Ontology languages comparison

[alex.shkotin]

John, Doug, and All,

Folowing your intresting discussion I asked Grok to compare OWL2, CYCL, Common Logic and Coq.

As I can't verify the answer in full, please, use this https://grok.com/share/bGVnYWN5_5fe4a83b-7fcf-4adf-8018-95f7ee841a0a

if you wish.

I added Coq becouse of https://github.com/GeoCoq/GeoCoq

Alex

[Alex Shkotin]

IN ADDITION

Disclaimer: Grok's raw response contains his own opinion, which may not coincide with the opinion of the person who created the prompt.

The initial discussion was in Russian.

Alex

Prompt: Let's compare the following languages as a means of creating formal ontologies: OWL2, CYCL, Common Logic, Coq. According to the following parameters: availability in open access, presence of a primer, presence of a syntactic analyzer, what are the means of knowledge processing, expressive power.

Grok:Here's the same comparison of the four languages/systems for building formal ontologies, now in English (updated as of November 20, 2025).

Parameter	OWL 2	CycL (Cyc language)	Common Logic (CL)	Coq
Open access availability	Fully open W3C standard. All specifications, tools, and ontologies are free.	OpenCyc (heavily reduced version) has been open since 2012; the full Cyc system (ResearchCyc) is closed-source and requires a paid/commercial license from Lucid.ai (as of 2023–2025). CycL documentation is only partially public.	ISO/IEC 24707:2018 standard. Specification is freely downloadable (after ISO registration), but not fully “open-source” in the usual sense. Implementations are mixed.	Completely free and open-source (OCaml codebase, LGPL license). All tools, source code, and ecosystem are free.
Availability of primers / good documentation	Excellent: official OWL 2 Primer (W3C), Manchester Syntax Primer, Protégé tutorials, books (“Semantic Web for the Working Ontologist”, etc.). Huge amount of learning resources.	Poor. Only old OpenCyc tutorials (2012–2017) and scattered wiki pages remain. No up-to-date, comprehensive public CycL primer exists (most documentation is internal).	Official standard + the book that is part of the standard. A few academic tutorials exist, but they are hard for beginners.	Outstanding: 600+ page official reference manual, several high-quality books (Software Foundations, Certified Programming with Dependent Types, Coq’Art), and dozens of courses/tutorials.
Parser / syntactic analyzer availability	Many mature parsers: Protégé (built-in), OWL API (Java), Apache Jena, RDF4J, owlready2 (Python), Manchester syntax parser, etc. All syntaxes (RDF/XML, Turtle, Manchester, Functional, DL) are supported.	OpenCyc had a SubL-based CycL parser. Today almost no maintained open-source parsers exist. The old opencyc-owl Java parser is dead and does not support modern CycL.	Several exist: CLIF ↔ TPTP converters, Hets supports CLIF, Vampire/Leo-III can read CLIF, some Prolog-based tools. But ready-to-use, user-friendly parsers are rare.	Built-in extremely powerful parser (Coq vernacular + LTac). Plus coq-serapi, coq-of-ocaml, standalone coq-parser libraries. Parsing support is top-tier.
Knowledge processing tools (reasoning, querying, editors, etc.)	Very rich ecosystem: • Reasoners: HermiT, Konclude, ELK, Openllet, FaCT++, TrOWL • Editors: Protégé (de facto standard), WebProtégé, VocBench • Query: SPARQL-DL, DL Query tab • APIs: OWLAPI, rdflib+owlrl (Python)	Full proprietary Cyc has by far the most powerful general-purpose inference engine of the four. Open alternatives are effectively dead. PyCyc wrapper exists but only works with the ancient OpenCyc.	Reasoners: Vampire, Leo-III, Paradox, Mace4 (for CLIF), Hets. Querying usually via TPTP or Prolog conversion. Almost no dedicated ontology editors (mostly text-based).	Coq is a proof assistant → reasoning = full interactive/automated theorem proving (Ltac, Auto, Hammer plugins). No SPARQL-style query, but you can program any extraction. Editors: CoqIDE, VsCode+VSCoq, ProofGeneral (Emacs).
Expressive power	OWL 2 DL = description logic SROIQ(D) → very high for ontologies (classes, properties, cardinality, property chains, datatypes), but still decidable. OWL 2 Full is undecidable.	CycL ≈ higher-order logic with modalities, microtheories, contexts, default & non-monotonic reasoning. Highest expressive power of the four (can formalize almost any commonsense or scientific concept). Undecidable.	Full sorted first-order logic with some second-order features (ISO Common Logic). Much more expressive than OWL 2 DL, but less than CycL (no built-in modalities or microtheories). Undecidable in general.	Based on Calculus of Inductive Constructions (CiC) → predicative higher-order intuitionistic logic with inductive/coinductive types. Comparable to HOL Light or Isabelle/HOL. Can formalize virtually anything (including ontologies as proven-correct libraries). Requires proofs.

