119 views

### Jon Awbrey

Jun 18, 2018, 3:36:47 PM6/18/18
to SysSciWG, Structural Modeling, Ontolog Forum @ GG
Ontolog, Systems Science, Structural Modeling Groups,

Spent the week moving the first load of our furniture to storage —
and boy! are my armchairs tiring! — so I'll have to pick up the
dangling threads later — but here's a bit of material I meant
to post last week ...

The middle ground between relations in general and the sign relations
we need to do logic, inquiry, communication, and so on is occupied by
triadic relations, also called ternary or 3-place relations.

Triadic relations are some of the most pervasive in mathematics,
over and above the importance of sign relations for logic, etc.

Here's a primer with examples from mathematics and semiotics:

Regards,

Jon

inquiry into inquiry: https://inquiryintoinquiry.com/
oeiswiki: https://www.oeis.org/wiki/User:Jon_Awbrey
isw: http://intersci.ss.uci.edu/wiki/index.php/JLA

### Ravi Sharma

Jun 18, 2018, 6:55:58 PM6/18/18
to ontolog-forum
Jon thanks good info

--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or unsubscribe to the forum, see http://ontologforum.org/info/
--- You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-forum+unsubscribe@googlegroups.com.

--
Thanks.
Ravi
(Dr. Ravi Sharma)
313 204 1740 Mobile

### Jon Awbrey

Jun 19, 2018, 10:36:26 AM6/19/18
Thanks, Ravi,

I chose those examples of triadic relations to be as simple
as possible without being completely trivial but they already
exemplify many features we need to keep in mind in all the more
complex cases as we use relational models of realistic phenomena
and objective domains.

The mathematical examples are typical of many in linguistic, logical,
axioms, definitions, equations, expressions, formulas, predicates, or
terms that denote the relations of interest. For example, we might be
discussing dyadic relative terms like “parent_of___” or “square_of___”
and triadic relative terms like “giver_of___to___” or “sum_of___and___”.

If we spend the majority of our time in contexts like that we may form
the impression that all the relational concepts we'll ever need can be
requisitioned off-the-shelf from pre-fab stock, no assembly required.

That's a pretty picture of our mental equipment. It may even be true
if we cook the data long enough and fudge the meaning of pre-fab down
to the level of amino acids or quarks or some other bosons on the bus.

As a practical matter, however, research pursued in experimental veins
tends to push the envelope of pre-fab concepts into surprisingly novel
realms of ideas.

I'll discuss the examples of sign relations when I get more time ...

Regards,

Jon

### Ferenc Kovacs

Jun 19, 2018, 1:31:40 PM6/19/18
Ravi, Jon,

For people who do not accept that relations in a natural language is represented by a verb, although the very same relation may be described in other linguistic (grammar) structures and who do not believe that the most generic word and unit of expressing ideas is an object grounded in sensory experience and some mental operations as separation and/or isolation (resulting in a name as well as a concept) it may be difficult to see that relations are operations whether they are indicated expressly or hidden otherwise.

Hence instead of triadic relations you have operands and operators where an operator is also a separator, yet altogether resulting in an operation. The most generic operation follows transitivity, in other words if you have a subject that has to be linked to an object reflecting associations between physical things and their names, objects and their properties, etc. in various sequences produced by the operation transit or change).

An SVO sentence pattern is an example of such transitivity in action. A logical phrase SP and SQ are other morphs with the operator disguised. Operators are normally not indicated as in „ab” (axb) or even in additive relations. In addition for example, what you make note of is the sum itself, and not the number of operations (addition) performed in the calculation (because they are the same in numeric terms). This is as important feature, although totally neglected, because it represents a compact way of denoting symbols and their relations. Nevertheless, when making a computer, you cannot disregard keeping track of the number operations performed.

Consequently, you have various kinds of objects all related by a verb that when fully specified, gives you the details of such a relation that enables you to visualize what takes place/is happening/or was the case. Comparison is the most fundamental mental operation that calls for two objects that are comparable on some grounds despite their differences in other respects. In ancient times comparison was physically performed by overlapping the two items in „space” (on a plane in fact, which was the forerunner of division, both a physical and a mental operation that proved to be very useful in a world where the distribution of goodies was a major component of social life.

The fact that you do (can) not keep track of the number of operations (repetitions of the operand and operators at different times and places) has lead us to two extreme solutions, namely the use of quantors without being able to check if any of those quantors are true or make sense. Fuzzy sets explain how you go round that problem, but interestingly enough the use of quantors is so popular that even syllogisms are built on them as some foolproof ground so logic and reasoning despite the fact that syllogism are nothing but a sort operation base on quantity, which may sound provoking but I need to explain it next time, should anyone be interested.

Ferenc Kovacs

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

### Matthew West

Jun 19, 2018, 2:32:17 PM6/19/18

Dear Colleagues,

Ravi, Jon,

For people who do not accept that relations in a natural language is represented by a verb, although the very same relation may be described in other linguistic (grammar) structures and who do not believe that the most generic word and unit of expressing ideas is an object grounded in sensory experience and some mental operations as separation and/or isolation (resulting in a name as well as a concept) it may be difficult to see that relations are operations whether they are indicated expressly or hidden otherwise.

[MW>] The mistake that people often make is to consider that a relation is something that is represented, rather than it being a way of representing something. That is, it is a mathematical structure that can be used in many different ways to represent many different things. It is not a fundamental ontological category.

Regards

Matthew West

### joseph simpson

Jun 19, 2018, 2:42:56 PM6/19/18
It may be useful to talk about "natural language relationships" and "mathematical relations."

A single mathematical relationship may represent many natural language relationships.

A natural language relationship may be used to further define and refine a specific mathematical relation application.

A mathematical relation may be used to further quantify and evaluate a natural language relationship.

