4/23/16
to ontolo...@googlegroups.com
This conversation makes me wonder whether any people here have implemented anything with Neo4j or any graph form of NoSQL database software?

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/D3412A82.14687B%25Frank.Guerino%40if4it.com.

For more options, visit https://groups.google.com/d/optout.



--

Frank Guerino

unread,
Apr 23, 2016, 2:36:43 PM4/23/16
to Ontolog Forum
Hi Doug,

We haven’t done so, yet, but it’s on our list.

I’m actually very interested in understanding Neo4j but have not taken the time to do so, yet.  We actually create our own Semantic Graph, directly through the compilation of data, based on processing rules.  The results of our graph are used to automatically generate documentation (e.g. JSON files for each element or groups of elements in the Graph) that can be used to feed any NoSQL repository, such as Neo4j.  We’re currently working on integrating our generated documentation into ElasticSearch, which is also a NoSQL repository, for multiple reasons.  For example:

  • facilitating 360 Degree context-based search
  • generating and having temporal history of any Graph available

I’m very curious to understand what putting our generated data into Neo4j would buy us.  If you know anything about the topic, I’d be very happy to discuss it with you, either on-Forum or off-Forum.

My Best,

Frank
--
Frank Guerino, Chairman
The International Foundation for Information Technology (IF4IT)
http://www.if4it.com
1.908.294.5191 (M)



HansTeijgeler

unread,
Apr 23, 2016, 3:50:19 PM4/23/16
to ontolo...@googlegroups.com

Hi Frank,

Let me explain how we model Relationships and relations in ISO 15926.

In ISO 15926 we have the entity types Relationship, ClassOfRelationship, and even ClassOfClassOfRelationship, each with subtypes.

These have two attributes | properties | relations (in EXPRESS | RDF | ISO 15926 )

Object1 << relation1 << Relationship >> relation2 >> Object2

A Relationship represents a fact (observed, asserted).
A Relationship can be classified (typed) with a ClassOfRelationship, that has the same structure, and that defines a requirement or rule.
For example:

CENTRIFUGAL-PUMP << hasClassOfWhole << ClassOfAssemblyOfIndividual_439839 >> hasClassOfPart >> IMPELLER
myPump << hasWhole << AssemblyOfIndividual_238743 >> hasPart >> myImpeller
AssemblyOfIndividual_238743 rdf:type ClassOfAssemblyOfIndividual_439839 .

The first line tells that centrifugal pumps shall have one or more impellers (the cardinalities are defined separately).
The second line tells that my individual pump has an individual impeller
The third line tells that the real-world situation in the second line complies with the rule or requirement defined in the first line

Your Marriage example is a bit more complex but very well doable.
First of all it shall be noted that ISO 15926 works with perdurants, "temporal partz" of individuals.

Here a temporal part of John Doe is related through a marriage relationship with a temporal part of Mary Smith.
The Activity of Marrying causes a marriage Event, which marks the Beginning (temporal bounding) of the applicable temporal parts of John and Mary.
In the case of a divorce:
The Activity of Divorcing causes a divorce Event, which marks the Ending of those temporal parts of John and Mary.

In case they re-marry (which sometimes happens) we have another instance of Marrying Activity that causes another marriage Event, which marks the beginning of other temporal parts of John and Mary.
All Beginnings and Endings of Individuals, or temporal parts thereof, are dateTime-stamped.

Forementioned marriage relationship can be classified (typed) with a class of relationship such as "Hetero Marriage relationship", 

Finally this: those Activities are taking place at a Location and at a dateTime, Persons and other Individuals participate in those Activities, each in their Role, and other Objects can be involved-by-reference (e.g. a law).
ISO 15926 is rich enough to allow modeling of all this in a consistent manner.

Resuming: above there was a marriage Relationship, a marrying Activity, and a marriage Event. Relationship are not dateTime-stamped, but relate temporal parts that are.

I step off my soapbox :)

Regards,
Hans
--
All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontolog Forum content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
---
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 post to this group, send email to ontolo...@googlegroups.com.
Visit this group at https://groups.google.com/group/ontolog-forum.

Doug McDavid

unread,
Apr 23, 2016, 4:01:11 PM4/23/16
to ontolo...@googlegroups.com
Thanks, Frank, and maybe this is an off-list topic.  There was a period of time when I specialized in a dbms that was developed by IBM Research and ultimately failed to come to market.  The key feature was the lack of any table structure, only subjects in categories connected via bi-directional, named relationships.  Of the several NoSQL entries, one dbms type is the graph database, of which Neo4j is an example.  Neo4j also eschews any form of table representation, opting instead for bi-directional, named, relationships.  I do know some people who have pushed it into the realm of reification of relationships as well, which is why I think it may be pertinent to this discussion.

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/5889214e-a11e-3352-541b-e7a8d79d0676%40quicknet.nl.

For more options, visit https://groups.google.com/d/optout.

John F Sowa

unread,
Apr 23, 2016, 6:22:40 PM4/23/16
to ontolo...@googlegroups.com
Chis, Pat, Melvin, Frank,

CM
> The analysis of belief contexts, and intensional contexts generally, has
> arguably been the most fraught topic in logic and the philosophy of
> language for the past 100+ years. And it is pretty much a certain that
> they are not easily representable by triples. At a minimum, the object
> of “believes" has to be some sort of first-class individual in its own right

PJH
> +1, Chris.
>
> And by the way, y'all need to be careful about the distinction between
> "X causes Y" and "that X causes Y". The first is (the form of) a
> statement, which would have a truth-value. The second is (the form of)
> a way to refer to a proposition. Trying to reduce this to 'triples'
> is way too crude and simplistic.

I completely agree with Chris and Pat. This is also the conclusion
of the IKRIS project that specified the IKL extension to Common Logic.

To illustrate Pat's point, following is a CLIF statement that
expresses the sentence "There exists a sky that is blue":

(exists (x) (and (sky x) (blue x)))

And following is the IKL sentence for the sentence "Sue believes
that there exists a sky that is blue":

(believes Sue (that
(exists (x) (and (sky x) (blue x))) ))

If you have a notation for any version of logic (RDF, OWL, FOL,
or whatever), you need to distinguish an object-level assertion
in the language from a metalevel assertion about an assertion.

The IKL operator named 'that' makes the distinction. Any IKL
sentence that uses the operator 'that' can be translated to an
English sentence with the IKL 'that' replaced by the English 'that'.

The converse, however, is not always true -- because most English
sentences don't have a well-defined mapping to any version of logic.

For more about the IKRIS project and the IKL extensions to CL,
see http://www.jfsowa.com/ikl/#s2

MC
> Do you believe that a Relationship Triple can also be treated
> as a Node and why?

An RDF triple asserts a proposition. In order to make an assertion
about a triple, you need to make the same distinction as IKL:

1. Wrap the triple (or set of triples) in parentheses (or some
other kind of enclosure).

2. Then put the word 'that' or some equivalent symbol at the
beginning (or the outside) of the wrapper for the triple
or set of triples.

3. As in the example above, the assertions (triples) inside the
wrapper are the object level, and the assertion about the
enclosed triple(s) is the metalevel.

But you need to distinguish metalanguage from nominalization.
For example, the noun 'marriage' is a nominalized form of the
verb 'marry'. You could translate both of the following
sentences

"Sue married Bob" and "There is a marriage of Sue and Bob"

to the same sentence in CLIF:

(married Sue Bob)

You don't need the IKL 'that' operator for this purpose.

John

Obrst, Leo J.

unread,
Apr 23, 2016, 9:14:06 PM4/23/16
to ontolo...@googlegroups.com, Christopher Menzel

Do you mean that “X causes Y” is a proposition (and so has a truth value), and that “that X causes Y” refers to a proposition (and so you don’t know its truth value)?

 

Thanks,

Leo

Obrst, Leo J.

unread,
Apr 23, 2016, 9:39:36 PM4/23/16
to ontolo...@googlegroups.com, Christopher Menzel

The usual linguistic notion is that you don’t know the truth value of an embedded statement, but that the verbs and other locutions of the main statement containing the embedded statement gives you some indication on how you should treat the embedded statement. Linguistically, verbs like “know”, “assert”, “regret”, etc., as opposed to “believe”, “think”, etc., allow the truth value of the embedded statement to “leak” out to the encompassing statement, i.e., as a kind of presuppostion. This is typically considered pragmatic construal, and analyses go back to (at least) Kartunnen (1973). See: http://plato.stanford.edu/entries/presupposition/. This is an intensional context, as Chris and Pat indicate.  Propositional attitudes are notoriously hard: http://plato.stanford.edu/entries/prop-attitude-reports/. In AI, folks such as Fagin et al [1] have attempted to unpack some of this.

 

Thanks,

Leo

 

[1] Fagin, Ronald; Halpern, Joseph Y.; Moses, Yoram; Vardi, Moshe Y. 1996.  Reasoning about Knowledge, MIT Press.

 

From: Obrst, Leo J.
Sent: Saturday, April 23, 2016 9:14 PM
To: ontolog-forum <ontolo...@googlegroups.com>; Christopher Menzel <chris....@gmail.com>
Subject: RE: [ontolog-forum] Are Relationship Triples also Nodes?

 

Do you mean that “X causes Y” is a proposition (and so has a truth value), and that “that X causes Y” refers to a proposition (and so you don’t know its truth value)?

 

Thanks,

Leo

 


Sent: Saturday, April 23, 2016 1:40 AM
To: ontolog-forum <ontolo...@googlegroups.com>; Christopher Menzel <chris....@gmail.com>

Christopher Menzel

unread,
Apr 23, 2016, 11:42:41 PM4/23/16
to Obrst, Leo J., ontolo...@googlegroups.com
I think Pat's main point was simply that the two belong to entirely distinct grammatical categories. "X causes Y" is a sentence, the sort of thing that can be true or false, whereas "that X causes Y" is a singular term, the sort of thing that refers to an entity of one sort or another. "That"-clauses, of course, are usually taken to refer to propositions (whatever you take propositions to be).

-c

John F Sowa

unread,
Apr 23, 2016, 11:58:18 PM4/23/16
to ontolo...@googlegroups.com
On 4/23/2016 9:13 PM, Obrst, Leo J. wrote:
> Do you mean that “X causes Y” is a proposition (and so has a truth
> value), and that “that X causes Y” refers to a proposition (and so
> you don’t know its truth value)?

Note the examples in my previous note:
> To illustrate Pat's point, following is a CLIF statement that
> expresses the sentence "There exists a sky that is blue":
>
> (exists (x) (and (sky x) (blue x)))
>
> And following is the IKL sentence for the sentence "Sue believes
> that there exists a sky that is blue":
>
> (believes Sue (that
> (exists (x) (and (sky x) (blue x))) ))

Since IKL is an extension of Common Logic, every CL sentence is
also an IKL sentence. If you have an IKL model M for those two
sentences, it will determine a truth value for each of them.

But in the second sentence, the following expression denotes
an entity of type Proposition:

(that (exists (x) (and (sky x) (blue x))))

A proposition may have a truth value, but a proposition is *not*
a truth value. In IKL, it is possible to write

(= p (that (exists (x) (and (sky x) (blue x)))))

This sentence says that p is a name for the proposition;
p is not a name for a truth value. If you write

(believes Sue p)

That says that Sue believes the proposition p. It does *not*
say that Sue believes a truth value. Sue may believe p even
if the truth value of p happens to be F.

> The usual linguistic notion is that you don’t know the truth value
> of an embedded statement, but that the verbs and other locutions
> of the main statement containing the embedded statement gives you
> some indication on how you should treat the embedded statement.

What you know or don't know is irrelevant to the IKL model theory.
The choice of verb (or relation in CL or IKL) is also irrelevant.

As another example, consider the sentence "2+2=4" and Fermat's
last theorem. The first one is easy to prove, but the second
was in doubt for a long time. But we now know that both of
them are true. Those two sentences happen to have the same
truth value, but they express very different propositions.

For more about propositions, see
http://www.jfsowa.com/logic/proposit.pdf

John


Pat Hayes

unread,
Apr 23, 2016, 11:58:46 PM4/23/16
to ontolog-forum, Simon Spero
Simon, 

new readers might find 


a more readable introduction. See especially http://www.ihmc.us/users/phayes/IKL/GUIDE/GUIDE.html#propositionNames

Pat

--
All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontolog Forum content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
---
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 post to this group, send email to ontolo...@googlegroups.com.
Visit this group at https://groups.google.com/group/ontolog-forum.

For more options, visit https://groups.google.com/d/optout.

Pat Hayes

unread,
Apr 24, 2016, 12:09:09 AM4/24/16
to ontolog-forum, Obrst, Leo J., Christopher Menzel
On Apr 23, 2016, at 8:13 PM, Obrst, Leo J. <lob...@mitre.org> wrote:

Do you mean that “X causes Y” is a proposition (and so has a truth value)

To avoid confusion arising from the X and Y, lets use a concrete example. "Leo is human" is a sentence - three words in sequence, a linguistic entity - which has a truth-value; in this case, True. Sentences are true or false; technically, they denote truth-values. In the conventional semantics of assertional logics, propositions do not play any role, but if you wanted to put them in there, then the sentence would *assert* that the corresponding proposition is true. 

, and that “that X causes Y” refers to a proposition (and so you don’t know its truth value)?

"That Leo is human" refers to a proposition, yes. In IKL it would be a proposition *name* which simply refers to the proposition without asserting or denying it. 

You may or may not know the truth-value of this proposition, but that has nothing to do with the distinction between a proposition name and a sentence. 

Pat


For more options, visit https://groups.google.com/d/optout.

Pat Hayes

unread,
Apr 24, 2016, 12:15:14 AM4/24/16
to ontolog-forum, Christopher Menzel, Obrst, Leo J.
On Apr 23, 2016, at 10:42 PM, Christopher Menzel <chris....@gmail.com> wrote:

I think Pat's main point was simply that the two belong to entirely distinct grammatical categories.

Yes. Which is a good sign that if you get these categories confused or accidentally merged, your analysis is going to get muddled or at best needlessly complicated. 

Pat


For more options, visit https://groups.google.com/d/optout.

Pat Hayes

unread,
Apr 24, 2016, 12:18:26 AM4/24/16
to ontolog-forum, Obrst, Leo J., Christopher Menzel
On Apr 23, 2016, at 8:39 PM, Obrst, Leo J. <lob...@mitre.org> wrote:

The usual linguistic notion is that you don’t know the truth value of an embedded statement,

?? "Bill doesn't believe that all mice are mammals" might be true or false, but I am sure the embedded sentence is true. 

If you mean that asserting the larger sentence does not in itself assert the embedded sentence, then yes, I agree. 

Pat


For more options, visit https://groups.google.com/d/optout.

Frank Guerino

unread,
Apr 24, 2016, 12:28:11 AM4/24/16
to Ontolog Forum
Hi Hans,

Thanks for taking the time to break this down.

I have a question.  You state that “A Relationship represents a fact (observed, asserted).”

Is such a statement accurate?  For example, for many years in our human history, people related “Earth” to “Flat” and related the “Sun” to a “Circular Orbit” and Revolves Around “Earth.”  It would therefore appear that not all relationships are facts.