Short conclusions

• For practical ontology engineering right now with a huge community and tooling → OWL 2 is unbeatable.

• CycL offers by far the greatest expressive power and comes with the massive pre-built Cyc ontology, but in 2025 open access is extremely limited — practically unusable without a commercial license.

• Common Logic is the “cleanest” powerful first-order standard (de jure), but the tooling ecosystem remains weak and editor-less.

• Coq is not specifically designed for ontologies, yet it allows you to build any ontology as a mathematically verified library; perfect when you need provably correct formalizations (formal verification, mathematics).

If you want the best balance of accessibility + tooling + expressiveness for most real-world tasks → OWL 2.

Maximum raw expressiveness and you’re willing to pay or negotiate a license → CycL.

Maximum mathematical rigor and proof guarantees → Coq.

Common Logic remains primarily a theoretical/standard reference with limited practical adoption.

чт, 20 нояб. 2025 г. в 18:18, alex.shkotin <alex.s...@gmail.com>:

--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/1f28a7dc-f02f-4d5f-9758-6b1dcdafa9bbn%40googlegroups.com.

[Giancarlo Guizzardi]

Alex,

These are languages that people use to make ontologies.

They are not ontology languages.

best,

Giancarlo

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAFxxROSf47Qm3-wYXLa6vKg8BWTu6Y8G6NdtrbOxpBUNGsqPVA%40mail.gmail.com.

[Ítalo Oliveira]

Hi guys,

SHACL has been increasingly used as an ontology specification language in the industry. Holger Knublauch (TopQuadrant) has dropped OWL completely in their projects, and he has been advocating for that. They argue it simplifies the codebase, in contrast to maintaining OWL and SHACL, and we lose nothing that really matters (or so they so). I wonder what your take is on this.

PS: That said, I second Giancarlo's opinion that these are not ontology languages, but languages that have been used to specify ontologies. They are not ontology languages because they do not commit to an ontology (besides their formal semantics, usually set theory). In comparison, OntoUML does exactly that by committing to a foundational ontology embodied in its grammar.

Best regards,

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAGG%2B%3DmXFiT%3DU3dfQv3dzjYEUDuCsmM%2B9UoCv%3D4co2b6%3DAttm5Q%40mail.gmail.com.

--

Ítalo Oliveira

Semantics, Cybersecurity, & Services Group (SCS)

Faculty of Electrical Engineering, Mathematics, and Computer Science

University of Twente

P.O. Box 217, 7500 AE Enschede, The Netherlands

E-mail address: i.j.dasil...@utwente.nl

Office: Zilverling 2034

Personal website 

Blog Not So Short Notes

[Alex Shkotin]

Giancarlo,

What does the term "ontology languages"  mean for you?

Alex

пт, 21 нояб. 2025 г. в 13:11, Giancarlo Guizzardi <gguiz...@gmail.com>:

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAGG%2B%3DmXFiT%3DU3dfQv3dzjYEUDuCsmM%2B9UoCv%3D4co2b6%3DAttm5Q%40mail.gmail.com.

[Giancarlo Guizzardi]

Alex,

In one line: a language which primitive reflects ontological distinctions and which grammar

reflects ontological constraints. 

best,

G

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAFxxRORpgjGfnRiQ9Cipd3O2M5x6%3DHBt97-wORA3k8ec%3Dc493Q%40mail.gmail.com.

[Alex Shkotin]

Please, give an example of such a language. The more the better.

пт, 21 нояб. 2025 г. в 14:20, Giancarlo Guizzardi <gguiz...@gmail.com>:

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAGG%2B%3DmX7SUqjHwbQRn9jPxMf2DuxrMwrK0MF55WeZZoY8G9RWQ%40mail.gmail.com.

[Ítalo Oliveira]

Alex, I gave you an example earlier: OntoUML.