Take care and have fun,

Joe

--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-forum+unsubscribe@googlegroups.com.

--
Joe Simpson

# All progress, therefore, depends on unreasonable people.”

George Bernard Shaw

### Ferenc Kovacs

Jun 19, 2018, 3:06:44 PM6/19/18
On Tuesday, June 19, 2018, 8:32:19 PM GMT+2, Matthew West <dr.matt...@gmail.com> wrote:

Dear Matt,
On Relation:  ... It is not a fundamental ontological category.

The three fairly widely recognized categories of an upper ontology are objects, relations and properties. Objects do have the possibility for verification through sensors of all kinds. Relations and properties are abstractions though, therefore they cannot be falsified that easily. It is also true that while objects tend to lend themselves to common classifications based on properties and relations, no such structures are known for properties and relations.

Unfortunately, relations are deemed to be unrelated to verbs and are used in a simplified manner to fit set theory, relations are not treated as complex patterns with all the specifications carried by verbs. Properties on the other hand allow for inversion with objects again to fit set theory interpretation of the world that is more rationally described by using statements and propositions with fully specified verbs.

a relation is something that is represented, rather than it being a way of representing something

A way of representing something (an object for instance) IS REPRESENTING, or performing an operation, not specified in more details, an answer to the how question that calls for a verb, which in turn calls for another verb, or an operation (originally meaning energy) to put in a generic form or relation. We do not know how to break down the web of everything is related to everything else but by comparing objects pairwise. Comparing is relating, and relating is one of the several mental operations that we are not aware of, except for the simplicity of induction, deduction and abduction.
Regards, Ferenc

--

### Jon Awbrey

Jun 19, 2018, 4:24:59 PM6/19/18
Joe,

I'll reply to your remarks first simply because they are the shortest
and they give me the quickest chance of clarifying a few simple points.

I use the word “relation” to mean a special type of mathematical object,
namely, a designated subset included within a cartesian product of sets.

Whatever else this definition of a relation may have going for or against it,
it does single out a class of formal structures that serve in good stead as
intermediary objects between the world of phenomena and our human capacity
for coping with whatever reality may emanate in them. So that is mainly
how I aim to use it here.

When I'm being careful, therefore, I'll try to use words that maintain
a distinction between objects, formal or otherwise, and the symbolic
modifications of media that we use to reference those objects.

For example:

> The mathematical examples are typical of many in linguistic, logical,
> axioms, definitions, equations, expressions, formulas, predicates, or
> terms that denote the relations of interest. For example, we might be
> discussing dyadic relative terms like “parent_of___” or “square_of___”
> and triadic relative terms like “giver_of___to___” or “sum_of___and___”.

I used “axioms, definitions, equations, expressions, formulas, predicates, terms”
along with “dyadic relative terms” and “triadic relative terms” for various sorts
of symbolic entities that serve to denote or describe formal objects of thought
and discussion, while I tried to reserve “relations” for the objects themselves.

Hope that helps to clear things up ...

Regards,

Jon

On 6/19/2018 2:42 PM, joseph simpson wrote:
> It may be useful to talk about "natural language relationships" and
> "mathematical relations."
>
> A single mathematical relationship may represent many natural language
> relationships.
>
> A natural language relationship may be used to further define and refine a
> specific mathematical relation application.
>
> A mathematical relation may be used to further quantify and evaluate a
> natural language relationship.
>
> Take care and have fun,
>
> Joe
>

oeiswiki: https://www.oeis.org/wiki/User:Jon_Awbrey

### joseph simpson

Jun 19, 2018, 5:50:52 PM6/19/18
Jon:

I am using the term "mathematical relation" in a slightly different manner.

For example, a binary relation is a set of ordered pairs of the elements of some other set.

The "part-of" natural language relationship is a well known natural language relationship.

However, is a part allowed to be part-of itself?

In some cases yes, in some cases no.

In the case where a part is not allowed to be part-of itself,
the logical properties for this natural language relationship are:
- irreflexive
- asymmetric
- transitive.

In the case where a part is allowed to be part-of itself,
the logical properties for this natural language relationship are:
- reflexive
- symmetric
- transitive. (equivalence)

or

- reflexive
- asymmetric
- transitive.

These additional logical (mathematical) relation characteristics
provide a mechanism to more clearly communicate the attributes
of the current part-of natural language relationship of interest.

Natural language and mathematics are different language types.

We have created the Augmented Model-Exchange Isomorphism (AMEI)
to support the creation of a catalog of natural language terms and their
isomorphic structured graphics and mathematical forms. See:

### John F Sowa

Jun 19, 2018, 6:13:09 PM6/19/18
Dear Matthew, Jon, Ravi, John Y, and Joe,

In March, I was the respondent for a talk on "The necessity of genuine
triadic relations" by Bill McCurdy at a philosophy convention in
San Diego. Since Bill and I are both Peirce fans, we were in strong

MW (in a response to Jon and Ravi)
> The mistake that people often make is to consider that a relation
> is something that is represented, rather than it being a way of
> representing something. That is, it is a mathematical structure
> that can be used in many different ways to represent many different
> things. It is not a fundamental ontological category.

I agree. To illustrate the issues, note the two different ways
of representing the sentence "Sue gives a child a book" in slide 3

On the left is a conceptual graph (CG), which represents the
verb 'gives' by a triadic relation named Gives. On the right is
a CG with a concept of type Give, which is linked by three dyadic
relations (Agent, Recipient, and Theme) to concepts for Sue, a
child, and a book.

This example illustrates three points:

1. For linguists, you can choose to represent verbs by a relation
or to a node that could represent the verb 'give' or the gerund
Giviing. The option depends on whatever ontology you choose.

2. For logicians, it illustrates the point that any triadic relatioin
can be replaced by three dyadic relations.

3. And for graph theorists, it shows that there is an unavoidable
triadic connection that will always be there, no matter which
choice of ontology you happen to choose.