You wrote: “Resuming: above there was a marriage Relationship, a marrying Activity, and a marriage Event. Relationship are not dateTime-stamped, but relate temporal parts that are.”

I agree with your statement that Relationships are not dateTime-stamped but I would view this a bit differently…

  • There are many Relationships…
    • Mary Smith is Married to John Doe
    • John Doe is Married to Mary Smith
    • Mary Smith is Spouse Of John Doe
    • John Doe is Spouse of Mary Smith
    • Mary Smith is Wife of John Doe
    • John Doe is Husband of Mary Smith
    • Mary Smith Married John Doe
    • John Doe Married Mary Smith
    • Etc.
  • There is a Marriage Node with Traits
    • Spouses = Mary Smith; John Doe (pointer to People/Person Nodes)
    • Start Date = 1/1/1999 (pointer to Date Node)
    • End Date = TBD (pointer to Date Node)
    • Facility = Facility XYZ (pointer to Facility Node)
  • There are many involved Nodes: See all nodes that the Marriage Node involves in its existence

In the above, Marriages are nodes in the Graph that involve (relate to) many other Nodes.  I hope this helps clarity.

By the way, could you please point me to common use systems that leverage EXPRESS | RDF | ISO 15926?  I’m very familiar with many of the most common systems used in most enterprises throughout all areas of business (e.g. Marketing, Sales, ERP, Finance, Support, etc.) but I’m not familiar with a single one that espouses such standards.

My Best,

Frank
--
Frank Guerino, Chairman
The International Foundation for Information Technology (IF4IT)
http://www.if4it.com
1.908.294.5191 (M)



Frank Guerino

unread,
Apr 24, 2016, 12:42:15 AM4/24/16
to Ontolog Forum
Hi Doug,

Your question about use of Neo4j caused me to spend some time exploring the technology, today.  I viewed multiple videos and read up on uses.  After doing so, my initial reaction to the technology is that we would have to:

  1. Manually create graphs within Neo4j, which appears to be very time consuming, or
  2. Automatically convert/integrate/migrate our own NOUNZ-generated graphs into Neo4j, which could be done with NOUNZ-JSON to Neo4j converters/modules.

This leads me to the question… Given that we already create the graph and that we already create views on top of the data, what would be the benefits of doing so?  My only answer to this is that such a conversion would allow the end user to leverage the Cypher Query Language to create queries of the data that we do not already perform and surface for them.  This implies that the end users would have to be Cypher-savvy… further implying a very small community of end users.

Do you believe my assessments to be incorrect and, if so, why (as I’d definitely want to understand any incorrect observations)?

My Best,

Frank
--
Frank Guerino, Chairman
The International Foundation for Information Technology (IF4IT)
http://www.if4it.com
1.908.294.5191 (M)



Frank Guerino

unread,
Apr 24, 2016, 12:45:49 AM4/24/16
to Ontolog Forum
Hi John,

Thanks for the contributions.  For my own education, could you please point me to common use systems that leverage CLIF?  As mentioned to Hans, I have pretty broad experience with many enterprise systems and have yet to come across any that leverage such technologies and standards mentioned in this thread.

My Best,

Frank
--
Frank Guerino, Chairman
The International Foundation for Information Technology (IF4IT)
http://www.if4it.com
1.908.294.5191 (M)

--
All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontolog Forum content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
---
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 post to this group, send email to ontolo...@googlegroups.com.

Frank Guerino

unread,
Apr 24, 2016, 12:49:20 AM4/24/16
to Ontolog Forum, Christopher Menzel
Hi Leo,

Question:  Given your response, how would you/we model the scenario where many people, for many years, believed (many even “knew”) the earth was flat, only to find out later that this was not the case?

My Best,

Frank
--
Frank Guerino, Chairman
The International Foundation for Information Technology (IF4IT)
http://www.if4it.com
1.908.294.5191 (M)

From: "Obrst, Leo J." <lob...@mitre.org>

HansTeijgeler

unread,
Apr 24, 2016, 1:53:30 PM4/24/16
to ontolo...@googlegroups.com

Hi Frank,

[FG] You state that “A Relationship represents a fact (observed, asserted).” Is such a statement accurate?  For example, for many years in our human history, people related “Earth” to “Flat” and related the “Sun” to a “Circular Orbit” and Revolves Around “Earth.”  It would therefore appear that not all relationships are facts.

[HT] A temporal part of the Earth was, in those days, asserted to be flat. That temporal part was ended at the time it was proven to be spherical, and a new temporal part began to be the placeholder of that information.

[FG] There are many Relationships…

      • Mary Smith is Married to John Doe
      • John Doe is Married to Mary Smith
      • Mary Smith is Spouse Of John Doe
      • etc
    [HT] In our lingo those would be called relations, or attributes, or rdf:Property. And as you know much better than me, those cannot be a subject nor object of another Property, unless you apply that afterthought reification. Relationships to us have two such properties, e.g. part and whole, and can be addressed.
    A case where a Relationship is the object of another Relationship is Approval of which the EXPRESS code is:
    ENTITY approval
        SUBTYPE OF (relationship);
        approved     :  relationship;
        approver     :  possible_individual;
    END_ENTITY;
    Attribute definitions
        approved     :  The Relationship that is approved in the Approval.
        approver     :  The PossibleIndividual that is the approver in the Approval.

    And this Approval can be classified (typed) with a ClassOfApproval:
    ENTITY class_of_approval
        SUBTYPE OF (class_of_relationship);
        class_of_approved  :    class_of_relationship;
        class_of_approver   :   class_of_individual;
    END_ENTITY;
    Attribute definitions
        class_of_approved  :   The ClassOfRelationship whose members are approved by the members of the class_of_approver.
        class_of_approver   :   The ClassOfIndividual whose members are the approvers of the ClassOfRelationship approved.

    [FG] By the way, could you please point me to common use systems that leverage EXPRESS | RDF | ISO 15926?  I’m very familiar with many of the most common systems used in most enterprises throughout all areas of business (e.g. Marketing, Sales, ERP, Finance, Support, etc.) but I’m not familiar with a single one that espouses such standards.

    [HT] ISO 15926 evolved from ISO10303-221 (STEP), and as a consequence the foundation model in ISO 15926-2 is defined in EXPRESS. ISO 15926 has as scope the representation of lifecycle information of facilities, ranging from myCar to a mega chemical plant, so all information from cradle to grave, often over a period of decades. ISO 10303 and EXPRESS appeared not to be the proper environment for that, so in the mid nineties we decided to start ISO 15926, in fact defining what the STEP folks called an Implementable ARM (Application Resource Model). In 2000 we decided to go for Semantic Web technologies as they developed in those days. We have a small set of Upper Ontologies defined in OWL:
    • ISO 15926-2 foundation model
    • ISO 15926-3 geometry and topology
    • ISO 15926-7 templates
    Instances thereof are defined in RDF, stored in quad stores, and queried with SPARQL.
    Actually our Upper Ontologies are built on top of OWL. Next to that we have the RDL (Reference Data Library) on the web (in RDF), and extensions thereof for standard object classes (e.g. ASTM, ANSI, BS, DIN, TEMA, etc) and user-defined classes (mainly for commercial products).

    Those templates are an answer to the unwieldy graphs that ISO 15926 produces because of its normalized nature with 201 entity types.
    In templates we write down the variables of a graph, but at the same time we define the semantics of the full graph in FOL.
    We were inspired by "Defining N-ary Relations on the Semantic Web" by Natasha Noy and Alan Rector.

    There are some 180 standard templates, each of them standardizes a relatively small and autonomous information chunk, like the template "IndividualHasPropertyWithValue" that has the following graph:


    and the following "signature":

    Role No

    Role Name

    Role Object Type

    1

    hasPropertyPossessor

    dm:PossibleIndividual

    2

    hasPropertyType

    dm:ClassOfProperty

    3

    valPropertyValue

    dm:ExpressReal

    4

    hasScale

    dm:Scale

    The FOL code for this template type is:

    IndividualHasPropertyWithValue(x1, x2, x3, x4) <->
    PossibleIndividual(x1) &
    ClassOfProperty(x2) &
    ExpressReal(x3) &
    Scale(x4) &
    exists u1 exists u2 exists u3 exists u4 exists u5 exists u6(
          PossibleIndividual(u1) &
          TemporalWholePartTemplate(u1, x1) &
          Property(u2) &
          ClassificationTemplate(u2, u1) &
          ClassificationTemplate(u2, x2) &
          PropertyQuantificationTriple(u3, u2, u4) &
          ClassificationTemplate(u3, x4) &
          RealNumber(u4) &
          ClassOfIdentificationTemplate(x3, u4) &
          PointInTime(u5) &
          BeginningTemplate(u5, u1) &
          ClassOfIdentificationTemplate(u6, u5) &
          RepresentationOfGregorianDateAndUtcTime(u6)) .

    An instance of this template could be: "Pump P-101 has a mass of 135.7 kg", which would result in something like:

    :T40D60747B02A40359BD43B7A96E845A5 rdf:type tpl:IndividualHasPropertyWithValue ;
        tpl:hasPropertyPossessor :TBC648553B09D4F88A89DE655D34B8521 ; # the GUID of P-101
        tpl:hasPropertyType rdl:RDS353339 ; # MASS
        tpl:valPropertyValue "135.7"^^xsd:decimal ;
        tpl:hasScale rdl:RDS1328669 ; # KILOGRAM
        meta:valEffectiveDate "2016-04-24T18:14:23Z"^^xsd:dateTime .

    Regards,
    Hans
    --
    All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontolog Forum content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
    ---
    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 post to this group, send email to ontolo...@googlegroups.com.
    Visit this group at https://groups.google.com/group/ontolog-forum.

    Obrst, Leo J.

    unread,
    Apr 24, 2016, 2:13:56 PM4/24/16
    to ontolo...@googlegroups.com, Christopher Menzel

    Ok, sure, I agree.

    Obrst, Leo J.

    unread,
    Apr 24, 2016, 2:17:57 PM4/24/16
    to ontolo...@googlegroups.com, Christopher Menzel

    Frank,

     

    If you are realist, then you will say that reality didn’t change, just our apprehension and understanding of it. Realists consider reality existing independent of human observers of it. The task of science is largely to increase our understanding of physical reality. In addition, with better instruments (microscopes, telescopes, etc.) we extend our apprehension of it.

     

    A good ontology should correspond to our best understanding of reality.

     

    Thanks,

    Leo

    Obrst, Leo J.

    unread,
    Apr 24, 2016, 3:12:31 PM4/24/16
    to ontolo...@googlegroups.com

    I’m redirecting this back to the ontology forum, rather than the ontology summit. Apparently Rob had problems posting to the latter.

     

    Thanks,

    Leo

     

    From: ontolog...@googlegroups.com [mailto:ontolog...@googlegroups.com] On Behalf Of Rich Cooper
    Sent: Sunday, April 24, 2016 2:35 PM
    To: ontolog...@googlegroups.com
    Subject: [Ontology Summit] FW: [ontolog-forum] Are Relationship Triples also Nodes?

     

    Since my google is stuck on rejecting my email, and won't let me post on o-forum, here is my response to a topic going on there.

     

    Leo wrote:>

     

    If you are realist, then you will say that reality didn’t change, just our apprehension and understanding of it. Realists consider reality existing independent of human observers of it. The task of science is largely to increase our understanding of physical reality. In addition, with better instruments (microscopes, telescopes, etc.) we extend our apprehension of it.

     

    A good ontology should correspond to our best understanding of reality.

     

    I suggest this is more the behavior of a physicist per se, than a scientist in general.  Science is so vastly much wider than physics!  Look at the total number of faculty in a physics department compared to the number in the entire science faculty! They all believe, as I do, that they are doing science, and many of them have little concern about physics details, preferring the depth of medicine, biology, psychology, sociology, geology, et ad finitum.  Yes, there is physics at the bottom of the turtle stack everywhere, but physics has no part in the other sciences' daily experiences. 

     

    Computer science, for example, is about the understanding of how machines can be programmed, organized, improved, and used to provide symbolic assistance to human users.  That has little to do with physics, yet lots to do with logic, math, classification, theories and so forth. 

     

    Sincerely,

    Rich Cooper,

    Rich Cooper,

     

    Chief Technology Officer,

    MetaSemantics Corporation

    MetaSemantics AT EnglishLogicKernel DOT com

    ( 9 4 9 ) 5 2 5-5 7 1 2

    http://www.EnglishLogicKernel.com


    For more options, visit https://groups.google.com/d/optout.

    --
    All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontology Summit content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
    ---
    You received this message because you are subscribed to the Google Groups "ontology-summit" group.
    To unsubscribe from this group and stop receiving emails from it, send an email to ontology-summ...@googlegroups.com.
    To post to this group, send email to ontolog...@googlegroups.com.
    Visit this group at https://groups.google.com/group/ontology-summit.
    To view this discussion on the web visit https://groups.google.com/d/msgid/ontology-summit/0c7901d19e58%2405704710%241050d530%24%40com.

    Obrst, Leo J.

    unread,
    Apr 24, 2016, 3:17:34 PM4/24/16
    to ontolo...@googlegroups.com

    Rich,

     

    It’s perhaps clearer with the physical sciences, but even the social and computational sciences use the physical sciences. In any case these others are intended to be based on reality, don’t you think?

    Obrst, Leo J.

    unread,
    Apr 24, 2016, 3:18:22 PM4/24/16
    to ontolo...@googlegroups.com

    Sorry, should be: “Rich”.

    Obrst, Leo J.

    unread,
    Apr 24, 2016, 4:45:12 PM4/24/16
    to ontolog...@googlegroups.com, ontolo...@googlegroups.com

    I think you are interested in data structures. That’s fine, for local notions you have control over. The problem comes when you try to generalize those, i.e., you invent database technology, and then go down that path of many structural model solutions, all different, and then try to apply those to larger user and computational units, i.e., partners, enterprises, communities, etc. However,  the common reality is what you scratch on eventually.

     

    Thanks,

    Leo

     

    From: ontolog...@googlegroups.com [mailto:ontolog...@googlegroups.com] On Behalf Of Rich Cooper
    Sent: Sunday, April 24, 2016 3:53 PM
    To: ontolo...@googlegroups.com
    Cc: ontolog...@googlegroups.com
    Subject: RE: [Ontology Summit] FW: [ontolog-forum] Are Relationship Triples also Nodes?

     

    LO:> Sorry, should be: “Rich”.

    RC:> No problem.  You stimulatingly wrote:

     

    It’s perhaps clearer with the physical sciences, but even the social and computational sciences use the physical sciences. In any case these others are intended to be based on reality, don’t you think?

     

    Thanks,

    Leo

     

    No, that is the problem I run into as a computer scientist.  In order to get the operational capability in software, there are REPRESENTATIONS of the real world, or of the interpretation of physical and nonphysical (information, theories, assertions) about nearly anything real or unreal.  It ranges from the physical world, such as models of quantum world situations for the reality focused, or the world soft sciences, which have only human observations of situations, possibly only social, not reality based. 

     

    Working with representations, as computer science does, has no direct relationship with reality.  That relationship is created in the mind of the programmer, who attempts (and usually fails) to document his thoughts so that other observers can try to understand what THEY think the program does. 