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAFxxRORNEpUapJd17weDEA%3D9Vpc-2zk9DkWc99y0c1TkPMAufQ%40mail.gmail.com.

[Alex Shkotin]

Sure, but I asked GG about his knowledge.

пт, 21 нояб. 2025 г. в 14:25, Ítalo Oliveira <italojso...@gmail.com>:

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAOEC0q-4xKyTXmZm-VF6z1n9dXooYASZT1Fb0erB5kqieAdwMg%40mail.gmail.com.

[Giancarlo Guizzardi]

Well, an example is OntoUML but I don't want to beat on my own drum

It has been designed in full commitment to the Unified Foundational Ontology (UFO).

In fact, it is a pattern language which real primitives are ontology design patterns

that represent the micro-theories comprising UFO

IMO this boils down to Nicola's Ontological level (https://www.researchgate.net/publication/2366917_The_Ontological_Level)

Ontology languages are ontological-level languages in his sense

I am not promoting either UFO or OntoUML here

I was just making an observation. had the headline been:

A comparison between (logical) languages used to create ontologies,

it would be totally fine. Maybe that is what you want to discuss. 

That is fine an potentially interesting

best,

G

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAFxxRORNEpUapJd17weDEA%3D9Vpc-2zk9DkWc99y0c1TkPMAufQ%40mail.gmail.com.

[Alex Shkotin]

Ítalo, formalizing knowledge in one or another formal language it's interesting to know about ontology languages.

пт, 21 нояб. 2025 г. в 14:25, Ítalo Oliveira <italojso...@gmail.com>:

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAOEC0q-4xKyTXmZm-VF6z1n9dXooYASZT1Fb0erB5kqieAdwMg%40mail.gmail.com.

[Alex Shkotin]

GG, thanks for clarification. But before I dig the URL. Is there any other language in your mind? not OntoUML. UFO is not a language.

Aligning terminology is a great deal before sharing knowledge.

Just from my point of view: when we are talking about one or another primitives we are talking about one or another theory. Language is more or less independent of things it describes.

And this topic is interesting itself: which primitives this particular language has, and how ontological they are.
Is owl:Thing an ontological primitive of OWL2?

And I agree with your reformulation. And the topic of "ontology languages" is very interesting. For me under the term "theories of Being".

My main question is: Is there any other language in your mind? not OntoUML. UFO is not a language.

Alex

пт, 21 нояб. 2025 г. в 14:29, Giancarlo Guizzardi <gguiz...@gmail.com>:

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAGG%2B%3DmUWtSzE7PoGZrnAFd4mt5zk_DoMf7QwV2nJz3AyzWiajQ%40mail.gmail.com.

[Alex Shkotin]

GG, your "grammar reflects ontological constraints" is very interesting. Please give an example.

пт, 21 нояб. 2025 г. в 14:20, Giancarlo Guizzardi <gguiz...@gmail.com>:

Alex,

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAGG%2B%3DmX7SUqjHwbQRn9jPxMf2DuxrMwrK0MF55WeZZoY8G9RWQ%40mail.gmail.com.

[Giancarlo Guizzardi]

On Fri, Nov 21, 2025 at 12:53 PM Alex Shkotin <alex.s...@gmail.com> wrote:

GG, thanks for clarification. But before I dig the URL. Is there any other language in your mind? not OntoUML. UFO is not a language.

Yes, UFO is not a language.

All languages inevitably make ontological commitments. Even FOL.

So, in this sense, my definition of an ontology language can be trivially satisfied in a boring way

The question is these languages commit to a suitable foundational layer on which we would like to

ground the phenomena being represented in that language

Take UML, for example. It does commit to an underlying ontology that is less neutral than the 

one we have for FOL. To make it simple: take the representation of part-whole relations in UML class diagrams.

It has primitives like aggregation, composition, meta-properties like shareability, life-time dependency, etc...

Less neutral than FOL. However, it commits to a bad ontology of parthood

(it has tons of problems but I don't want to digress here)

Suppose now that we have a good mereological theory, then this can be coded in the language's

primitives and associated grammatical constraints. An example of the latter is one that makes

an expression in that language that does not satisfy, e.g., weak supplementation grammatically incorrect.

  

Aligning terminology is a great deal before sharing knowledge.

absolutely

 

Just from my point of view: when we are talking about one or another primitives we are talking about one or another theory. Language is more or less independent of things it describes.