JA
> The middle ground between relations in general and the sign relations
> we need to do logic, inquiry, communication, and so on is occupied by
> triadic relations, also called ternary or 3-place relations.

I agree. But if you say that to a logician, you'll get the response,

See slide 2 for the kind of response that Peirce, Bill McCurdy, and I
would make to that point: Peirce used the term 'triad', not triadic
the graph will still contain irreducible triadic nodes. In existential
graphs, you will get a "teridentity", which consists of three lines
linked together. In CGs, you get a concept node with three links.

No matter how you transform the graph, you'll always have something

MW
> Existence here is not about whether you can kick something, but

I agree. And that raises many important issues about logic and
ontology that would take too long to explain in this note.

Short summary: languages and logics consist of signs, and some
of those signs express propositions, some of which are laws.
For more, see http://jfsowa.com/pubs/signs.pdf

JY
> I find that having separate classes for Information Entity and
> Information Realization helps to know what I am representing.
> The fact that it is dependent continuant helps but may not be
> the dominant characteristic. So there is nothing preventing
> addition of these classes as subclasses in BFO?

Barry classified information artifacts as dependent continuants.
That makes the information contained in the book "depend" on
the printed or otherwise represented copies.

But as I said in a previous note, that is a half-vast theory of
signs. It lets you talk about sign tokens (copies of information),
but not about sign types (the kinds of information contained in
those sign tokens). Philosophers from antiquity to the present
have gone into huge amounts of detail about signs and systems of
signs such as languages. See http://jfsowa.com/pubs/signs.pdf

In that article, I quote the philosopher David Armstrong, who
mentioned a “distinction that practically all contemporary
philosophers accept... It is the distinction between token and
type” by Charles Sanders Peirce.

Joe S
> It may be useful to talk about "natural language relationships"
> and "mathematical relations." A single mathematical relationship
> may represent many natural language relationships. A natural
> language relationship may be used to further define and refine
> a specific mathematical relation application. A mathematical
> relation may be used to further quantify and evaluate a natural
> language relationship.

Yes. There are many ways to translate an NL text to some version
of logic for different purposes in different contexts. Slide 2
of triads.pdf is just one example.

John

### Jon Awbrey

Jun 29, 2018, 11:44:15 AM6/29/18
Cf: Sign Relations, Triadic Relations, Relations • 6

On 6/19/2018 5:50 PM, joseph simpson wrote:
JS:
> Jon:
>
> Thanks for the additional information.
>
> I am using the term "mathematical relation" in a slightly different manner.
>
> For example, a binary relation is a set of ordered pairs of the elements of
> some other set.

That is the first definition I learned for binary relations.

Slightly more generally, a binary relation L is a subset of
a cartesian product X × Y of two sets, X and Y. In symbols,
L ⊆ X × Y. Of course X and Y could be the same, but that's
not always the case.

pretty much interchangeably in application to relations, but
I developed a bias toward dyadic on account of computational
contexts where binary is reserved for binary numerals.

Once again, partly due to computational exigencies, I would
now regard this first definition as the weak typing version.

The strong typing definition of a k-place relation includes
the cartesian product X_1 × … × X_k as an essential part of
its specification. This serves to harmonize the definition
of a k-place relation with the use of mathematical category
theory in computer science.

When I get more time, I'll go through the material I linked
on relation theory in a slightly more leisurely manner ...

• Relation Theory
http://intersci.ss.uci.edu/wiki/index.php/Relation_theory
https://en.wikiversity.org/wiki/Relation_theory

Regards,

Jon

JS:

### joseph simpson

Jun 29, 2018, 12:44:48 PM6/29/18
to Jon Awbrey, structura...@googlegroups.com, Sys Sci, Ontolog Forum @ GG
Jon:

I think we are generally on the same page with respect to the definition of a "mathematical relation."

Another definition of mathematical relation from "Constraint Theory - Multidimensional Mathematical Model Management"  by George Friedman, 2005, is:

".. a relation between a set of variables is defined as that subset within the product set of the variables which satisfies that relation."

A relation between or among a set of variables is not restricted to binary relations (as you point out.)

Further, a relation is a constraint on the variation of the variable values.

Viewing a relation as a constraint on variation directly aligns the definitions of a relation and a system.

A system is defined as a constraint on variation.

Take care and have fun,

Joe

### joseph simpson

Jun 29, 2018, 3:25:37 PM6/29/18
to structura...@googlegroups.com, Sys Sci, Ontolog Forum @ GG, Jon Awbrey, mjs...@gmail.com
Jack stated:

"How do you confirm that your presumed constraint makes the resulting system Fit For Purpose? Friedman did."

Jack, great question and observation.

It will take some context building to create an environment where the question can be answered.

However, the primary value and operational purpose, I see in Constraint Theory, is the reduction in resources needed to evaluate and analyze any given collection of relations (equations.)

Given a collection of equations that describe systems and/or the needed integrated system behavior, Constraint Theory may be used as a guide to develop an efficient empirical data collection process that supports a specific system design process.

It might be useful to review Chapter 1 in the Constraint Theory book which places Constraint Theory in the realm of a decision support tool.  General information needed to design and produce the product system of interest is encoded into mathematical relations.  I have been thinking about discussing the first chapter, "Motivations," on the Structural Modeling and System Science lists as I believe that George's work is one of the next steps past the ISM software developed by Warfield.

Constraint Theory works with collections of relations.

One of the first steps in the analysis of Chapter 1, Motivations is the categorization of the relation types given in the example of low dimension.

Two relation categories will be used, linear relations and  non-linear relations.

In any case, my plan is to address these issues as the structural modeling material is developed.

Now I will address this area sooner, rather than later.

Bottom line: it is all about reducing cost by applying logical analysis.