     

    Randy Jensen demonstrated that the software development problem is communication among programmers; the lack of a single vision thereof.  We have yet to demonstrate much agreement among ontologists about what ontology is, so why would we jump to the conclusion that we all experience the same reality?

     

    It's very valuable for programmers and their users to have databases and FOL tools and symbolic representations such as equations to work with, when designing representations.  But what in reality that representation corresponds to is often missing.  It is "supposed to" correspond to the things that people experience, not to the reality they experience it of.  And the said people are diverse, so their interpretations are diverse. 

     

    Aside: Exceptions are in embedded machine systems, such as car computers, where they are mapped to physical switches, sensors and displays in a much more traceable way.  Even then, only the driver and passengers interpret the representations, each in unique ways, which may or may not be what the program designers expected. 

     

    So the world of representations is a very different one than the physical world.   Yet it relates to the sciences (and humanities) in every field that has tried it so far. 

     

    So ontologies of representations are, IMHO, valid or exploratory ontologies are very useful for representations, regardless whether they are directly mapped to anything real.  Instead, they are at the most detailed, mapped to messages that never exit the program. 

     

    I suggest that we go with "ontology is the observation, classification and distinction among objects from a representation." 

     

    Note that an observation is itself a representation. 


    For more options, visit https://groups.google.com/d/optout.

    --
    All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontology Summit content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
    ---
    You received this message because you are subscribed to the Google Groups "ontology-summit" group.
    To unsubscribe from this group and stop receiving emails from it, send an email to ontology-summ...@googlegroups.com.
    To post to this group, send email to ontolog...@googlegroups.com.
    Visit this group at https://groups.google.com/group/ontology-summit.

    Frank Guerino

    unread,
    Apr 24, 2016, 6:09:00 PM4/24/16
    to Ontolog Forum
    Hi Hans,

    In your last response, you stated: "A temporal part of the Earth was, in those days, asserted to be flat.”  However, in your previous response, you stated: “A Relationship represents a fact (observed, asserted).”

    I’m a bit confused.  Are you still stating that a Relationship is a fact?  I understand that all Relationships are assertions but I cannot see how all Relationships are facts.

    In my previous response to you, I asked you the question: “Could you please point me to common use systems that leverage EXPRESS | RDF | ISO 15926?  I’m very familiar with many of the most common systems used in most enterprises throughout all areas of business (e.g. Marketing, Sales, ERP, Finance, Support, etc.) but I’m not familiar with a single one that espouses such standards.”

    Your response was to get into standards, design and syntax.  My question is still open.  Could you please point me to common/mainstream applications that leverage such standards to share data with each other?  I do not see mainstream business applications leveraging such standards, or even planning to do so.  Therefore, I’m trying to understand their existing business uses to solve real business problems so can you point to such examples?

    Frank Guerino

    unread,
    Apr 24, 2016, 6:18:07 PM4/24/16
    to Ontolog Forum, Christopher Menzel
    Hi Leo,

    You wrote: “A good ontology should correspond to our best understanding of reality.”

    A fair enough answer.  I think part of the issue is that we have gaps in understanding what the brain really does in order to handle specific cases and the best we can do, as humans, is identify, implement, and test the available options for handing similar situations, given the tools and technologies available at the time for doing so.

    Frank Guerino

    unread,
    Apr 24, 2016, 6:30:44 PM4/24/16
    to Ontolog Forum
    This is in response to Rich Cooper via Leo Orbst…

    Hi Rich,

    You Wrote: “Computer science, for example, is about the understanding of how machines can be programmed, organized, improved, and used to provide symbolic assistance to human users.  That has little to do with physics, yet lots to do with logic, math, classification, theories and so forth.”

    I would respectfully suggest that machines have significant limitations and, often, the better answer for solving logic problems is to go back and “modify the machine”, which leads to things like “modifying the semiconductors within the machines”, which leads to the science of physics for doing so.

    It is not a magical convenience that, approximately every 18 months, new faster machines with more memory are available to scientists in order to solve problems for similar costs.

    HansTeijgeler

    unread,
    Apr 25, 2016, 9:07:53 AM4/25/16
    to ontolo...@googlegroups.com

    Hi Frank,

    What about an asserted fact? Can you give me an example of a Relationship that is not representing a fact, actual or asserted?

    Yes, an example of a mainstream system using ISO 15926 is Bentley ( https://www.bentley.com/en/products/brands/openplant ) .

    Other companies working on it in different levels: AVEVA, Bechtel, Bentley Systems, Black & Veatch, Consolidated Contractors Company (CCC), CH2MHILL, The Dow Chemical Company, Emerson, Fluor, GlencoIS, Hatch, Intergraph, Noumenon, Siemens, and WorleyParsons.

    Fiatech has the following projects going on:

    Roadmap Element 9, Information Management

    ISO 15926 Information Patterns (IIP) – Reg Hunter

    Development of consensus-based ISO 15926 template information patterns and creation of data layer software for key industry tools

    Development of consensus-based ISO 15926 template information patterns and creation of data layer software for key industry tools A

    Exploring Piping Management through Project Information Flow (PIF) – Ray Topping

    Objective is to enable usage of existing ISO 15926 tools to achieve smooth piping data flow from design to construction and, ultimately, to owner/operator handover. Will also attempt to approximate the cost and schedule savings achieved with implementing the data exchange tools and best practices.

    This project will define the requirements of piping data from design to construction (with handover as the ultimate goal), including addressing piping specifications, geometry standards, material standards, and drafting rules and symbols. Project scope includes developing a neutral file format compatible with ISO 15926 for piping isometric data exchange. R

    Capturing Equipment Data Requirements Using ISO 15926 & Assessing Conformance (EDRC) – Ray Topping

    To establish a common understanding across industry projects of how to use ISO 15926 and how to assess software conformance to specified ISO 15926 data structures and capabilities. Collaboration with MIMOSA on the OGI Pilot is a cornerstone of this project.

    To establish a common foundation for representing, exchanging, and sharing equipment data using ISO 15926 and to demonstrate how to assess software implementations compliance. R

    ISO 15926 Information Models and Proteus Mappings (IIMM) – Ray Topping

    Information models for process and instrumentation diagrams (P&IDs) and 3D models will be documented with ISO 15926 classes and templates. The information models will then be mapped to the Proteus Schema.

    Interoperability between Proteus Schema XML and Part 8 OWL XML deployments will enable transparent access to intelligent design information. This will save major costs and time delays incurred in the current manual, non-interoperable data and handover exchange processes. A

    David Price may give you other examples.

    Regards,

    Hans

    Frank Guerino

    unread,
    Apr 25, 2016, 11:35:13 AM4/25/16
    to Ontolog Forum
    Hi Hans,

    You wrote: “What about an asserted Fact? Can you give me an example of a Relationship that is not representing a fact, actual or asserted?”

    If we have the Semantic Relationship such as:

    Flat is related as a Characteristic of Earth

    The relationship is definitely an assertion (as in “evidential data available”).  It is not, however, a fact (as in “true”).

    I think you’re alluding to a situation where data available are treated as “facts, at hand,” such as in the case when an investigator says, “Let’s deal with the facts we have available.” even though he/she really means, “Let’s deal with the data we have available.”  Is this correct?

    Thanks for the example.  I checked out the Bentley site and their products.  It appears that use of the standard is currently in very narrow/niche areas.  Is this interpretation correct?  Again, I am by no means implying that the standard is not valuable or that it couldn’t be used in other areas.  I’m simply trying to understand its current uses and applications.

    Chris Menzel

    unread,
    Apr 25, 2016, 11:36:15 AM4/25/16
    to ontolog-forum
    In probably my first defense of anything Rich has ever said, his claim was that computer science is about how understanding how machines can be programmed, organized, etc. It is perhaps just a quibble, but the physics of semiconductors is an aspect of computer engineering and it's at least arguable that that is a distinct (albeit of course closely related) discipline.

    -chris


    HansTeijgeler

    unread,
    Apr 25, 2016, 12:11:16 PM4/25/16
    to ontolo...@googlegroups.com

    Hi Frank,

    Whatever, for an engineer it is a fact. If it would not be asserted to be true to him/her, it would not be in his/her computer system.

    In case you want details about the Bentley products please contact Glen....@bentley.com

    Regards,

    Hans

    Matthew West

    unread,
    Apr 25, 2016, 4:05:00 PM4/25/16
    to ontolo...@googlegroups.com

    Dear Hans,

    The fact would be the assertion that a relationship is true.

    Regards

    Matthew

    Christopher Menzel

    unread,
    Apr 25, 2016, 4:14:13 PM4/25/16
    to ontolo...@googlegroups.com
    Asserting that something is true adds nothing to simply asserting it. If “φ is true” is a fact, so is φ. Nor does anything become a fact in virtue of being asserted to be true. If I assert that it is true that the earth is larger than Jupiter, I have not asserted a fact, I’ve simply said something false.

    -chris


    and the following "signature":

    William Frank

    unread,
    Apr 25, 2016, 5:28:46 PM4/25/16
    to ontolog-forum
    Christoper is of course correct that asserting something is true adds nothing to simply asserting except, as he points out, we have a metalanguage indirection.  He says:

    if "φ is true" is a fact, so is φ.  This implies that φ is an assertion.   

    But, we can even in that jumble that is English, say "That polly is a parrot is true".  In this case, if  φ  = 'that polly is a parrot' then φ is NOT an assertion, it is a proposition, which might be true or false  And about which we can say things.  A previous part of this thread seemed to be completely aware of this, and was discussing the difficult problem of dealing with assertions *about* propositions, (I wonder if φ) (I hate it that φ) I doubt that φ , I believe that φ, etc.  Let us imagine φ.  Assume φ,  

    So, when Mathew said

    The fact would be the assertion that a relationship is true.

    I understood this to mean that he was regarding a relationship instance to be a proposition, such as 

    'that John and Sally are married' 

    not an assertion, 

    which you get in English by taking out the 'that', which syntactically implies an assertion, without it being explicit, just an accident of our languauge, but in a fully analytic langauge, by instead adding an explicit assertion morpheme, such as, in English, "I assert" "is true", etc.  



    doug foxvog

    unread,
    Apr 26, 2016, 1:34:55 AM4/26/16
    to ontolo...@googlegroups.com
    On Sun, April 24, 2016 00:28, Frank Guerino wrote:
    > Hi Hans,
    >
    > Thanks for taking the time to break this down.
    >
    > I have a question. You state that "A Relationship represents a fact
    > (observed, asserted)."

    Note that if something is asserted to be a fact, it need not be so.

    > ...
    > You wrote: "Resuming: above there was a marriage Relationship, a marrying
    > Activity, and a marriage Event. Relationship are not dateTime-stamped, but
    > relate temporal parts that are."

    Actually, relationships that are not timelessly true need some kind of
    time stamps -- either on the relationship, or on temporal slices of their
    arguments. The relationship does not need a time stamp if the arguments
    have them.

    E.g., to avoid timestamping the relationship, you could have the
    relationship being:

    * The time slice of Mary Smith from 1/1/1999 CE until the end of Marriage
    Situation #37387078043 is Married to the time slice of John Doe from
    1/1/1999
    CE until the end of Marriage Situation #37387078043.

    with Marriage Situation #37387078043 being the Marriage Node referred to
    below.

    > I agree with your statement that Relationships are not dateTime-stamped
    > but I would view this a bit differently
    >
    > * There are many Relationships
    >> * Mary Smith is Married to John Doe
    >> * John Doe is Married to Mary Smith
    >> * Mary Smith is Spouse Of John Doe
    >> * John Doe is Spouse of Mary Smith
    >> * Mary Smith is Wife of John Doe
    >> * John Doe is Husband of Mary Smith
    >> * Mary Smith Married John Doe
    >> * John Doe Married Mary Smith

    Note that each of these needs a time stamp to be true in some possible
    world. They are not relationships that were true 20,000 years ago. Nor
    were they true on the date of birth of the older of the two parties.
    Using the data from below, each of these relationships has a Start date of
    1/1/1999 CE and all but the last two an End Date yet to be determined but
    after 26/4/2016. Note that if one of the parties has a sex change, the
    fifth or sixth relationship may have a different end date than
    relationships 1-4.

    > * There is a Marriage Node with Traits
    >> * Spouses = Mary Smith; John Doe (pointer to People/Person Nodes)
    >> * Start Date = 1/1/1999 (pointer to Date Node)
    >> * End Date = TBD (pointer to Date Node)
    >> * Facility = Facility XYZ (pointer to Facility Node)
    > * There are many involved Nodes: See all nodes that the Marriage Node
    > involves in its existence

    > In the above, Marriages are nodes in the Graph that involve (relate to)
    > many other Nodes. I hope this helps clarity.
    ...

    Matthew West

    unread,
    Apr 26, 2016, 2:02:10 AM4/26/16
    to ontolo...@googlegroups.com

    Dear Douglas,
    The fact is that it was asserted.
    Regards
    Matthew West


    To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/c6cd67ad85e40249ee5f7a4f906f50d2.squirrel%40emailmg.ipage.com.

    doug foxvog

    unread,
    Apr 26, 2016, 2:10:09 AM4/26/16
    to ontolo...@googlegroups.com

    Pat,

    Note that these two sentences have the same meaning:
    * "Most scientists believe that X causes Y."
    * "Most scientists believe X causes Y."

    You can model this several ways. But each way that i see models the
    proposition as a first order individual.

    You could model the second sentence as having an elided "that". You could
    model it the first sentence as having its "that" restricting the meaning
    of "believe" (e.g., removing the possible meaning of believing an
    authority). You could model them both as having further elisions and
    being shortened forms of:
    * "Most scientists believe the truth of the statement 'X causes Y'."

    If we want to model this in logic, let's skip the complication of the
    subject by replacing it with "John".

    Then we have:
    * John believes the truth of Propostion8765056203.
    * Propostion8765056203 is the relationship 'X causes Y'.

    This can be done with nested triples, if you use a language that allows
    implicit quoting of arguments. Then if you have a predicate
    "believesProposition" with an implicitly quoted second argument, all you
    need is:

    * (believesProposition John (causes X Y))

    If you don't allow implicitly quoted arguments, but allow for a quoting
    operator, then you use that as well as your triples:

    * (believesProposition John (Quote (causes X Y)))
    or
    * (believesProposition John (TheProposition (causes X Y)))

    If you disallow a quoting operator either, then you cannot do this with
    triples. You can express it with a functional expression and a single
    triple, though:

    * (believesProposition John (TheProposition causes X Y))

    -- doug foxvog

    > From: Pat Hayes
    > Sent: Saturday, April 23, 2016 1:40 AM

    > +1, Chris.
    >
    > And by the way, y'all need to be careful about the distinction between "X
    > causes Y" and "that X causes Y". The first is (the form of) a statement,
    > which would have a truth-value. The second is (the form of) a way to refer
    > to a proposition. Trying to reduce this to 'triples' is way too crude and
    > simplistic.