And this topic is interesting itself: which primitives this particular language has, and how ontological they are.
Is owl:Thing an ontological primitive of OWL2?

class, objectproperty, dataproperty are

if you mean by Thing TOP in the DL sense, then yes

And I agree with your reformulation. And the topic of "ontology languages" is very interesting. For me under the term "theories of Being".

I think it is a very interesting topic indeed 

My main question is: Is there any other language in your mind? not OntoUML.

Given what I wrote above, I would rephrase the question upside down:

1) what are the foundational ontologies we would like to ground our domain-level work on?

2) how many of these foundational ontologies have modeling languages that would support that process? 

 

To answer your original question, there are other languages that were designed with the aim

of doing that in different ways, for example:

I think Dov Dori's OPM in a sense aimed at doing that

John's Conceptual Graphs were also informed by his own ontology

UFO is not a language.

yes, correct. UFO is not a language 

-G

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAFxxRORZfa83om-4xkE8WosH67_NNCtXcUnXPC8gjAGkf6RbOw%40mail.gmail.com.

[Giancarlo Guizzardi]

On Fri, Nov 21, 2025 at 12:58 PM Alex Shkotin <alex.s...@gmail.com> wrote:

GG, your "grammar reflects ontological constraints" is very interesting. Please give an example.

My example of parthood in the previous email is an example of that.

Here is another one. Suppose in your ontology of types you have a notion of Role

such that:

1) Roles are sortals that specialize a unique Kind

(e.g., all students are people)

2) Roles are defined in a relational context

(e.g., in order for one to be a student, they must be enrolled in an educational institution)

3) Roles are contingent

(e.g., no student is necessarily - in a modal sense - a student)

4) Roles cannot supertype a Kind

...

Then we can design a grammar in which, not only we have primitives that

represent Kinds, Roles, relational conditions, etc but these primitives

of the language are governed by grammatical constraints such that a

model in that language that breaks rules 1-4 above are ungrammatical models

-G

 

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAFxxROQCXrtcdbh3r-RuK0F%2Bdj4jShqXB%3DPJ8K1PapMpv34Low%40mail.gmail.com.

[John F Sowa]

Alex, Giancarlo, Italo, Doug, List,

Before getting into details, we should start with a definition of what we mean by ontology, an ontology language, and their relationship to logic, natural languages, the people who specify or modify an ontology, and the people who develop tools and technology for working with ontology for designing, implementing, and using software based on that ontology.

Definition:   An ontology is a specification of the concepts and relations of some body of knowledge or subject matter in a version of formal logic that can be used for reasoning about the subject and for developing software that processes data represented in that logic.

My book on Conceptual Structures (1984 was one of the first to use the word ontology instead of epistemology for specifying the concepts and terminology used to design and implement knowledge based systems.  I have collaborated with various people who have specified ontologies for systems of various kinds.   For the past 25 years, I have been working with Arun Majumdar and others in AI startup companies dedicated to the development of AI systems that use large, rich ontologies.   So I'll talk about our experience and how it might help others on their projects.

Fundamental issue:  Some people have complained that Common Logic is too complex and unreadable for most users and developers.  That is probably true for the internal representation.  But we allow users and developers to read, write, and update ontologies in their preferred natural languages -- English, Russian, Portuguese,, Chinese, or whatever.

We don't expect anyone to learn the internal notation.  They can just say what they mean in their preferred language.  The system translates that to the internal form.  Then it transates the internal form to a precise statement in their preferred language and asks "Is this what you mean?"   The users can say "yes" or add whatever revisions or corrections in their preferred language.

The users never have to learn what subset or version of their preferred language is acceptable for ontology.  But after a while, they learn what kinds of statements are immediately accepted and what kinds require more back and forth revisions.

 

[John F Sowa]

I  accidentally hit SEND before I finished that note.

But it's after 11 pm and there is much more to say and many references to cite.

So it;s better to leave that for tomorrow.

Good night.      

[Giancarlo Guizzardi]

Dear John and others,

Apologies for replying to your unfinished email

(and thanks for the detailed message despite the late hour).

I am replying half-way because I think the discussion is going 

to a different direction than the one I intended.

I have nothing against CL. Your 1984 Books is a classic. Great great book.

I read it for the first time 30 years ago and it was one of the texts that got me excited about

knowledge representation.