Take care, be good to yourself and have fun,

Joe

On Fri, Jun 29, 2018 at 10:20 AM, Jack Ring wrote:
How do you confirm that your presumed constraint makes the resulting system Fit For Purpose? Friedman did.

--
You received this message because you are subscribed to the Google Groups "Structural Modeling" group.
To unsubscribe from this group and stop receiving emails from it, send an email to structural-modeling+unsub...@googlegroups.com.
To post to this group, send email to structural-modeling@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Structural Modeling" group.
To unsubscribe from this group and stop receiving emails from it, send an email to structural-modeling+unsub...@googlegroups.com.
To post to this group, send email to structural-modeling@googlegroups.com.

### John F Sowa

Jun 29, 2018, 10:17:26 PM6/29/18
On 6/29/2018 11:44 AM, Jon Awbrey wrote:
> a binary relation L is a subset of a cartesian product X × Y
> of two sets, X and Y. In symbols, L ⊆ X × Y.

That is a purely extensional definition. It doesn't allow you
to distinguish relations that happen to have different intensions,
but the same extension.

For example, people used to think that the relation swan(x) and the
relation white-swan(x) were identical because they never saw a swan
of any other color. But when they discovered black swans, they had
to recognize that the two definitions represented distinct sets.

More generally, a definition of a relation R(x1,...,xN) by intension
is a rule or criterion for determining whether R is true of some
N-tuple. But a definition by extension is just a set of N-tuples.

You can have multiple relations with distinct intensions, but the
same extension. For more about intensions and extensions, the
logician Alonzo Church wrote a very nice three-page discussion:
http://jfsowa.com/logic/alonzo.htm

In those three pages, Church talks about definitions of functions,
but the same principle holds for definitions of relations.

John

### joseph simpson

Jun 30, 2018, 10:20:45 AM6/30/18
to structura...@googlegroups.com, Sys Sci, Ontolog Forum @ GG, Jon Awbrey, mjs...@gmail.com
Jack:

In "Constraint Theory," Chapter 1 Motivations, page 4  an example of low dimension is given.

The context for the example is  "A decision-making manager authorized to initiate the preliminary design of a new system..."

In this example the chief systems engineer defined a "total systems optimization criterion, T."

T is given by the mathematical equation:

T = PE/C

where:
P = political index of acceptability
E = system effectiveness
C = life cycle cost

A few items of interest at this point.

The item of interest (T) is known and encoded in a mathematical relation.
Unlike the initial context in structural modeling (Warfield's work), where the item of interest is unknown and a natural language relationship is used by a group of individuals to determine the structure of the item of interest.

The initial item is described at a high level of abstraction using existing knowledge and a formal mathematical equation.

The item description is further elaborated by providing additional mathematical equations that decrease the level of abstraction as well as add additional existing knowledge to the item of interest description.  Table 1-1, page 5 provides a list of equations and variables.

Table 1-1 lists six (6) equations [three (3) linear equations and three (3) nonlinear equations.].

Table 1-1 lists eight (8) variables, however two (2) variables used in the equations appear to be missing.  These variables are 1) X and 2) Dmax.

On page six (6) the mathematical model is described as having six (6) equations and eight (8) variables.

It appears that the model has six (6) equations and ten (10) variables.  This is a point that is open for discussion. Why were X and Dmax not included in the model analysis.

Page 6 also states, ".. there should be two  "degrees of freedom.""

Even with the term "degrees of freedom" in quotes it is not clear how the concept of degrees of freedom applies to a mathematical model that is composed of both linear and nonlinear equations.

Clearing up the number of variable in the example model as well as addressing the specific use of "degrees of freedom" in this case will help create a common context to continue the discussion of Constraint Theory.
To unsubscribe from this group and stop receiving emails from it, send an email to structural-modeling+unsubscribe...@googlegroups.com.

To post to this group, send email to structural-modeling@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Structural Modeling" group.
To unsubscribe from this group and stop receiving emails from it, send an email to structural-modeling+unsubscribe...@googlegroups.com.

To post to this group, send email to structural-modeling@googlegroups.com.

--
Joe Simpson

# All progress, therefore, depends on unreasonable people.”

George Bernard Shaw

### joseph simpson

Jul 1, 2018, 1:21:10 PM7/1/18
to structura...@googlegroups.com, Sys Sci, Ontolog Forum @ GG, Jon Awbrey, mjs...@gmail.com
Jack:

The questions related to the number of variables and the application of the "degrees of freedom" may be put aside as a distraction, for awhile.

Another way of looking at this specific knowledge encoding problem is to identify and apply a structuring natural language relationship that may provide some insight into the structure of the group of equations.

The current problem space is populated with two basic object types, 1) variables and 2) equations.

The variables are a basic type not composed of any other types.

The equations are a composite type composed of variables, constants, numeric values and mathematical operators.

We will focus on the basic type, variables.

What is the structuring relationship of interest between and among the given variables?

There are a few candidate structuring natural language relationships that could be used, but "dependent-on" was selected for the initial analysis.

The natural language relationship dependent-on has the following logical properties:
- irreflexive
- asymmetric
- transitive

Starting with "T" at the top of a sheet of paper, a directed graph of the variable dependencies may be constructed.

Once the variable dependencies are encoded in this manner, recursive cycles of dependencies appear in some variable combinations.

If you trace the dependency thread from the T at the top of the page down a single thread to a terminal variable, without encountering a "dependency cycle", then you arrive at a variable terminal set of P, M, A, Dmax and X.

The three variables, P, M, and A are included in the book as part of the allowable computation group.  T = f(E,P), T = f(M,P) and T = f(A,P).

Due to the fact that X and Dmax were not included in the variable set in the book, they are not addressed.

It should be easy to check the and see if X and Dmax are, in fact allowed in the computations.