    > Pat Hayes
    >
    > On Apr 22, 2016, at 10:36 AM, Christopher Menzel
    > ontolog-foru...@googlegroups.com<mailto:ontolog-foru...@googlegroups.com>.
    > To post to this group, send email to
    > ontolo...@googlegroups.com<mailto:ontolo...@googlegroups.com>.
    > Visit this group at https://groups.google.com/group/ontolog-forum.
    > To view this discussion on the web visit
    > https://groups.google.com/d/msgid/ontolog-forum/15DF6E68-5F54-42F4-A7EC-DC7974E30215%40gmail.com<https://groups.google.com/d/msgid/ontolog-forum/15DF6E68-5F54-42F4-A7EC-DC7974E30215%40gmail.com?utm_medium=email&utm_source=footer>.
    > For more options, visit https://groups.google.com/d/optout.
    >
    > ------------------------------------------------------------
    > IHMC (850)434 8903 home
    > 40 South Alcaniz St. (850)202 4416 office
    > Pensacola (850)202 4440 fax
    > FL 32502 (850)291 0667 mobile (preferred)
    > pha...@ihmc.us<mailto:pha...@ihmc.us>
    > http://www.ihmc.us/users/phayes
    >
    >
    >
    >
    >
    > --
    > All contributions to this forum by its members are made under an open
    > content license, open publication license, open source or free software
    > license. Unless otherwise specified, all Ontolog Forum content shall be
    > subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
    > ---
    > 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 post to this group, send email to
    > ontolo...@googlegroups.com<mailto:ontolo...@googlegroups.com>.
    > Visit this group at https://groups.google.com/group/ontolog-forum.
    > To view this discussion on the web visit
    > https://groups.google.com/d/msgid/ontolog-forum/64F3BC9D-C7A4-4BBF-A748-A81D96B44305%40ihmc.us<https://groups.google.com/d/msgid/ontolog-forum/64F3BC9D-C7A4-4BBF-A748-A81D96B44305%40ihmc.us?utm_medium=email&utm_source=footer>.
    > For more options, visit https://groups.google.com/d/optout.
    > --
    > All contributions to this forum by its members are made under an open
    > content license, open publication license, open source or free software
    > license. Unless otherwise specified, all Ontolog Forum content shall be
    > subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
    > ---
    > 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 post to this group, send email to
    > ontolo...@googlegroups.com<mailto:ontolo...@googlegroups.com>.
    > Visit this group at https://groups.google.com/group/ontolog-forum.
    > To view this discussion on the web visit
    > https://groups.google.com/d/msgid/ontolog-forum/CY1PR09MB0826CBDCDBD50EE60B86333ADD610%40CY1PR09MB0826.namprd09.prod.outlook.com<https://groups.google.com/d/msgid/ontolog-forum/CY1PR09MB0826CBDCDBD50EE60B86333ADD610%40CY1PR09MB0826.namprd09.prod.outlook.com?utm_medium=email&utm_source=footer>.
    > For more options, visit https://groups.google.com/d/optout.
    >
    > --
    > All contributions to this forum by its members are made under an open
    > content license, open publication license, open source or free software
    > license. Unless otherwise specified, all Ontolog Forum content shall be
    > subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
    > ---
    > 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 post to this group, send email to
    > ontolo...@googlegroups.com<mailto:ontolo...@googlegroups.com>.
    > Visit this group at https://groups.google.com/group/ontolog-forum.
    > To view this discussion on the web visit
    > https://groups.google.com/d/msgid/ontolog-forum/D341C7EB.146A5C%25Frank.Guerino%40if4it.com<https://groups.google.com/d/msgid/ontolog-forum/D341C7EB.146A5C%25Frank.Guerino%40if4it.com?utm_medium=email&utm_source=footer>.
    > For more options, visit https://groups.google.com/d/optout.
    >
    > --
    > All contributions to this forum by its members are made under an open
    > content license, open publication license, open source or free software
    > license. Unless otherwise specified, all Ontolog Forum content shall be
    > subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
    > ---
    > 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 post to this group, send email to ontolo...@googlegroups.com.
    > Visit this group at https://groups.google.com/group/ontolog-forum.
    > To view this discussion on the web visit
    > https://groups.google.com/d/msgid/ontolog-forum/CY1PR09MB082690E46B005AA868299C3FDD610%40CY1PR09MB0826.namprd09.prod.outlook.com.

    HansTeijgeler

    unread,
    Apr 26, 2016, 8:28:58 AM4/26/16
    to ontolo...@googlegroups.com

    "We don't believe things because they are true, things are true because we believe them."

    Hans

    Frank Guerino

    unread,
    Apr 26, 2016, 8:57:08 AM4/26/16
    to Ontolog Forum
    Hello all,

    I need to amend one of my responses…

    Hans wrote: “Resuming: above there was a marriage Relationship, a marrying Activity, and a marriage Event. Relationship are not dateTime-stamped, but relate temporal parts that are.”

    I Responded: "I agree with your statement that Relationships are not dateTime-stamped but I would view this a bit differently…"

    I need to correct this, as I believe Relationships “always” have time stamps (even if we don’t implement such timestamps in our models).

    1. At a minimum, relationships will always have a timestamp for when they come into existence.
    2. I believe they can also have timestamps for end of validity, meaning they stop being assertions and exist simply as historical memories.
    3. If we believe that relationships can have traits, then we must believe such traits can change and, therefore, it is possible we could also infer a modification timestamp.
    As Graphs/Networks move through time, we see new relationships evolve and establish themselves (at specific points in time).  I believe these new relationships help reinforce or invalidate existing older relationships.

    I apologize for any confusion.

    I need to stop replying in the middle of the night, when I’m low on sleep and loaded with with allergy medications!

    HansTeijgeler

    unread,
    Apr 26, 2016, 9:19:06 AM4/26/16
    to ontolo...@googlegroups.com

    Frank,

    The persisting problem is that whatever I explain, you ignore that.

    I explained to you what our definition is of Relationship, and that we work with temporal parts, and that relationships are between temporal parts, and that the timestamps are given to those temporal parts.

    Rather than claiming that I was wrong and you are right, please keep in mind that we work from two different assumption/definitions.

    In the old days of ISO 15926, which was then called ISO 10303-221, we had Associations instead of Relationships, and no temporal parts, and yes, these Associations were timestamped, as you wrote below.


    Since you still work with relations (our lingo) I wonder whether you really timestamp each of those when working in RDF. That would mean that ALL relations (rdf:Property) would have to be reified. :-((

    Regards,

    Hans

    --
    All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontolog Forum content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
    ---
    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 post to this group, send email to ontolo...@googlegroups.com.
    Visit this group at https://groups.google.com/group/ontolog-forum.

    Christopher Menzel

    unread,
    Apr 26, 2016, 2:20:01 PM4/26/16
    to ontolo...@googlegroups.com
    I believe that's one of the dumbest sayings I've ever read. :-)

    -chris

    Christopher Menzel

    unread,
    Apr 26, 2016, 2:36:00 PM4/26/16
    to ontolo...@googlegroups.com
    On 25 Apr 2016, at 4:28 PM, William Frank <william...@gmail.com> wrote:
    Christoper is of course correct that asserting something is true adds nothing to simply asserting except, as he points out, we have a metalanguage indirection.  He says:

    if "φ is true" is a fact, so is φ.  This implies that φ is an assertion.

    I never said "This implies that φ is an assertion" nor would I have said it, as "assertion" is too vague — it is ambiguous between a certain speech act, viz., the utterance of a declarative sentence, the sentence uttered in such an act, and the proposition expressed by the sentence uttered in such an act. In what I wrote, φ can be understood either to be a name for declarative sentence or an expression for a proposition like a that-clause, depending on what you take the bearers of truth values to be.

    But, we can even in that jumble that is English, say "That polly is a parrot is true".  In this case, if  φ  = 'that polly is a parrot' then φ is NOT an assertion, it is a proposition, which might be true or false.

    If φ = 'that Polly is a parrot', then φ is a noun phrase that (as commonly understood) denotes a proposition. If φ = that Polly is a parrot, i.e., if φ is what 'that Polly is a parrot' denotes, then φ is a proposition.

    -chris

    John F Sowa

    unread,
    Apr 26, 2016, 2:42:25 PM4/26/16
    to ontolo...@googlegroups.com
    On 4/26/2016 2:19 PM, Christopher Menzel wrote:
    > I believe that's one of the dumbest sayings I've ever read. :-)
    >
    >> "We don't believe things because they are true, things are
    >> true because we believe them."

    Unfortunately, some things are both dumb and dangerous.
    Not so long ago, some politicians in positions of power said
    > If you repeat something often enough, it becomes true.

    It's dangerous because (a) the people who said it knew
    it was false, (b) what they really meant was
    > If you repeat something often enough, gullible voters
    > will believe that it's true.

    And (c) they were right in that assumption.

    John

    Frank Guerino

    unread,
    Apr 26, 2016, 4:50:32 PM4/26/16
    to Ontolog Forum
    Hi Hans,

    You wrote: “I explained to you what our definition is of Relationship, and that we work with temporal parts, and that relationships are between temporal parts, and that the timestamps are given to those temporal parts.”

    Please do not believe that I am ignoring you.  I am not.  I understand that you work with relationships that are between temporal parts and I respect that.  However, please know that while you believe that and work with such assumptions, I do not.  My own view is that the Relationship is also temporal.  For example:

    • If we have the Semantic Relationship "Jane is Married to John” (and all equivalent relationships such as “Is Spouse Of” and their directional inversions), we consider, both, Jane and John to be independent temporal parts.
    • These independent temporal parts age with time and their states change with time.
    • The relationships associated with their marriage (all corresponding to a new event Node called a Marriage) all occur at a specific point in time that is independent of their individual temporal changes (although it locks in at specific points in both of their temporal changes, meaning they have exact ages at the time of the relationship creations).

    You wrote: “Rather than claiming that I was wrong and you are right, please keep in mind that we work from two different assumption/definitions.”

    Han, please know that I never made such a claim.   I simply stated my beliefs.  As a matter of fact, I openly acknowledge that neither of us has any evidence to prove that either of us is right or wrong.  In fact, we could both be right or both be wrong.  I simply engaged looking for opinions and views.

    You wrote: “Since you still work with relations (our lingo) I wonder whether you really timestamp each of those when working in RDF. That would mean that ALL relations (rdf:Property) would have to be reified. :-(("

    Also, please be reminded that we don’t use RDF because we have yet to see any integration case that justifies the business value of doing so.  No systems we interact with or plan to interact with use or plan to use RDF or any of the ISO standards (to our knowledge).  I acknowledge that this may change, one day, but has not as of today.

    You had defined relations in your text…

    Object1 << relation1 << Relationship >> relation2 >> Object2

    CENTRIFUGAL-PUMP << hasClassOfWhole << ClassOfAssemblyOfIndividual_439839 >> hasClassOfPart >> IMPELLER

    myPump << hasWhole << AssemblyOfIndividual_238743 >> hasPart >> myImpeller


    The core Graph/Network constructs we currently work with are:
    1. Noun Types (Ontology for defining Nodes and also represent Type Nodes)
    2. Noun Type Metadata (Defines internal traits of each Noun Type, where each trait is an Attribute Node)
    3. Noun Instances (Equivalent to Nodes of specific Noun Types)
    4. Semantic Relationships (which bind Subject Noun Instances to Object Noun Instances through Predicates) [e.g. “John Doe is a Product Owner of Product XYZ”]
    5. Edges (a.k.a. Predicates or Relationship Descriptors that explain how any two Nodes are bound) [e.g. “Product Owner” in the previous Semantic Relationship example].  Edges, like Types, are a form of Node.
    6. (There are others but they are inconsequential for this thread.)

    We already reify Edges (#5), as they point to all similar Edges.

    My original question to the community was about whether or not an entire Semantic Relationship (#5) could point to something else.  I believe, for example, that a Semantic Relationship can point to its history over time, at a minimum, and was simply looking for more opinions/views on this.

    I hope this helps clarify.

    My Best,

    Frank
    --
    Frank Guerino, Chairman
    The International Foundation for Information Technology (IF4IT)
    http://www.if4it.com
    1.908.294.5191 (M)
    From: HansTeijgeler <hans.te...@quicknet.nl>
    Reply-To: Ontolog Forum <ontolo...@googlegroups.com>
    Date: Tuesday, April 26, 2016 at 9:19 AM
    To: Ontolog Forum <ontolo...@googlegroups.com>
    Subject: Re: [ontolog-forum] Are Relationship Triples also Nodes?

    Frank Guerino

    unread,
    Apr 26, 2016, 5:07:36 PM4/26/16
    to Ontolog Forum
    Hi Matthew,

    In a previous post, you made a statement that got me thinking and I’m just now getting around to a response.  You wrote: “If you take a nodes and edges view of the world then in principal any edge can be replaced by a node and two more edges. The problem is where to stop?”

    I do not know that I’ve seen the same thing.  An edge, such as a Predicate, can be treated as a specific type of Node but our research has shown us that it has a limited set of what it can point to, which is all Semantic Relationships that also use that Edge.  For example, if we have a Semantic Relationship with an Edge = “Subject Matter Expert”, such as:

    Resource "Main, Mary" is related as a "Subject Matter Expert” of Business Domain “Sales"

    Our work has shown us that the predicate “Subject Matter Expert” can point only to the set of all other similar predicates that exist within all other Semantic Relationships that use it (see link).  The implication is that this set is always a subset of the greater Network/Graph.

    Does this logic make sense or am I missing something?

    My Best,

    Frank
    --
    Frank Guerino, Chairman
    The International Foundation for Information Technology (IF4IT)
    http://www.if4it.com
    1.908.294.5191 (M)


    From: Matthew West <dr.matt...@gmail.com>
    Reply-To: Ontolog Forum <ontolo...@googlegroups.com>
    Date: Thursday, April 21, 2016 at 4:50 PM
    To: Ontolog Forum <ontolo...@googlegroups.com>
    Subject: Re: [ontolog-forum] Are Relationship Triples also Nodes?

    Dear Frank,
    If you take a nodes and edges view of the world then in principal any edge can be replaced by a node and two more edges. The problem is where to stop? The thing to note is that you can say something about nodes (more edges to other nodes) but not about edges. In my analysis my conclusion was that there was no more information to record after you had identified the state playing the role in a relationship. So no point going beyond that. That is how ISO15926 works. You can always have convenient short cuts across that base layer. In SO 15926 these are templates.
    Regards
    Matthew West



    Pat Hayes

    unread,
    Apr 26, 2016, 9:51:02 PM4/26/16
    to Frank Guerino, ontolog-forum
    On Apr 26, 2016, at 4:07 PM, Frank Guerino <Frank....@if4it.com> wrote:

    Hi Matthew,

    In a previous post, you made a statement that got me thinking and I’m just now getting around to a response.  You wrote: “If you take a nodes and edges view of the world then in principal any edge can be replaced by a node and two more edges. The problem is where to stop?”

    I do not know that I’ve seen the same thing.  An edge, such as a Predicate, can be treated as a specific type of Node but our research has shown us that it has a limited set of what it can point to, which is all Semantic Relationships that also use that Edge.  For example, if we have a Semantic Relationship with an Edge = “Subject Matter Expert”, such as:

    Resource "Main, Mary" is related as a "Subject Matter Expert” of Business Domain “Sales"

    Our work has shown us that the predicate “Subject Matter Expert” can point only to ...