In any case, this is a direct connection between the work of John and George.

There are some simple evaluation and parsing rules that might be applied to this general "dependency tree" algorithm to engage a wide range of computational system types.

By selecting the proper natural language relationship, the structure of a group of equations may be evaluated and analyzed.

In this case, the components of the total system structural graph that are directed acyclic graphs appear to hold the key to "computational allowability".

The components of the total system structure that create dependency cycles appear to hold the key to computational constraint generation.

Mathematics provides a wide range of valuable knowledge encoding techniques.

Using structural modeling to more closely align natural language relationships  with formal mathematical relations has the capability to provide insight and increase communication in a number of areas.

### joseph simpson

Jul 1, 2018, 1:54:14 PM7/1/18
to structura...@googlegroups.com, Sys Sci, Ontolog Forum @ GG, Jon Awbrey, mjs...@gmail.com
Jack:

The concepts and terms were selected to support the primary context of "Triadic Relations."

The selected terms are the terms used by George in Constraint Theory.

A system is a group of objects organized by a given relationship.

In this case, the selected objects are "variables."

In this case, the natural language organizing relationship is "dependent on."

Take care and have fun,

Joe

On Sun, Jul 1, 2018 at 10:31 AM, Jack Ring wrote:
Joseph,
Rather than variables and equations the fundamentals may be Operands and Operators. Each can be simple, compound or complex. What you describe strikes me as a special case, not a system perspective.
Is this possible?
Jack

To unsubscribe from this group and stop receiving emails from it, send an email to structural-modeling+unsub...@googlegroups.com.

To post to this group, send email to structural-modeling@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Structural Modeling" group.
To unsubscribe from this group and stop receiving emails from it, send an email to structural-modeling+unsub...@googlegroups.com.

To post to this group, send email to structural-modeling@googlegroups.com.

### joseph simpson

Jul 2, 2018, 2:06:28 PM7/2/18
to structura...@googlegroups.com, Sys Sci, Ontolog Forum @ GG, Jon Awbrey, mjs...@gmail.com
Jack:

A brief discussion of some aspects of the example of low dimension from Chapter 1 of "Constraint Theory," is included in Technical Report 23.

Technical Report 23 is scheduled to be posted to Research Gate in a day or two.

A key, fundamental aspect of the example evaluation is the language used in the evaluation and communication of the evaluation results.

The referenced example has many specific issues when evaluated in formal mathematical terms.

For example, are the six presented equations in the collection of equations.

Two of the equations appear to address the same variable.

Does this make a duplicate entry in the collection of equations and therefore not a set?

The collection of equations contains both linear and nonlinear equations.

Does this mix of equation types matter when formal mathematics is used to evaluate the collection of equations?

If we use mathematics as the object language of general systems and natural language as the metalanguage to discuss the system object language, then these questions may need to be answered using a language other than mathematics.

Structural modeling uses natural language to evaluate system structure.

In this case, all of the listed issues associated with formal mathematical evaluation of the collection of equations become irrelevant when the "depends on" natural language relationship is used to analyze the collection of equations.

The missing variables are highlighted in this natural language analysis.

The duplicate variable equations are combined into one factor.

The composite variable E is highlighted and shown to be comprised of other variables.

In my mind it more about language type than the level of abstraction associated with the terms.

In this specific case we have associated logical properties with our natural language relationship.

Correct logic has always been a feature of proper, effective rhetoric.

Logic appears to be an area that spans mathematics and natural language.

In any case, the brief evaluation of Constraint Theory in Technical Report 23, is just a start.

Take care, be good to yourself and have fun,

Joe

On Mon, Jul 2, 2018 at 10:13 AM, Jack Ring wrote:
Joe,
Be clear that George was addressing a model as in a simulator so the terms he selected make sense in that context.
When addressing a general model of system structure I suggest using more generic terms.
Jack

### Jon Awbrey

Jul 6, 2018, 11:20:19 AM7/6/18
Ferenc, Ontologists ...

Our thoughts live in natural and artificial languages the
way fish swim in natural and artificial bodies of water.

One of the lessons most strikingly impressed on me by my first year physics course
or the mass of collateral reading I did at the time was to guard against the errors
that arise from “projecting the properties and structures of any language or symbol
system on the external world”. This was mentioned especially often in discussions
of quantum mechanics — it was a common observation that our difficulties grasping
wave-particle duality might be due to our previous conditioning to see the world
through the lenses of our subject-predicate languages and logics. Soon after, I
learned about the Sapir-Whorf hypothesis, and today I lump all those cautionary
tales under the heading of GRAM (“Grammar Recycled As Metaphysics”).

Back to packing ...
More later ...

Jon

On 6/19/2018 1:31 PM, 'Ferenc Kovacs' via ontolog-forum wrote:
> Ravi, Jon, For people who do not accept that relations in a natural language is represented by a verb, although the
> very same relation may be described in other linguistic(grammar) structures and who do not believe that the most

### Ferenc Kovacs

Jul 7, 2018, 5:29:16 AM7/7/18
Jon
As I have written before the fact that we can move about in space and (time), both physically (in fact) and mentally enables us to switch points of views that are described in geometric terms resulting in points, two of which connect as a line and three of which connect as a plane. Lines have a bipolar setup, but moving on up and down is the same as a circle, or a cycle. Words are invented to represent the end points and experience shows that such a setup is contrast required for recognition of points in language too. Movement is thus closed, may be bidirectional and is the most likely structure of associations, the elements or repetition, substitution and recursion. Moving on a line connecting three points means moving on a plane that cannot be represented by words arranged in a planar fashion in a natural language manifested linearly.
The reason or explanation may be that the number of linguistic units may not be arbitrarily (more than two or three), they may not be connected in any kind of chunks at will either, because in thinking you need to follow a particular direction and  you must stay in the confinement of a single plane (three points). It is true whether you manipulate words or numbers, it is easy to repeat addition and multiplication, etc. of twos, but not threes which requires changing directions or leaving the plane of communicative operations. As I once already quoted this passage:
\Every morning in the early part of the above-cited month, on my coming down to breakfast, your (then) little brother William Edwin, and yourself, used to ask me: `Well, Papa, can you multiplytriplets?' Whereto I was always obliged to reply, with a sad shake of the head: `No, I can only addand subtract them.' " The problem, of course, was that there exists no 3-dimensional normed division algebra. He really needed a 4-dimensional algebra.