    Well now, "can point only to" is a very strong claim. In an RDF graph, node names and edge names (IRIs) are treated identically so a single IRI may be both a node and an edge, and both roles are independent of what it "points to", if that terminology is interpreted to mean what an HTTP request will yield from the Web. In fact, the semantics of the IRI in its role in RDF is completely separated from its role in Web data access, so it can point to (literally) anything. 

    If your claim, above, is true as stated, then your research work would seem to have shown that RDF is incoherent, which, with respect, I doubt. 

    Pat Hayes

    the set of all other similar predicates that exist within all other Semantic Relationships that use it (see link).  The implication is that this set is always a subset of the greater Network/Graph.

    Does this logic make sense or am I missing something?

    My Best,

    Frank
    --
    Frank Guerino, Chairman
    The International Foundation for Information Technology (IF4IT)
    http://www.if4it.com
    1.908.294.5191 (M)


    From: Matthew West <dr.matt...@gmail.com>
    Reply-To: Ontolog Forum <ontolo...@googlegroups.com>
    Date: Thursday, April 21, 2016 at 4:50 PM
    To: Ontolog Forum <ontolo...@googlegroups.com>
    Subject: Re: [ontolog-forum] Are Relationship Triples also Nodes?

    Dear Frank,
    If you take a nodes and edges view of the world then in principal any edge can be replaced by a node and two more edges. The problem is where to stop? The thing to note is that you can say something about nodes (more edges to other nodes) but not about edges. In my analysis my conclusion was that there was no more information to record after you had identified the state playing the role in a relationship. So no point going beyond that. That is how ISO15926 works. You can always have convenient short cuts across that base layer. In SO 15926 these are templates.
    Regards
    Matthew West



    --
    All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontolog Forum content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
    ---
    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 post to this group, send email to ontolo...@googlegroups.com.
    Visit this group at https://groups.google.com/group/ontolog-forum.

    For more options, visit https://groups.google.com/d/optout.

    ------------------------------------------------------------
    IHMC                                     (850)434 8903 home
    40 South Alcaniz St.            (850)202 4416   office
    Pensacola                            (850)202 4440   fax
    FL 32502                              (850)291 0667   mobile (preferred)

    Frank Guerino

    unread,
    Apr 27, 2016, 12:07:25 AM4/27/16
    to Ontolog Forum
    Hi Pat,

    I wrote: “Our work has shown us that the predicate “Subject Matter Expert” can point only to"

    You Responded: “Well now, "can point only to" is a very strong claim.”

    Please know that my response was not to imply that our research covered all definitive permutations.  It was meant as a statement that "so far" we have not seen anything else that an Edge can point to.  “So far," we’ve seen that an Edge can point to its sphere of Equivalents (e.g. Synonyms and Preferred Terms) and !Equivalents (e.g. Antonyms).

    You wrote: “In an RDF graph, node names and edge names (IRIs) are treated identically so a single IRI may be both a node and an edge, and both roles are independent of what it "points to", if that terminology is interpreted to mean what an HTTP request will yield from the Web. In fact, the semantics of the IRI in its role in RDF is completely separated from its role in Web data access, so it can point to (literally) anything. “

    This is interesting and I’m trying to understand it, better.  I certainly agree that Edges are a form of a Node, as are Types and Instances of Types.  However, I have yet to see them (Edges) point to anything other than Equivalents or !Equivalents, in a Graph.

    In my earlier post, I gave an example using the Edge = “Subject Matter Expert”, which pointed to all other Edges = “Subject Matter Expert".  Could I please ask you to provide me with one or more examples of other things Edge can point to?  In other words, if we stick with the same example:

    Resource "Main, Mary" is related as a "Subject Matter Expert” of Business Domain “Sales"

    …what else do you believe the Edge = “Subject Matter Expert” can point to?

    My Best,

    Frank
    --
    Frank Guerino, Chairman
    The International Foundation for Information Technology (IF4IT)
    http://www.if4it.com
    1.908.294.5191 (M)

    HansTeijgeler

    unread,
    Apr 27, 2016, 3:49:35 AM4/27/16
    to ontolo...@googlegroups.com

    Hi Chris,

    That is true (to you), because you believe it :-))

    Hans

    --
    All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontolog Forum content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
    ---
    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 post to this group, send email to ontolo...@googlegroups.com.
    Visit this group at https://groups.google.com/group/ontolog-forum.

    John F Sowa

    unread,
    Apr 27, 2016, 8:49:57 AM4/27/16
    to ontolo...@googlegroups.com
    On 4/27/2016 3:49 AM, Hans Teijgeler wrote:
    > That is true (to you), because you believe it :-))

    The phrase 'to you' is a necessary, but confusing addendum.

    The model-theoretic semantics for all versions of logic -- including
    informal logics stated in natural languages -- depends on a domain D,
    over which the quantifiers range. They include the two basic
    quantifiers, some and every, and any other quantifiers that are
    defined or definable in terms of them: most, many, few, 1, 2, 3...

    For any logic L (formal or informal, crisp or fuzzy) and domain D,
    the semantics of L determine the truth value of any sentence in L.

    Given that explanation, you could say that the phrase 'to you' is
    a short way of implying "Within the domain of your beliefs and
    your personal way of thinking, the sentence you believe is true."

    It's better to say "You can believe that if you want to, but it
    ain't necessarily so."

    John

    +1 613-791-2183

    unread,
    Apr 27, 2016, 11:40:07 AM4/27/16
    to ontolo...@googlegroups.com

    0q

    To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/D343ADEF.1470F5%25Frank.Guerino%40if4it.com.

    +1 613-791-2183

    unread,
    Apr 27, 2016, 11:40:08 AM4/27/16
    to ontolo...@googlegroups.com

    0q

    On Apr 25, 2016 11:35 AM, Frank Guerino <Frank....@if4it.com> wrote:
    To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/D343ADEF.1470F5%25Frank.Guerino%40if4it.com.

    Christopher Menzel

    unread,
    Apr 27, 2016, 12:30:02 PM4/27/16
    to ontolo...@googlegroups.com
    On Apr 27, 2016, at 2:49 AM, HansTeijgeler <hans.te...@quicknet.nl> wrote:
    > That is true (to you), because you believe it :-))

    “φ is true to X” is nothing more than “X believes φ”. So if “true to us” is all you mean by “true” in “things are true because we believe them”, I obviously have no objection, since that of course renders the aphorism logically trivial (and hence completely uninteresting).

    -chris

    > ------------------------------------------------------------------------
    >
    > On 26-4-2016 20:19, Christopher Menzel wrote:
    >> I believe that's one of the dumbest sayings I've ever read. :-)
    >>
    >> -chris
    >>

    HansTeijgeler

    unread,
    Apr 27, 2016, 12:59:41 PM4/27/16
    to ontolo...@googlegroups.com
    So that is settled.
    Hans

    Pat Hayes

    unread,
    Apr 27, 2016, 8:54:45 PM4/27/16
    to ontolog-forum, Frank Guerino
    On Apr 26, 2016, at 11:07 PM, Frank Guerino <Frank....@if4it.com> wrote:

    Hi Pat,

    I wrote: “Our work has shown us that the predicate “Subject Matter Expert” can point only to"

    You Responded: “Well now, "can point only to" is a very strong claim.”

    Please know that my response was not to imply that our research covered all definitive permutations.  It was meant as a statement that "so far" we have not seen anything else that an Edge can point to.  “So far," we’ve seen that an Edge can point to its sphere of Equivalents (e.g. Synonyms and Preferred Terms) and !Equivalents (e.g. Antonyms).

    When you say "... can point to...", you seem to be making a claim about possibility: if X can not point to Y, then X's pointing to Y is *impossible*. That is an absolute claim, nothing about "so far" in it. Perhaps you use the English word "can" in a different way, however. 

    It is also possible that I have misunderstood what you mean by "point to"; I confess to finding your usage slightly opaque at this point. In the graph model I am most intimately familiar with, the RDF graph, the names used to label nodes and arcs are all IRIs, formerly called URIs or URLs, which are also Web "addresses" (this usage is frowned upon by Web engineers, so I use scare quotes) ie which "point to" things on the Web. Here, "point to" means that the pointee will be returned as the payload of an HTTP GET request using the pointer as the address. (This is the Web version of the familiar and venerable computer science usage of referring to a file address as a "pointer".) As I said previously, this "pointing to" by an IRI is completely independent of and separate from the semantic import of that same IRI when used as a graph name (node or edge) in RDF. 

    However, perhaps you mean "point to" in the sense in which edges are seen as directed lines or "arrows" pointing from one node to another. But if so, then there is no room for any doubt about what an edge points to: it points to the node at the sharp end of its arrowhead, so I had assumed you did not mean this. 


    You wrote: “In an RDF graph, node names and edge names (IRIs) are treated identically so a single IRI may be both a node and an edge, and both roles are independent of what it "points to", if that terminology is interpreted to mean what an HTTP request will yield from the Web. In fact, the semantics of the IRI in its role in RDF is completely separated from its role in Web data access, so it can point to (literally) anything. “

    This is interesting and I’m trying to understand it, better.  I certainly agree that Edges are a form of a Node

    I would rather say that edges and nodes are distinct roles that a name can play, and that it may play both roles. The semantic import of a name is somewhat different in each role, however, so it is important not to confuse them. 

    , as are Types and Instances of Types.  However, I have yet to see them (Edges) point to anything other than Equivalents or !Equivalents, in a Graph.

    Well, for example, one informal (ie not imposed by the normative standards) usage, which is however becoming quite widespread, is that each IRI "points to" (in the HTTP GET sense) a document which more or less formally defines its meaning, either by giving a defining ontology or perhaps only a textual indication of what the name is intended to refer to. This seems to be different from your experience, if I follow you correctly. 

    In my earlier post, I gave an example using the Edge = “Subject Matter Expert”, which pointed to all other Edges = “Subject Matter Expert".  Could I please ask you to provide me with one or more examples of other things Edge can point to?  In other words, if we stick with the same example:

    Resource "Main, Mary" is related as a "Subject Matter Expert” of Business Domain “Sales"

    …what else do you believe the Edge = “Subject Matter Expert” can point to?

    Per the above, it might identify a document which provides a definition of, or an ontology which formally axiomatizes, the notion expressed by this usage of "subject matter expert". 

    Pat 


    For more options, visit https://groups.google.com/d/optout.

    Frank Guerino

    unread,
    Apr 28, 2016, 10:06:05 AM4/28/16
    to Ontolog Forum
    Hi Pat,

    Let me start by answering your last statement, first, as I believe it will help add context.

    You wrote: “However, perhaps you mean "point to" in the sense in which edges are seen as directed lines or "arrows" pointing from one node to another. But if so, then there is no room for any doubt about what an edge points to: it points to the node at the sharp end of its arrowhead, so I had assumed you did not mean this.”

    Yes, in a sense this is what I mean.  A Node can point to many similar Nodes (homogeneous) or different Nodes (heterogeneous).  In the example:

    • Person Doe, Jane is related as a Subject Matter Expert to Business Domain Marketing

    The edge “Subject Matter Expert” is a type of Node, itself, that can point to other Edges.  For example, it can point:

    • To other Edges of the same descriptive Predicate = “Subject Matter Expert”
    • To its defined Equivalents, such as “SME” or “Expert”
    • To its defined anti-Equivalents (!Equivalents), such as “Amateur”  (NOTE: “opposites” are common examples of !Equivalents)

    This allows, for example, to traverse from “Subject Matter Expert” to “Expert” to all Node Instances that explicitly relate to “Expert", in addition to those that explicitly relate to "Subject Matter Expert”.

    You wrote: “When you say "... can point to...", you seem to be making a claim about possibility: if X can not point to Y, then X's pointing to Y is *impossible*. That is an absolute claim, nothing about "so far" in it. Perhaps you use the English word "can" in a different way, however.”

    In a sense, yes.  “Can only point to” represents a bounded set.  When someone makes the statement “Those cherries” it infers specific cherries (a subset) in the set of all cherries, which are a subset of all fruits).   “Those cherries” points to a specific bounded set of cherries and not to anything else.

    It is the same with the “Subject Matter Expert” example.  When using such a Predicate, it appears we can only find and access a specific bounded set of things that are 1 Degree of Separation away from it, in the Graph (i.e. things it points to).  It cannot directly point to things that are 2+ Degrees of Separation, away, and therefore we cannot access such “payloads” without first traversing through its 1st Degree related Nodes.

    Structurally, a Graph allows us the flexibility to bind anything to anything. However, the brain and nature don’t appear to work that way.  Specific things are bound to other specific things in very specific ways, implying bounded sets of relationships.  Hence “can only point to.”

    I qualify all of the above with… It is what I’ve seen based on years of work and I clearly may be wrong and am willing to learn new things.  :-)

    Pat Hayes

    unread,
    Apr 28, 2016, 11:05:48 AM4/28/16
    to ontolog-forum, Frank Guerino
    On Apr 28, 2016, at 9:06 AM, Frank Guerino <Frank....@if4it.com> wrote:

    Hi Pat,

    Let me start by answering your last statement, first, as I believe it will help add context.

    You wrote: “However, perhaps you mean "point to" in the sense in which edges are seen as directed lines or "arrows" pointing from one node to another. But if so, then there is no room for any doubt about what an edge points to: it points to the node at the sharp end of its arrowhead, so I had assumed you did not mean this.”

    Yes, in a sense this is what I mean.

    OK, but in that case, your next sentence does not make sense:

     A Node can point to many similar Nodes (homogeneous) or different Nodes (heterogeneous).

    If your nodes and edges are part of a graph, and if "point to" is understood in the above sense, then an edge can only point to a single node. If you have a hypergraph, then indeed an edge can "point to" several other nodes. Is this what you mean? But then I see no reason why a hyperedge might not point to anything in the graph. Perhaps by 'edge' you mean to refer to all edges labelled with the same 'predicate' name? Ie to a set of edges rather than a single edge?

     In the example:

    • Person Doe, Jane is related as a Subject Matter Expert to Business Domain Marketing

    The edge “Subject Matter Expert” is a type of Node, itself, that can point to other Edges.  For example, it can point:

    • To other Edges of the same descriptive Predicate = “Subject Matter Expert”
    • To its defined Equivalents, such as “SME” or “Expert”
    If it points to SME, then why would this not be saying that SME was the actual object of "Subject Matter Expert" rather than an equivalent? Or do you have several different kinds of "pointing"?

    • To its defined anti-Equivalents (!Equivalents), such as “Amateur”  (NOTE: “opposites” are common examples of !Equivalents)

    This allows, for example, to traverse from “Subject Matter Expert” to “Expert” to all Node Instances that explicitly relate to “Expert", in addition to those that explicitly relate to "Subject Matter Expert”.

    You wrote: “When you say "... can point to...", you seem to be making a claim about possibility: if X can not point to Y, then X's pointing to Y is *impossible*. That is an absolute claim, nothing about "so far" in it. Perhaps you use the English word "can" in a different way, however.”

    In a sense, yes.  “Can only point to” represents a bounded set.  When someone makes the statement “Those cherries” it infers specific cherries (a subset) in the set of all cherries, which are a subset of all fruits).   “Those cherries” points to a specific bounded set of cherries and not to anything else.

    "those cherries" is an indexical usage. I don't see the relevance of that to our discussion here. 

    It is the same with the “Subject Matter Expert” example.

    Do you mean that "subject matter expert" is understood to be an indexical? 

     When using such a Predicate, it appears we can only find and access a specific bounded set of things that are 1 Degree of Separation away from it, in the Graph (i.e. things it points to).

    Not sure what you mean by "find and access" here. But if you are saying that only some things are related by any given relation (Predicate, if you like), then yes of course that is true, but that is not to say that the set of such things is somehow limited or restricted a priori. Relations can relate all kinds of things, and can themselves be related to other things in all kinds of ways. 

     It cannot directly point to things that are 2+ Degrees of Separation, away, and therefore we cannot access such “payloads” without first traversing through its 1st Degree related Nodes.

    Structurally, a Graph allows us the flexibility to bind anything to anything. However, the brain and nature don’t appear to work that way.

    Really? It seems to me that nature has no bearing on this, and that our brains are quite capable of dealing with arbitrary graph structures. Mine is, at any rate :-)

    Pat


    For more options, visit https://groups.google.com/d/optout.

    John F Sowa

    unread,
    Apr 28, 2016, 12:10:19 PM4/28/16
    to ontolo...@googlegroups.com
    Frank and Pat,

    FG
    > Structurally, a Graph allows us the flexibility to bind anything
    > to anything.

    Yes, and so does a linear notation such as predicate calculus and
    the many linear and diagrammatic notations that map to and from it.

    FG
    > A Node can point to many similar Nodes (homogeneous) or
    > different Nodes (heterogeneous).

    Yes. That's a useful option in many, many versions of logic.

    FG
    > However, the brain and nature don’t appear to work that way.
    > Specific things are bound to other specific things in very
    > specific ways, implying bounded sets of relationships.

    PJH
    > Really? It seems to me that nature has no bearing on this,
    > and that our brains are quite capable of dealing with arbitrary
    > graph structures.

    I agree. The human brain supports natural languages, which can
    express anything that can be said in any artificial notation of
    any kind. That includes every version of mathematics, symbolic
    logic, diagrams, the Semantic Web...

    FG
    > Hence “can only point to.”

    With a gesture, a facial expression, or a way of behaving,
    people can express or suggest aspects of a situation that
    nobody has yet expressed precisely in logic or even natural
    languages.

    But every detail of every digital computer can be completely
    and definitively defined in first-order predicate calculus.
    Every programming language compiled to or interpreted by
    any digital computer can also be fully and definitively
    specified by FOL.

    Therefore, any notation that may be implemented in any
    programming language that runs on any digital computer can
    be specified in logic. It's impossible to implement a method
    of "pointing to" that cannot be represented in logic.

    FG
    > I clearly may be wrong and am willing to learn new things.

    OK. I'll suggest a reading list that cites 100+ documents
    that include many more references: http://www.jfsowa.com/ikl/

    Section 1 starts in the 1980s with the AI work on knowledge
    bases and the DB conceptual schema for databases. It cites
    the early work on computational ontologies, the projects that
    led to the Semantic Web, and the original documents for the
    SW and the DARPA project (from 2000 to 2005).

    Section 2 covers the IKRIS Project (Interoperable Knowledge
    Representation for Intelligence Support). It was sponsored by
    the DoD in 2005 and 2006, and it brought together an impressive
    group of researchers in AI, logic, ontology, and the SW. The
    references cite the original documents and later articles.

    Sections 3 and 4 cover more recent developments.

    For the history of semantic networks and related graph notations,
    see http://www.jfsowa.com/pubs/semnet.pdf

    Existential graphs by C. S. Peirce implement a version of logic
    that is a superset of the DRS notation developed by Hans Kamp
    for natural language semantics. EGs also have a direct mapping
    to and from the IKRIS notation. For an overview, see
    http://www.jfsowa.com/talks/egintro.pdf

    John

    Frank Guerino

    unread,
    Apr 28, 2016, 12:30:05 PM4/28/16
    to Ontolog Forum
    Hi Pat,

    Unfortunately, the word “Node” gets overused and lost in the translation, sometimes.  Let me clarify a little more by identifying types of Nodes.  Some examples include:

    • A “Type or Classification Node”.  A Type Node points to one or more Type Instance Nodes.  E.g. “Fruits” is a type (as opposed to "Vegetables”)and will point to all Fruits in the graph.
    • A “Type Instance Node”.  This is a specific Instance of a Type Node.  A very  specific Apple is an example.
    • An "Edge (or Arc) Node”.  Examples is a “Subject Matter Expert” which can exist as an Edge/Arc between a Person Type Instance Node and a Topic Type Instance Node.
    One type of "Semantic Relationship” is a bridge between Type Instance Nodes, via an Edge Node.  For example:

    E.g Person Doe, Jane is a Subject Matter Expert to Topic XYZ

    Another type of a "Semantic Relationship” is a bridge between two Edge Nodes, via another Edge Node.  For example:

    E.g. A Subject Matter Expert is related to a SME Definition Node which helps support and define the relationships between one SME and all other Subject Matter Experts.

    You wrote: “If your nodes and edges are part of a graph, and if "point to" is understood in the above sense, then an edge can only point to a single node."

    I don’t’ know if I agree with this statement.  In the above two examples of Semantic Relationships, we can get to

    • Former Example: the two Type Instances Nodes in that specific Subject Matter Expert instance
    • In The Latter Example: all other Subject Matter Experts with the help of a Definition Node (a whole other topic of conversation because: Do we get to other SMEs through or with parallel use of the Definition Node?).

    However, going back to my earlier statement… I believe that the Edge Node “Subject Matter Expert” can point to multiple things but definitely has a bounded/limited set of things it can point to.

    You wrote: "Do you mean that "subject matter expert" is understood to be an indexical?”

    I have never used “indexical” but it appears that the answer to this question is “yes.”  I gather that “Type Nodes” (see definition above) would also be examples of Indexical Nodes.  (Thanks for teaching me this term!)

    Thinking about “Indexical" a bit further… I would suggest that all Nodes are Indexical as all Nodes point to one or more things and an Index is, in short, defined to be something that provides quick and organized access to a limited set of things that are all part of a broader set of things.  In fact, while in theory a single node can exist with no bindings, I don’t believe there is an example (in reality) that doesn’t relate to at least one other node (after we observe/analyze it).

    You wrote: “Relations can relate all kinds of things, and can themselves be related to other things in all kinds of ways. “

    I agree but I would qualify that separating the data structure from reality sheds some light.  A Data Structure that is a Graph can relate anything to anything.  However, in reality and to my knowledge, we have no evidence to prove that everything is related to everything else.  Instead, it appears that specific things are related to other specific things.

    I wrote: “Structurally, a Graph allows us the flexibility to bind anything to anything. However, the brain and nature don’t appear to work that way.”  And you responded: “Really? It seems to me that nature has no bearing on this, and that our brains are quite capable of dealing with arbitrary graph structures. Mine is, at any rate :-)”

    My point is that when we learn something (in nature) we only bind what we’ve learned together with other things we’ve learned (all in the past).  The possibility of relating everything to everything else certainly exists (the empty Graph data structure) but do we really relate anything, arbitrarily, to any and all other things before we learn/derive them?  Or, do we relate things after we learn/derive them?  Our brains appear to only know things after we observe them and after we store and create such relationships.  Otherwise, I believe we’d all know the future.

    Also, if we start with the reverse and make the hypothesis that nature binds everything to everything else, I believe we have no way of ever proving it because the tools we use (our brains and computers) can only create relationships within and between the sets of data we’ve learned… which is “always” limited/bound.

    (A very fun conversation.  Thank you!)

    My Best,

    Frank
    --
    Frank Guerino, Chairman
    The International Foundation for Information Technology (IF4IT)
    http://www.if4it.com
    1.908.294.5191 (M)



    From: Pat Hayes <pha...@ihmc.us>
    Reply-To: Ontolog Forum <ontolo...@googlegroups.com>
    Date: Thursday, April 28, 2016 at 11:05 AM
    To: Ontolog Forum <ontolo...@googlegroups.com>, Frank Guerino <Frank....@if4it.com>
    Subject: Re: [ontolog-forum] Are Relationship Triples also Nodes?

    Frank Guerino

    unread,
    Apr 28, 2016, 3:03:07 PM4/28/16
    to Ontolog Forum
    Hi John,

    Your site (http://www.jfsowa.com/ikl/) is extremely interesting.  I’ve been there before but not recently.  I especially appreciate how you’ve taken the time to associate similar/related concepts in a paragraph by paragraph story-based approach.  There is a great deal for KM professionals to learn from what you’ve done and how you’ve done it.

    This being said, there are also decades of knowledge rolled up into your site, making it difficult (despite all your added efficiencies) to sit and digest everything over a short and economical period of time.

    You wrote: “With a gesture, a facial expression, or a way of behaving, people can express or suggest aspects of a situation that nobody has yet expressed precisely in logic or even natural languages.”

    Bodies are nothing more than bio-machines (i.e. Tools) for minds.  We’re figuring out, every day, how to decompose and mimic these things, digitally + mechanically, every day.  For example, a couple of days ago I watched a news clip on Sign Language Gloves that when worn during signing allow direct wireless control of a voice translator to pure English.

    You wrote: “It's impossible to implement a method of "pointing to" that cannot be represented in logic.”

    I agree.  However, this brings up the point that the logic representation can often be implemented different ways, as witness by the similar features being implement in different systems, with different technologies, in different ways.  There is often no guarantee that any one way is any better than any other until we collect enough historical data, over long term adoption and use.  I feel this way about things like RDF.  It’s nice but there’s no guarantee it is the best solution and data only shows that most mainstream applications have no intentions of adopting it.  Getting back to the point, just because something is represented does not imply it is represented 1) the only way or 2) the best way.

    I wrote: "However, the brain and nature don’t appear to work that way. Specific things are bound to other specific things in very specific ways, implying bounded sets of relationships.”  

    Pat Responded: “Really? It seems to me that nature has no bearing on this and that our brains are quite capable of dealing with arbitrary graph structures."

    You wrote: "I agree.  The human brain supports natural languages, which can express anything that can be said in any artificial notation of any kind.  That includes every version of mathematics, symbolic logic, diagrams, the Semantic Web…”

    Please see my last response to Pat.  In it, I try to point out that the “capability” of what the brain can do and the “actuality” of what it actually does are two very distinct things.  Aside from genetic encoding, our brains are born empty… without the understanding of Colors like Red or Blue, without the understanding of Fruits and Vegetables, and without the relationships that allow us to understand that many Fruits known as Apples are Red (but not all).  What things we learn and relationships we establish are only those we collect and create, over time.

    I would also present the theory of a Node that points to everything in the universe.  One faction would argue that no such Node exists.  Other factions might argue that such a Node is called “God” (what I call the “The God Node”).

    Thanks, again, for the contributions and the references.

    My Best,

    Frank
    --
    Frank Guerino, Chairman
    The International Foundation for Information Technology (IF4IT)
    http://www.if4it.com
    1.908.294.5191 (M)



    --
    All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontolog Forum content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
    ---
    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 post to this group, send email to ontolo...@googlegroups.com.

    Gary Berg-Cross

    unread,
    Apr 28, 2016, 3:10:21 PM4/28/16
    to ontolog-forum

    When the conversation gets to discussing a god node it may have gone an abstraction too far.

    Gary

    To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/D347CA7C.149C1F%25Frank.Guerino%40if4it.com.

    Frank Guerino

    unread,
    Apr 28, 2016, 3:35:12 PM4/28/16
    to Ontolog Forum
    Hi Gary,

    Regardless of the religious/spiritual implications or its possible labels, I believe the question is still a valid one:  "Is it possible to have a single Node that points to everything in the universe?”

    This, in turn, comes with other questions: “If so, is it even worth trying to implement such a Node?” or “What would the benefit of implementing such a Node be?”

    I personally don’t believe there is any benefit (even if it does exist).

    My Best,

    Frank
    --
    Frank Guerino, Chairman
    The International Foundation for Information Technology (IF4IT)
    http://www.if4it.com
    1.908.294.5191 (M)

    From: Gary Berg-Cross <gberg...@gmail.com>
    Reply-To: Ontolog Forum <ontolo...@googlegroups.com>
    Date: Thursday, April 28, 2016 at 3:10 PM
    To: Ontolog Forum <ontolo...@googlegroups.com>
    Subject: Re: [ontolog-forum] Are Relationship Triples also Nodes?

    Simon Spero

    unread,
    Apr 28, 2016, 3:46:57 PM4/28/16
    to ontolog-forum

    Taking rdfs as an example:

    ICEXT(I(rdfs:Resource)) = IR

    Simon

    --
    All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontolog Forum content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
    ---
    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 post to this group, send email to ontolo...@googlegroups.com.
    Visit this group at https://groups.google.com/group/ontolog-forum.
    To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/D347DCC0.149DCC%25Frank.Guerino%40if4it.com.

    John F Sowa

    unread,
    Apr 28, 2016, 4:04:33 PM4/28/16
    to ontolo...@googlegroups.com
    On 4/28/2016 3:02 PM, Frank Guerino wrote:
    > Bodies are nothing more than bio-machines (i.e. Tools) for minds.
    > We’re figuring out, every day, how to decompose and mimic these
    > things, /digitally + mechanically/, every day.

    Nobody has a clue about how to achieve fully human ability.
    For over half a century, the best estimate has always been
    "in about 50 years". That is still the median estimate today.
    My more realistic guess is "sometime in the 22nd century".
    See http://www.jfsowa.com/talks/nlu.pdf

    > For example, a couple of days ago I watched a news clip on
    > Sign Language Gloves that when worn during signing allow direct
    > wireless control of a voice translator to pure English.

    Translating ASL (American Sign Language) to English is somewhat
    easier than translating one NL to another. As anyone who has
    used Google translate knows, there's a long way to go before
    we get FAHQMT (Fully Automatic High-Quality Machine Translation).
    See nlu.pdf

    > I try to point out that the “capability” of what the brain can do
    > and the “actuality” of what it actually does are two very distinct
    > things.

    Yes. The capability of the hardware by 2050 might be able to run
    a fully human simulation -- but *only if* somebody knew how to
    program it to achieve that.

    Please note: The nematode C. elegans has a total of only 302
    neurons. The "wiring diagram" of C. elegans is completely known.
    It's probably possible to simulate its neural network with today's
    hardware -- but nobody knows where to begin. Neuroscientists
    today do not fully understand the workings of a single neuron.
    http://wormatlas.mobify.me/hermaphrodite/nervous/Neuroframeset.html

    The so-called Deep Neural Networks are just beginning to scratch
    the surface. And their "artificial neurons" are simple switches
    that have no resemblance to real neurons.

    And by the way, don't believe the hype you hear about neural networks.
    That hype machine is designed to generate funding. It's unrelated
    to what neuroscientists know.