ijkijk 􀀀1:

He showed that they were a normed division algebra and used this to express the product of two sums of eight perfect squares as another sum of eight perfect squares: the `eight squares theorem' [51 http://www.ams.org/journals/bull/2002-39-02/S0273-0979-01-00934-X/S0273-0979-01-00934-X.pdf

Ferenc

Ferenc Kovacs
Cell phone: 36-70-6784000
Landline:36-87-782831

--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-forum+unsub...@googlegroups.com.

### joseph simpson

Jul 12, 2018, 1:28:21 AM7/12/18
to structura...@googlegroups.com, Sys Sci, Ontolog Forum @ GG, Jon Awbrey, mjs...@gmail.com
Jack:

During the presentation of Technical Report 23, "System Structure and Behavior," it was noted that the current presentation of Constraint Theory could be improved.

The report -- available at:

-- began to analyze the Constraint Theory examples.

If there is interest, we could continue the evaluation of this material in more detail.

Would that be a valuable exercise?

### Ronald Stamper

Jul 12, 2018, 7:20:58 AM7/12/18
Dear colleagues,

May I call upon your interest in triadic relations to assist me?
A key result of our research programme is an approach to semantics that is proving very successful in practice.
Nevertheless, as Refutationists, following Popper, we always look for ways to falsify our hypotheses and theory.
One key hypothesis states that, in our semantics, triadic and higher order relations are not needed; they are always
composed of binary relationship, not of an artificial character, but ‘real’ physical, social or semiological relationships.

I would greatly appreciate any suggestions of triadic relationships that threaten to refute that hypothesis.

Regards,

Ronald Stamper

--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.

### McCarthy, William

Jul 12, 2018, 9:59:31 AM7/12/18

OK, I will try this M1 UML example from “Brian’s Gemini Music” – part of a revenue business process from a database class with REA (Resource-Event-Agent) modeling.

Jobs (gigs) are entertainment events (occurrents) that last 1-4 hours  (weddings, reunions, corporate functions, booze cruises, etc.).  Players are people who may play multiple roles (DJ, singer, dancer, officiant, etc,) on a single job.  The entrepreneur needs to know who did what for how long on every gig.

Bill McCarthy

Michigan State University

Ronald Stamper

Jack

Jack

To unsubscribe from this group and stop receiving emails from it, send an email to structural-mode...@googlegroups.com.
To post to this group, send email to structura...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Structural Modeling" group.

To unsubscribe from this group and stop receiving emails from it, send an email to
To post to this group, send email to

--

Joe Simpson

# All progress, therefore, depends on unreasonable people.”

George Bernard Shaw

--

Joe Simpson

# All progress, therefore, depends on unreasonable people.”

George Bernard Shaw

--

Joe Simpson

# All progress, therefore, depends on unreasonable people.”

George Bernard Shaw

--
You received this message because you are subscribed to the Google Groups "Structural Modeling" group.

To unsubscribe from this group and stop receiving emails from it, send an email to structural-mode...@googlegroups.com.
To post to this group, send email to structura...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Structural Modeling" group.

To unsubscribe from this group and stop receiving emails from it, send an email to structural-mode...@googlegroups.com.
To post to this group, send email to structura...@googlegroups.com.

--

Joe Simpson

# All progress, therefore, depends on unreasonable people.”

George Bernard Shaw

--
You received this message because you are subscribed to the Google Groups "Structural Modeling" group.

To unsubscribe from this group and stop receiving emails from it, send an email to structural-mode...@googlegroups.com.
To post to this group, send email to structura...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Structural Modeling" group.

To unsubscribe from this group and stop receiving emails from it, send an email to structural-mode...@googlegroups.com.
To post to this group, send email to structura...@googlegroups.com.

### Matthew West

Jul 12, 2018, 11:21:07 AM7/12/18

Dear William,

I can provide an answer to this example.

If you look at the attached space time diagram you can see that I have transformed the diamond in your diagram into a state_of_person (clearly not a triadic relationship) that is a participant in an activity (Gig) that is classified by a particular role. The state of course has a start and end time, the start time for the person is their DoB. I have not done the qualification or rate for the job, but these do not affect the triadicity of the relation and can be relatively easily added.

By the way, the arrows in the diagram represent classification relationships, and whole part is represented by things being geometrically part of.

Regards

Matthew West

Information  Junction

Mobile: +44 750 3385279

Skype: dr.matthew.west

This email originates from Information Junction Ltd. Registered in England and Wales No. 6632177.

Registered office: 28, Connemara Crescent, Whiteley, Fareham, Hampshire, PO15 7BE.

image002.png
Presentation1.jpg

### John F Sowa

Jul 12, 2018, 11:54:39 AM7/12/18
On 7/12/2018 7:20 AM, Ronald Stamper wrote:
> One key hypothesis states that, in our semantics, triadic and
> higher order relations are not needed; they are always composed
> of binary relationship, not of an artificial character, but ‘real’
> physical, social or semiological relationships.
>
> I would greatly appreciate any suggestions of triadic relationships
> that threaten to refute that hypothesis.

It is true that predicate calculus allows you to replace

But -- and this is a huge **BUT** -- there are two points to note:

1. If you represent your logic in a graph notation (as Peirce
and many linguists do), you will note that there is still
a syntactic change, not a semantic change.