    John

    Pat Hayes

    unread,
    Apr 28, 2016, 8:38:43 PM4/28/16
    to ontolog-forum, Frank Guerino
    On Apr 28, 2016, at 11:30 AM, Frank Guerino <Frank....@if4it.com> wrote:

    Hi Pat,

    Unfortunately, the word “Node” gets overused and lost in the translation, sometimes.  

    Amen to that :-)

    Let me clarify a little more by identifying types of Nodes.

    Whoa. First, tell us what you mean by "node". I have been assuming that you were referring to some kind of graph-based notation for expressing assertions. There are many of these notations, of course, ranging from Pierce's existential graphs through various kinds of semantic network, conceptual graphs, concept maps to contemporary notations like RDF or UML. Details vary, and some are more expressive than others, but the idea basic to all of them is that a node-edge-node triple amounts to an atomic sentence asserting that a binary relation holds between two things. Your examples seem to fit this interpretation also. 

    If this is more or less right, then in such a graph the *syntactic* roles of nodes and edges are fixed by the topology of the graph structure itself. So what do you mean by a "type" of node? What distinguishes one graph node as being of a different type from another?  

     Some examples include:

    • A “Type or Classification Node”.  A Type Node points to one or more Type Instance Nodes.  E.g. “Fruits” is a type (as opposed to "Vegetables”)and will point to all Fruits in the graph.
    When you say that this node "points to" other nodes, do you mean that they are linked by an edge in the graph? If this is your intended meaning of "points to" then most of our earlier conversation has been totally confused :-)

    RDF (actually RDFS) does something similar using rdf:type as the edge label, but in that case the arc is directed from the individuals to the type node (called a 'class' in RDFS and OWL.)

    • A “Type Instance Node”.  This is a specific Instance of a Type Node.  A very  specific Apple is an example.
    • An "Edge (or Arc) Node”.  Examples is a “Subject Matter Expert” which can exist as an Edge/Arc between a Person Type Instance Node and a Topic Type Instance Node.
    One type of "Semantic Relationship” is a bridge between Type Instance Nodes, via an Edge Node.  For example:

    E.g Person Doe, Jane is a Subject Matter Expert to Topic XYZ

    There is a standard terminology here which you could use to avoid misunderstanding. This is a *sentence*, which is typically being *asserted*, ie claimed to be true; the word or words labeling the edge name a *relation* which (the sentence asserts) holds between two *individuals* (the things named by the names on the first and last nodes.) (Sometimes people use the terminology "proposition", or even "fact", instead of "sentence", but these usages are not strictly correct and can give rise to a lot of needless confusion and pointless debate.) Your observation that an edge is a type of node amounts to the claim that relations can be treated as individuals, a long-standing point of debate among logicians. 

    Another type of a "Semantic Relationship” is a bridge between two Edge Nodes, via another Edge Node.  For example:

    E.g. A Subject Matter Expert is related to a SME Definition Node which helps support and define the relationships between one SME and all other Subject Matter Experts.

    You wrote: “If your nodes and edges are part of a graph, and if "point to" is understood in the above sense, then an edge can only point to a single node."

    I don’t’ know if I agree with this statement.

    I now think we were misunderstanding one another. 

     In the above two examples of Semantic Relationships, we can get to

    • Former Example: the two Type Instances Nodes in that specific Subject Matter Expert instance
    • In The Latter Example: all other Subject Matter Experts with the help of a Definition Node (a whole other topic of conversation because: Do we get to other SMEs through or with parallel use of the Definition Node?).

    However, going back to my earlier statement… I believe that the Edge Node “Subject Matter Expert” can point to

    There are at least three senses of "point to" which could be understood here. If we have two nodes A and B, linked by an edge with the label R, ie the sentence A R B in graph form, we could say 
    (1) the node A points to B
    (2) the edge labelled with R points to B
    (3) R points to B
    (4) the entire triple A R B points to B
    (5) the name R can be used to access some information from elsewhere (ie not in this triple at all) which it "points to". 

    I was, earlier, understanding "points to" in sense 5 when discussing RDF, then when trying to understand your usage I assumed you meant sense (2). Earlier in this message you seem to be using sense (1) or perhaps (2), but here you seem to be using sense (3) or perhaps (4). Can you clarify more precisely what it is you do mean?

    multiple things but definitely has a bounded/limited set of things it can point to.

    You wrote: "Do you mean that "subject matter expert" is understood to be an indexical?”

    I have never used “indexical” but it appears that the answer to this question is “yes.”  I gather that “Type Nodes” (see definition above) would also be examples of Indexical Nodes.  (Thanks for teaching me this term!)

    I find this very hard to believe. I was referring to the concept described here: https://en.wikipedia.org/wiki/Indexicality 

    Thinking about “Indexical" a bit further… I would suggest that all Nodes are Indexical as all Nodes point to one or more things and an Index is, in short, defined to be something that provides quick and organized access to a limited set of things that are all part of a broader set of things.  In fact, while in theory a single node can exist with no bindings, I don’t believe there is an example (in reality) that doesn’t relate to at least one other node (after we observe/analyze it).

    You wrote: “Relations can relate all kinds of things, and can themselves be related to other things in all kinds of ways. “

    I agree but I would qualify that separating the data structure from reality sheds some light.  A Data Structure that is a Graph can relate anything to anything.  However, in reality and to my knowledge, we have no evidence to prove that everything is related to everything else.

    Well, in the sense that some relation or other holds between them, I would claim that indeed everything IS related to everything else. Cite me two things and I will tell you a relation that holds between them. Now, not all relations are worth writing down or spending any time with, for sure: but to say that there is NO relation between two things is a very strong and actually rather occult claim. 

     Instead, it appears that specific things are related to other specific things.

    I wrote: “Structurally, a Graph allows us the flexibility to bind anything to anything. However, the brain and nature don’t appear to work that way.”  And you responded: “Really? It seems to me that nature has no bearing on this, and that our brains are quite capable of dealing with arbitrary graph structures. Mine is, at any rate :-)”

    My point is that when we learn something (in nature) we only bind what we’ve learned together with other things we’ve learned (all in the past).  The possibility of relating everything to everything else certainly exists (the empty Graph data structure) but do we really relate anything, arbitrarily, to any and all other things before we learn/derive them?  

    The relations hold whether or not we describe, or even know, them. Jupiter is orbiting the sun, and it was doing that long before anyone or anything knew or thought about it, and likely will be doing so long after all life on Earth is extinct. 

    Pat

    --
    All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontolog Forum content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
    ---
    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 post to this group, send email to ontolo...@googlegroups.com.
    Visit this group at https://groups.google.com/group/ontolog-forum.

    For more options, visit https://groups.google.com/d/optout.

    Jack Park

    unread,
    Apr 28, 2016, 10:53:12 PM4/28/16
    to ontolo...@googlegroups.com
    Hi Frank,

    Full disclosure, I wear a "topic mapper" hat. Through that lens, relationships are first class nodes -- in that parlance, relation nodes are topics as are actor nodes.

    I like to think that every relationship that has any meaning at all -- that is, worthy of representing -- likely has a biography attached to it.

    Cheers
    Jack

    On Thu, Apr 21, 2016 at 9:14 AM, Frank Guerino <Frank....@if4it.com> wrote:
    Hello All,

    I’d like the community's opinions on the following:

    Given a Graph/Network that is composed of Nodes, Edges (a.k.a. Links) and Relationships….
    Defining a Node as an entity that has descriptive traits and state.
    Defining an Edge as the descriptive connection between any two Nodes (e.g. a Predicate)…
    Defining a Relationship as consisting of a Subject Node that is linked through an Edge to an Object Node (a.k.a. a Triple)…

    Do you believe that a Relationship Triple can also be treated as a Node and why?

    If we decompose the Relationship as a set of 3 components, each component can point to one or more things.  For example, if we have a Relationship…

    {
        “Subject Node”: “Tom Smith”,
        “Predicate”: “Product Owner”,
        “Object Node”: “Product ABC”
    }

    1. The Subject Node points to all data about "Tom Smith”.
    2. The Object Node points to all relevant data about "Product ABC”.
    3. The Predicate points to all Relationships that contain the Predicate “Product Owner”.

    (NOTE: Each of the above elements (Nodes & Predicates) can also point to other things like Voids or Holes, even though the above example does not show it.)

    (NOTE: We’ve already proven all of these pointer types in our research so we know they do, in fact, exist and work as described above.)

    While the Relationship can exist as an independent entity, on its own, if it is treated as a Node what else do you think that set would/could point to?  In other words, what can that independent Triple, as a whole and singular entity, also point to?

    Thanks, in advance, for any thoughts you offer.

    My Best,

    Frank
    --
    Frank Guerino, Chairman
    The International Foundation for Information Technology (IF4IT)
    http://www.if4it.com
    1.908.294.5191 (M)

    --
    All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontolog Forum content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
    ---
    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 post to this group, send email to ontolo...@googlegroups.com.
    Visit this group at https://groups.google.com/group/ontolog-forum.

    John F Sowa

    unread,
    Apr 28, 2016, 11:39:05 PM4/28/16
    to ontolo...@googlegroups.com
    Jack and Frank,

    JP
    > I like to think that every relationship that has any meaning
    > at all -- that is, worthy of representing -- likely has
    > a biography attached to it.

    I agree. And Common Logic treats everything as a first-class entity.
    For any x, you can use x as an individual, a relation, or a function.
    IKL goes one step further by adding metalanguage to Common Logic.

    If you like graph or network notations, Peirce's existential graphs
    can be used to represent full CL or eve IKL. Peirce's EGs with
    CL or IKL semantics are a superset of Kamp's DRS notation for
    natural language semantics. For any overview, see the intro
    to EGs: http://www.jfsowa.com/talks/egintro.pdf

    If you prefer Topic Map notation, you can map the EG representation
    to a TM-like style -- that gives you Kamp's style of NL semantics
    plus the CL treatment of relations as first-class entities, plus
    the IKL metalanguage features. And you can adapt it to any graph
    notation you prefer.

    FG
    > Let me clarify a little more by identifying types of Nodes. Some
    > examples include:
    >
    > * A “*Type or Classification Node*”. A Type Node points to one or
    > more Type Instance Nodes. E.g. “Fruits” is a type (as opposed to
    > "Vegetables”)and will point to all Fruits in the graph.
    > * A “*Type Instance Node*”. This is a specific Instance of a Type
    > Node. A very specific Apple is an example.
    > * An "*Edge (or Arc) Node*”. Examples is a “Subject Matter Expert”
    > which can exist as an Edge/Arc between a Person Type Instance Node
    > and a Topic Type Instance Node.

    There is nothing to clarify. A variable or name in logic can
    be used for any or all of those purposes (and many more).

    CL or IKL let you do that.

    John

    Edward Barkmeyer

    unread,
    May 1, 2016, 2:34:31 PM5/1/16
    to ontolo...@googlegroups.com

    At the risk of entering into this discussion a week late:

     

    What Hans says here is very confused.  It is the assertion that the earth is flat that is temporal in nature.  That assertion was commonly, but by no means universally, held for many centuries, but no temporal part of the Earth was ever flat.   I frankly do not believe (!) that ISO 15926 provides the mechanism for dealing with belief, nor that it was ever intended to.

     

    OTOH, Hans is quite right (in spirit) when he says:

    "We don't believe things because they are true, things are true because we believe them."

    I would substitute “perceive” for “believe”.  Ultimately what modelers, analysts, design engineers, and plant managers state as “facts” are their perceptions of reality, based on whatever science, measurement, mathematics, and other knowledge systems they use.  Knowledge engineering is not about absolute truth, but rather about the capture of available “knowledge”, all of which is based on human cognition.  The purpose of recording these “facts” is to enable inferences and to guide action.  And it all comes down to “doing the best we can with what we know”.

     

    As Chris and Pat pointed out, when we expressly want to talk about situations that we don’t necessarily take to be “fact” in this usual sense, we are off into another logical world.  But reifying relationships per se is just a possible formal representation of a “fact” as a “state of affairs” object that is taken to exist, which is the Tarski interpretation of the corresponding relationship proposition being True. 

     

    And OBTW, in that view, a “reference to a proposition” may be construed as a reference to a category (subclass) of “state of affairs” that may or may not be empty, and may in some cases have multiple instances. 

     

    -Ed

     

    From: ontolo...@googlegroups.com [mailto:ontolo...@googlegroups.com] On Behalf Of HansTeijgeler
    Sent: Sunday, April 24, 2016 1:53 PM
    To: ontolo...@googlegroups.com
    Subject: Re: [ontolog-forum] Are Relationship Triples also Nodes?

     

    Hi Frank,

    My Best,

     

    Frank

    --
    Frank Guerino, Chairman
    The International Foundation for Information Technology (IF4IT)
    http://www.if4it.com
    1.908.294.5191 (M)

     

     

     

    From: HansTeijgeler <hans.te...@quicknet.nl>


    Reply-To: Ontolog Forum <ontolo...@googlegroups.com>
    Date: Saturday, April 23, 2016 at 3:50 PM
    To: Ontolog Forum <ontolo...@googlegroups.com>

    Subject: Re: [ontolog-forum] Are Relationship Triples also Nodes?

     

    Hi Frank,

    --

    All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontolog Forum content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
    ---
    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 post to this group, send email to ontolo...@googlegroups.com.
    Visit this group at https://groups.google.com/group/ontolog-forum.


    For more options, visit https://groups.google.com/d/optout.

     

    --

    All contributions to this forum by its members are made under an open content license, open publication license, open source or free software license. Unless otherwise specified, all Ontolog Forum content shall be subject to the Creative Commons CC-BY-SA 4.0 License or its successors.
    ---
    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 post to this group, send email to ontolo...@googlegroups.com.
    Visit this group at https://groups.google.com/group/ontolog-forum.

    Christopher Menzel

    unread,
    May 1, 2016, 2:57:37 PM5/1/16
    to ontolo...@googlegroups.com
    On 1 May 2016, at 1:34 PM, Edward Barkmeyer <ebark...@thematix.com> wrote:
    At the risk of entering into this discussion a week late:
     
    What Hans says here is very confused.  It is the assertion that the earth is flat that is temporal in nature.  That assertion was commonly, but by no means universally, held for many centuries, but no temporal part of the Earth was ever flat.   I frankly do not believe (!) that ISO 15926 provides the mechanism for dealing with belief, nor that it was ever intended to.
     
    OTOH, Hans is quite right (in spirit) when he says:

    "We don't believe things because they are true, things are true because we believe them."

    I would substitute “perceive” for “believe”.  Ultimately what modelers, analysts, design engineers, and plant managers state as “facts” are their perceptions of reality, based on whatever science, measurement, mathematics, and other knowledge systems they use.  Knowledge engineering is not about absolute truth, but rather about the capture of available “knowledge”, all of which is based on human cognition.  The purpose of recording these “facts” is to enable inferences and to guide action.  And it all comes down to “doing the best we can with what we know”.