2. Linguists use the word 'obligatory' to distinguish semantic
triads that cannot be eliminated -- i.e., if you delete one of
the dyads, the semantics will imply that something is missing.

For example, look at the attached give.jpg, which shows two ways
of representing the sentence "Sue gives a child a book."

The graph on the left represents the verb 'gives' with a triadic
relation. And the graph on the right represents 'give' with a
simple node that is attached to three dyadic relations for the
agent (Agnt), recipient (Rcpt), and theme (Thme).

Linguists say that the verb 'give' has three obligatory adjuncts.
If you omit one (for example, "I gave \$15"), the resulting sentence
implies that there was some unstated recipient.

As another example, if you say "I gave at the office", linguists
say that "at the office" is an optional adjunct, and that there
are two missing obligatory adjuncts: the theme and recipient.

As an example of a non-obligatory triad, consider the sentence
"Bill ate a sandwich for lunch." This would give you a triadic
graph that is similar to give.jpg. But "for lunch" is an optional
adjunct. If you omit it, there is no obligatory constraint that
would imply that something is missing.

Summary: (1) A graph notation or a parse tree for a linear notation
will show that a triadic connection cannot be eliminated by using
only dyadic relations. (2) Linguists have shown that some words
deleted, the semantics will imply that something is missing.

John
give.jpg

### Matthew West

Jul 12, 2018, 1:42:23 PM7/12/18
Dear John,
Saying that an activity is a relationship is simply a category error. An activity is not something that relates things, it is something that has participants, and given that the minimum number of participants is one (e.g. a person whistling) and the largest number is a lot (consider a football match, not just the 22 players, but officials, and don't forget each person in the crowd watching) saying that an activity is a relationship just does not work.
The example you give below, is actually about constraints, where you are requiring that certain participants exist for an activity to be valid. I agree that a relationship is a way to ensure this, it is not the only way that it can be done.
Regards
Matthew

-----Original Message-----
Sent: 12 July 2018 16:55
Subject: Re: [ontolog-forum] [structural-modeling] Triadic Relations

### William Frank

Jul 12, 2018, 3:59:57 PM7/12/18
to ontolog-forum
! +1

People who learn a way to cast something of category c1 as something of category c2 often say that the thing *is* or *really is* of category c2.   And, as you example shows, an activity is one category, and a relationship another.   YEt, there are mappings between assertions made using the different categories.

That is, there are a dualities among many categories, so that the activity of getting married, for instance, has a dual in the event of a given marriage taking place, which in turn has a dual in a given marriage relationship obtaining among some of the participants in the event.   The relationships, in turn, provide constraints on future activities and events, as you mention, such as the fact that in most current cultures, you can't get married if you are already married.  And, people who believe that some upper ontology, that has usually a small given set C of categories, is the 'right' one, or sufficient, naturally re-cast any ci not in C as another cj that *is* in C.    I have always wondered at the cavalierness of logician's willingness to treat 'is the father of', and 'gave something to' as both of the same category (relation).    I see this as the same pattern as people who take a course in which they learn that all the other boolean connective can be defined in terms of 'nor' as demonstrating that all the others are somehow 'secondary', or uncessessary'.

Instead, I believe that if one's goal is to understand and model how to reason, and to be able to systematize knowledge in a domain about which we then *can* reason, so dependent on definitions of terms in the language, it is more effective to rationally reconstruct an upper ontology with which people work in one or more natural languages, using all the categories we happen to find, and then define duality relationships among these categories, and only then decide what subset of categories we want to recast all the others into, for some given purpose.   I think this is hard; I don't know how to do it, though I believe I have seen it done one or a few languages at a time.   But it seems to me that deciding ahead of time we only have things and relationships, say, or only facts, and then going about and showing how to do the re-casting, sweeps this hard and important problem of understanding the categories of thought and experience people use, under the rug, as well as generating irreconcilable difference of opinion, founded as these differences seem to be  on no clear and comprehensive purpose or consequent set of requirements.

Wm

To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-forum+unsubscribe@googlegroups.com.

--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-forum+unsubscribe@googlegroups.com.

### John F Sowa

Jul 12, 2018, 9:06:37 PM7/12/18
to ontolog-forum
Pat, Leo, Matthew, and William,

I want to emphasize that this issue has *nothing* to do with
a graph notation or a linear notation for logic. It is a
fundamental semantic issue that happens to be easier to see
in a graph.

But as I said, it can also be seen in the parse tree of any
linear notation. And it can even be seen in the patterns
of variables that occur in an algebraic notation.

PJH
> Triadicity - the need for a three-place relation - is an
> accidental side-effect of using one particular style of
> graphical notation.

No. It happens to be easier to see in a graph notation,
but Peirce discovered the issues in 1869 by analyzing Kant's
table of 12 = 4 x 3 categories. In those days, nobody had
a notation for predicates with more than one argument.

Fundamental distinction: In Kant's table, the third item
in each of the four triads was some kind of law or law-like
principle. In general, the third item answers the question
of how and why the first and second are related.

In 1870, Peirce published his relational *algebra*, which
was the first *linear* notation for relations with two or
more arguments. His triads came first, and he invented
a linear notation for representing them.

Frege (1879) invented the first notation for full FOL.
He didn't cite Peirce (1870), but he used Peirce's word
'valence' for the number of arguments. (Peirce's BS degree,
by the way, was in chemistry.)

Peirce (1885) invented the modern algebraic notation for
predicate calculus. Peano (1895) adopted it and replaced
Peirce's choice of symbols -- but he cited Peirce, not Frege,
as the source. For details, see the article by Hilary Putnam:
http://jfsowa.com/peirce/putnam.htm

Peirce did not develop his existential graphs until 1897.
For the next 14 years, he did most of his logic with EGs.
And he used them to explain his ontology.

Leo
> the usual distinction is between argument and adjunct...
> Adjuncts are always optional, but arguments are either
> necessary or optional.

Yes. I admit a temporary lapse there. When I was developing
conceptual graphs, I had not yet seen Peirce's existential
graphs. My primary sources were the semantic networks in AI
and Lucien Tesnière's dependency graphs.

The terms that Tesnière used (long before most other linguists)
are actant/circonstant. For my ontology, I translated those
terms as participant/circumstance. I think that pair is better
for semantics than argument/adjunct, which is too syntactic.

I apologize for not using my own ontology.

MW and WF,
> Saying that an activity is a relationship is simply a category error.

I certainly agree, and I endorse the +1 by William.

But we need to make a sharp distinction between the terminology used
for logic, the terminology used to describe natural languages, and
the terminology used for the ontology of the subject matter.

MW
> An activity is not something that relates things, it is something
> that has participants,

Yes. That is my preferred ontology. I represent all content words
in NLs by concept nodes in conceptual graphs. The concepts that
represent actions or activities are linked by relations named Agent,
Patient, Theme, Recipient, Instrument...

And a concept of type Donate is represented in the same way
whether it's expressed by the verb 'donate', the gerund 'donating',
or the nominalized verb 'donation'.

But in the logic, we have symbols called predicates that represent
entities called relations, which have instances called relationships.
Those relationships should not be confused with relationships in
the subject matter (even though some subject matter relationships
might be represented by logical relationships).

I definitely do not like to represent a verb such as 'give'
by a triadic relation. I used that as an example because there
are some people (even occasionally some linguists) who do that.

John

### Jon Awbrey

Jul 13, 2018, 9:54:14 AM7/13/18
Ronald, All …

Carpet installers have taken over my office — just barely have WiFi and phone, so being briefest for now …

Concrete examples are very instructive in this arena, and when it comes to refutations a single counter-example serves to puncture many a fallacious general proposition.

Here’s a couple of articles, illustrated with concrete examples of a fundamental character, that provide a bare minimum introduction to triadic relations along with their reducibility and irreducibility properties.

• Relation Reduction

Regards,

Jon

### Ravi Sharma

Jul 14, 2018, 3:05:02 PM7/14/18
to ontolog-forum
What is the difference between Predicates, triples (as special case!) and triadic relations?
Sanskrit grammar has 8 Cases (these are modifications of Noun or Object) and German 4 and I think Russian might have 6? Are these related to Triadic?
Regards

Ronald Stamper

Jack

Jack

To unsubscribe from this group and stop receiving emails from it, send an email to structural-modeling+unsub...@googlegroups.com.

To post to this group, send email to structural-modeling@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Structural Modeling" group.
To unsubscribe from this group and stop receiving emails from it, send an email to structural-modeling+unsub...@googlegroups.com.

To post to this group, send email to structural-modeling@googlegroups.com.

--
Joe Simpson

# All progress, therefore, depends on unreasonable people.”

George Bernard Shaw

--
Joe Simpson

# All progress, therefore, depends on unreasonable people.”

George Bernard Shaw

--
Joe Simpson

# All progress, therefore, depends on unreasonable people.”

George Bernard Shaw

--
You received this message because you are subscribed to the Google Groups "Structural Modeling" group.
To unsubscribe from this group and stop receiving emails from it, send an email to structural-modeling+unsub...@googlegroups.com.

To post to this group, send email to structural-modeling@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Structural Modeling" group.
To unsubscribe from this group and stop receiving emails from it, send an email to structural-modeling+unsub...@googlegroups.com.

To post to this group, send email to structural-modeling@googlegroups.com.

--
Joe Simpson

# All progress, therefore, depends on unreasonable people.”

George Bernard Shaw

--
You received this message because you are subscribed to the Google Groups "Structural Modeling" group.
To unsubscribe from this group and stop receiving emails from it, send an email to structural-modeling+unsub...@googlegroups.com.

To post to this group, send email to structural-modeling@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Structural Modeling" group.
To unsubscribe from this group and stop receiving emails from it, send an email to structural-modeling+unsub...@googlegroups.com.

To post to this group, send email to structural-modeling@googlegroups.com.

--
Joe Simpson

# All progress, therefore, depends on unreasonable people.”

George Bernard Shaw

--
Joe Simpson

# All progress, therefore, depends on unreasonable people.”

George Bernard Shaw

--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-forum+unsubscribe@googlegroups.com.

--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-forum+unsubscribe@googlegroups.com.

--
Thanks.
Ravi
(Dr. Ravi Sharma)
313 204 1740 Mobile

### Azamat Abdoullaev

Jul 15, 2018, 7:15:27 AM7/15/18
to ontolog-forum
The question to the point.
Broadly, it is a distinction of two basic worldviews: Aristotelian and Kantian.
Or, if things and their properties (or substances and their attributes) OR relations of all sorts, make the prime elements of the world, its nature and knowledge.
Accordingly, there is the difference between a subject-predicate logic and relational logic.
Unlike predicates, relations are classified as to the nature of related entities, as well as wrt formal properties, number, symmetry, transitivity or reflexivity.
Don't know about formal logic, but all advanced science, and mathematics, is mostly about studying of the relationships of things.
Global Ontology is to have Relationship as its prime entity, while Data Ontology, as the prime concept.
Regards

Ronald Stamper

Jack

Jack

To unsubscribe from this group and stop receiving emails from it, send an email to structural-modeling+unsubscri...@googlegroups.com.

To post to this group, send email to structural-modeling@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Structural Modeling" group.
To unsubscribe from this group and stop receiving emails from it, send an email to structural-modeling+unsubscri...@googlegroups.com.

To post to this group, send email to structural-modeling@googlegroups.com.