    No Ed, not you! :-) Please oh please let us not pervert and bastardize the word "true" by giving any credence whatever to the silly aphorism above! And note — you don't! Your point has to do with the words "fact" and "knowledge", and that is a very different matter altogether. Now, I for one, would prefer to be a hard liner about those words as well and insist that a fact is simply a true proposition and that one can only know things that are true, but the fact is :-) that "fact" in KE does tend to mean simply a belief based upon what one has observed or accepted from a reliable source and that such "facts" constitute the "knowledge" about a given domain. But "fact" and "knowledge" in these senses are not absolute — the "facts" are acknowledged to be revisable and, hence, it is acknowledged that "fact" really only means something like "warranted belief" and that one's "knowledge" is simply a collection of such beliefs. Truth, however, is another matter entirely; truth is independent of our ability to know it and it seems to me that we must keep that as an irrevocable backdrop of our inquiry. We strive for truth, and that's what keeps us honest about our "facts" and what underscores the importance of revising them as the world dictates.

    -chris

    Obrst, Leo J.

    unread,
    May 1, 2016, 6:13:03 PM5/1/16
    to ontolo...@googlegroups.com

    The main problem as I see it is that we only see ontology by way of epistemology. I.e., things exist however they exist, but we only see them by perception and then belief based on perception, cogitation, etc. However, the things don’t change. Our descriptions of them do.

    John F Sowa

    unread,
    May 1, 2016, 10:15:39 PM5/1/16
    to ontolo...@googlegroups.com
    Chris, Ed, and Leo,

    This thread is mixing philosophical and practical issues.

    CM
    > Please oh please let us not pervert and bastardize the word "true"
    > by giving any credence whatever to the silly aphorism [quoted by Ed].
    > ...
    > I for one, would prefer to be a hard liner about those words as
    > well and insist that a fact is simply a true proposition and that
    > one can only know things that are true...

    I agree. Unfortunately, the world tends to be mushy or fuzzy.

    CM
    > it is acknowledged that "fact" really only means something like
    > "warranted belief" and that one's "knowledge" is simply a
    > collection of such beliefs.

    Yes. I made some related points in an article on "What is
    the source of fuzziness" for Lotfi Z's Festschrift:
    http://www.jfsowa.com/pubs/fuzzy.pdf

    Note Figure 3 (page 5) about relating the world to a theory
    by means of a Tarski-style model. Every axiom of the theory
    may have denotation True in terms of the model. But it is
    extremely rare for any model to be an exact approximation to
    any part of the physical world (or universe).

    Leo
    > The main problem as I see it is that we only see ontology by way
    > of epistemology...

    But it's important to clarify the difference between the two.
    In terms of Figure 3 in fuzzy.pdf, a formal ontology is a
    theory, as shown on the right. The truth or falsity of any
    statement in the theory is determined by a Tarski-style
    evaluation in terms of a model.

    Epistemology addresses the issues about relating the model
    to some part or aspect of the world on the left of Fig. 3.

    Scientific methodology is a branch of applied epistemology,
    which is devoted to analyzing, quantifying, and reducing
    the errors between the model and the world.

    John

    Bruce Schuman

    unread,
    May 1, 2016, 11:49:44 PM5/1/16
    to ontolo...@googlegroups.com

    > http://www.jfsowa.com/pubs/fuzzy.pdf

     

    Fascinating article, reminds me of why I spent 30 years under the spell of "Conceptual Structures"

     

    Of course, I am always tested by a commitment to top-down ("drill-down") intentional linearity, by which I affirm the claim that language is no more fuzzy than the accuracy of our measuring instrument and the persistence of our commitment to be accurate.  Stipulative definition turns the empirical bottom-up approach to everything on its head.  Just gazing around yourself with no method -- how could your thinking be other than fuzzy? 

     

    >  Unfortunately, the world tends to be mushy or fuzzy.

     

    Well -- is it the world that is fuzzy....  I think the exact quote is -- "the world is a continuum of shadings"

     

    Concepts are inventions of the human mind used to construct a model of the world. They package reality into discrete units for further processing, they support powerful mechanisms for doing logic, and they are indispensable for precise, extended chains of reasoning. But concepts and percepts cannot form a perfect model of the world, -- they are abstractions that select features that are important for one purpose, but they ignore details and complexities that may be just as important for some other purpose. Leech (1974) noted that "bony structured" concepts form an imperfect match to a fuzzy world. People make black and white distinctions when the world consists of a continuum of shadings.       Conceptual Structures, p.344

     

    Yes – but those shadings are cognitive acts, and lines drawn to serve human intention.  We need a mathematical semantics of the bones – that’s the algebraic mystery. 

     

    And that skeletal theory of semantics is clearly driven by human motivation – we draw the lines in particular places “for some reason” – some reason that serves our intention…

     

    In this sense, the world is a continuum (with NO distinctions), and the “shadings” are cognitive constructions created within the perceiving mind, to serve some immediate convenience.

     

    Chapter 2 of CS, “Psychological Evidence”, begins with a discussion of “percepts”.   This book taught me the basics of semantics.  I could never read one word of Chomsky, and never had the slightest clue why anybody thought it was interesting….

     

    p.29 says “An important question is whether perception is a bottom-up process that first matches smaller units and combines them into larger ones [“chunking”] – or whether it is a top-down process that first matches large units and then fills in smaller details.

     

    “Some evidence favors a top-down approach and other evidence favors a bottom-up approach.  Pritchard et al found evidence both for the stability of the whole, as Gestalt theory maintains, and for the existence of parts for building the whole.  They concluded that both approaches “are valid and complement one another.”  This psychological conclusion has also emerged as a basic principle of A.I.: top-down reasoning is a goal-directed process that imposes a tightly-controlled organization [human intentional stipulation!]; bottom-up reasoning is a data-directed [and context-free, yikes] or stimulus-directed process that leads to more diffuse chains of associations.”

     

    Want to know why global civilization is a mess?  It’s them freaking “diffuse chains of association arising spontaneously absolutely anywhere” in crazed and annoyed competition.  The Bible guys called it Babel.  It might be the exact opposite of Logos…

     

    “The two approaches may be combined in bidirectional reasoning, which is originally triggered by some stimulus in the data, but which then invokes a high-level goal that controls the rest of the process.  Some problems are more naturally handled by one approach or the other, but the combination of the two is especially powerful.”   CS, p30

     

    You hit it out of the park with that one.  The entire bony-structured framework is organized on that primary dimension of direction.  Why is that SO hard for the bears to see….

     

    (too much industrial commitment to bottom-up, eh?)

     

    But as I read the “What is the Source of Fuzziness?” PDF – I see what looks like a strong and persuasive clue.  Part 3, Relating Patterns to Patterns says

     

    In science, collections of patterns form theories. In other fields, they are called models, blueprints, project plans, or syndromes. Whatever they’re called, collections of patterns are expressed in notations for which precision is important. Yet scientists are always aware of the experimental error, which they try to limit by carefully controlled experiments. Engineers express their frustration in a pithy slogan: All models are wrong, but some are useful. To bridge the gap between theories and the world, Figure 3 shows a model as a Janus-like structure, with an engineering side facing the world and an abstract side facing a theory.

     

    That is how it works.  See the world that way, and the huge semantic mudslide of global evolution gets washed away.  That’s the jewel in the lotus.

     

    Now – just get the dimensional structure of the bones worked out – and everything devolves to measurement.  Stipulate it (because, like we’ve said 500 times, the parsing is ad hoc) through Q&A dialogue right on down to “acceptable error tolerance” and you’ve worked out your agreement…

     

    A Janus-like structure – what the hippies call “holons” – is a “a part that is also a whole” – facing in two (“bi-directional”) directions – up the part/whole hierarchy to abstract theory, or down, to the concrete world.

     

    We gotta stop arguing this tension (“which side is right”), and get an “all of the above” model figured out.

     

    Bruce Schuman

    http://origin.org/one/spectrum.cfm

     

     

    -----Original Message-----
    From: ontolo...@googlegroups.com [mailto:ontolo...@googlegroups.com] On Behalf Of John F Sowa
    Sent: Sunday, May 1, 2016 7:16 PM
    To: ontolo...@googlegroups.com
    Subject: Re: [ontolog-forum] Are Relationship Triples also Nodes?

     

    Chris, Ed, and Leo,

     

    This thread is mixing philosophical and practical issues.

     

    CM

    > Please oh please let us not pervert and bastardize the word "true"

    > by giving any credence whatever to the silly aphorism [quoted by Ed].

    > ...

    > I for one, would prefer to be a hard liner about those words as well

    > and insist that a fact is simply a true proposition and that one can

    > only know things that are true...

     

    I agree.  Unfortunately, the world tends to be mushy or fuzzy.

     

    CM

    > it is acknowledged that "fact" really only means something like

    > "warranted belief" and that one's "knowledge" is simply a collection

    > of such beliefs.

     

    Yes.  I made some related points in an article on "What is the source of fuzziness" for Lotfi Z's Festschrift:

    http://www.jfsowa.com/pubs/fuzzy.pdf

     

    Note Figure 3 (page 5) about relating the world to a theory by means of a Tarski-style model.  Every axiom of the theory may have denotation True in terms of the model.  But it is extremely rare for any model to be an exact approximation to any part of the physical world (or universe).

     

    Leo

    > The main problem as I see it is that we only see ontology by way of

    > epistemology...

    image001.png
    thespectrum.png

    Frank Guerino

    unread,
    May 2, 2016, 2:00:31 PM5/2/16
    to Ontolog Forum
    Hi Pat,

    You wrote: “Whoa.  First, tell us what you mean by a “node.” … but the idea basic to all of them is that a node-edge-node triple amounts to an atomic sentence asserting that a binary relation holds between two things. Your examples seem to fit this interpretation also.”

    Yes, this is the model we follow.  What you call a Sentence we call a Semantic Relationship.

    You wrote: “If this is more or less right, then in such a graph the *syntactic* roles of nodes and edges are fixed by the topology of the graph structure itself. So what do you mean by a "type" of node? What distinguishes one graph node as being of a different type from another? ”

    Yes, I agree.  However, there are different contexts for which certain Nodes exist.  In the example:

    • A categorization “Apple” points to all Apples in a set but not all other elements in a set.
    • A categorization “Food” points to all things considered as Food in the set and includes Apples.

    In this example categorizations have different purpose than an instance of a specific Apple.  We see them act as type or categorization pointers that point to many things, simultaneously, that are representations of such categorizations.

    It is the same with descriptive Predicates.  They have purpose and can only point to a set of things that are tied the context of their existence.  For example, in the following Semantic Relationship (a.k.a. Sentence):

    • Color Red is Related as a Primary Color to Apple XYZ (where XYZ is a specific Apple)
    The descriptive Predicate “Primary Color”, which is the label of an Edge, can only point to:

    • its Subject (Color Red)
    • Its Object (Apple XYZ)
    • Its Equivalents
      • Other Edges of Label = “Primary Color”
      • Synonyms
      • Preferred Terms
      • Etc.
    • Its !Equivalents
      • Antonyms (non in this case)
      • Etc.

    A node that has a purpose for Categorization exists for a very different purpose than a node that is a descriptive Predicate.  This is what I meant by “types of nodes.”

    You wrote: “There are at least three senses of "point to" which could be understood here. If we have two nodes A and B, linked by an edge with the label R, ie the sentence A R B in graph form, we could say 
    (1) the node A points to B
    (2) the edge labelled with R points to B
    (3) R points to B
    (4) the entire triple A R B points to B
    (5) the name R can be used to access some information from elsewhere (ie not in this triple at all) which it "points to”. “

    Given my description above, I believe we have some fundamental differences.  For example, I believe that:

    (1): a node A can never directly point to B without first going through a labeled Edge.  In other words, it can only point through a descriptive Predicate that provides context/meaning.
    (2) and (3): See my description what a descriptive Predicate can point to, above.

    You wrote: “Well, in the sense that some relation or other holds between them, I would claim that indeed everything IS related to everything else. Cite me two things and I will tell you a relation that holds between them. Now, not all relations are worth writing down or spending any time with, for sure: but to say that there is NO relation between two things is a very strong and actually rather occult claim."

    I agree and I do concede that sometimes we leverage “Implied Predicates” because humans don’t care about seeing certain Predicates.  This is seen in the fact that you can click on a categorization called “Countries” that points directly to Indexes, which allow access to different organized sets of Countries.  However, there are no specific Predicates that bridge between the categorization and the Indexes.  It would not be intuitive or make sense for most people to access Indexes through a descriptive Predicate called “Index Of”.

    My Best,

    Frank
    --
    Frank Guerino, Chairman
    The International Foundation for Information Technology (IF4IT)
    http://www.if4it.com
    1.908.294.5191 (M)




    From: Pat Hayes <pha...@ihmc.us>
    Reply-To: Ontolog Forum <ontolo...@googlegroups.com>
    Date: Thursday, April 28, 2016 at 8:38 PM
    To: Ontolog Forum <ontolo...@googlegroups.com>, Frank Guerino <Frank....@if4it.com>
    Subject: Re: [ontolog-forum] Are Relationship Triples also Nodes?

    Frank Guerino

    unread,
    May 2, 2016, 2:03:22 PM5/2/16
    to Ontolog Forum
    Hi Jack,

    Thanks for the contribution.

    I agree with your view.  As mentioned in my last response to Pat, I agree with him (and you) that not all relationships are worthy of representing.  However, that being said, I’m sure the brain creates and maintains many relationships between things that we would, on the surface, never consider useful.

    My Best,

    Frank
    --
    Frank Guerino, Chairman
    The International Foundation for Information Technology (IF4IT)
    http://www.if4it.com
    1.908.294.5191 (M)

    From: Jack Park <jack...@gmail.com>
    Reply-To: Ontolog Forum <ontolo...@googlegroups.com>
    Date: Thursday, April 28, 2016 at 10:53 PM
    To: Ontolog Forum <ontolo...@googlegroups.com>
    Subject: Re: [ontolog-forum] Are Relationship Triples also Nodes?

    John F Sowa

    unread,
    May 2, 2016, 10:55:34 PM5/2/16
    to ontolo...@googlegroups.com
    On 5/1/2016 11:49 PM, Bruce Schuman wrote:
    > Fascinating article, reminds me of why I spent 30 years
    > under the spell of "Conceptual Structures"

    I'm happy that you were impressed by the book.

    >> "People make black and white distinctions when the world consists
    >> of a continuum of shadings." Conceptual Structures, p.344
    >
    > Yes – but those shadings are cognitive acts, and lines drawn to serve
    > human intention. We need a mathematical semantics of the bones –
    > that’s the algebraic mystery.

    Those "bones" are the subject matter of physics. Nearly every
    branch of mathematics has been used to represent theories of
    physics. And many of the most fertile branches of mathematics
    were inspired by and developed in order to represent and solve
    problems in physics and engineering.

    > the world is a continuum (with NO distinctions), and the “shadings”
    > are cognitive constructions created within the perceiving mind

    Although the world is continuous, it is not uniform. There
    are many regions with sharp edges and abrupt changes. Just
    take a walk in the woods, and you'll bump into or fall off
    places where the first derivative is very far from zero.

    Falling off a cliff is an objective reality that other people
    can observe and verify -- even if they failed 9th grade algebra.

    > Want to know why global civilization is a mess?

    There are many reasons. But I doubt that learning more
    algebra -- by itself -- would make much of a difference.

    John
    Reply all
    Reply to author
    Forward
    0 new messages