Now, back to ontology. Allow me to propose a different definition.

In the context of our discussion here, an ontology is  a domain model

that results from a process of ontological analysis.

It is a semantic contract  that tells the world what your worldview is about that domain.

What is assumed to exist and with which nature/definition, and what are the state of affairs  

deemed acceptable in that domain. 

Yes, it contains concepts and relations and constraints...

Constraints are fundamental for ensuring (as much as possible)

that the models of your ontology are the intended ones

(i.e., that your representation is neither over- nor under- constrained).

For that reason, a formal language plays a fundamental role here

and the more expressive it is, the more helpful it is to help us achieve that goal.

Moreover, ideally, the representation should be expressive enough in order not to

collapse relevant distinctions among state of affairs in the world into the same logical model. 

So, being able to express whatever constraints one wants is necessary but not sufficient.

The real challenge and most important bit is knowing which constraints to express

We obviously need to constrain the representation  w.r.t. the right set of intended state of affairs. 

For that, logic does not help us much. For that we need ontological analysis. 

So, I don't think the primary role of logic in ontology is to support reasoning

and that it necessarily needs to help us to build software to process data.

Of course, these are important applications of an ontology. Once we get the ontology right,

we can create MANY new artifacts derived from it, each optimized to support particular

reasoning (among other) tasks. All fine. 

Now, let me go back to the original discussion.

Given the aforementioned role of ontological analysis in this process.

Ontology (capital O) is of fundamental value here.

Of course, we can look at our best Ontological theories about 

mereology, causality, dependence, identity, etc. and do ontological

analysis with them using a more (ontologically neutral) language, i.e., 

a language whose underlying ontology differs immensely from those theories. 

OR we can create languages whose underlying ontologies reflect exactly those

Ontological theories. As I said before, languages that embodied in their primitives

and grammar the relevant ontological distinctions, constraints and patterns put forth

by those theories. 

This language would have an analogous effect to what Domain-Specific

Modeling Languages would do in Software Engineering, namely, create uniformity and agility

in the development of representations using those Ontological theories - which is something

we empirically demonstrate, by the way. Moreover, some of those constraints that we need 

to approximate our logical models to the intended ones come for free here via a theory inclusion mechanism

(as they are embedded in the language).

So, in my view, the discussion I seem to have initiated boils down to the following statements:

a) An ontology needs to precisely characterize a domain including the state of affairs deemed acceptable according to the laws of that domain

b) Expressive logical languages are extremely useful for writing the constraints that approximate the logical models of our ontology to those state of affairs 

c) logical languages are not that helpful for telling us what those state of affairs are; ontological Analysis is needed for that

d) Ontological theories are fundamental for supporting ontological analysis

e) Every language presupposes some Ontology

f) a language which underlying Ontology matches the Ontological theories we want to use (as per c and d) will make our lives much more productive promoting constraint reuse (from those theories), uniformity and agility in engineering our ontologies based on those theories 

Again, apologies for the long email. I think this helps with the discussion.

Thanks

best,

Giancarlo

--

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

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/5caf133aca3c4adca504ea41a9c47322%409c989340b0c64fc293b1af2c88d929f3.

[Fahad Khan]

Perhaps this is a dumb question but how is it possible for FOL to make ontological commitments if any formula in FOL can be negated? I get that it can limit the kinds of ontologies one can describe but this isn't the same thing.

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAGG%2B%3DmU%2BtdHjDDrbr_8KVb8UZq83VvXBmLoDeLoO3ALMVM5%2BhA%40mail.gmail.com.

[Giancarlo Guizzardi]

On Sat, Nov 22, 2025 at 10:16 AM Fahad Khan <anasf...@gmail.com> wrote:

Perhaps this is a dumb question but how is it possible for FOL to make ontological commitments if any formula in FOL can be negated?

Fahad, this is a very interesting question.

Apologies because I am on the move and won't be able to enter into this (other argument now)

but I recommend taking a look at https://www.google.com/url?sa=t&source=web&rct=j&opi=89978449&url=http://www.columbia.edu/~av72/papers/LNC_2014.pdf&ved=2ahUKEwjW2MO0vYWRAxX8h_0HHSTgIfAQFnoECCcQAQ&usg=AOvVaw0UfgczAQsAgcdyZM_Crejm

 

I get that it can limit the kinds of ontologies one can describe but this isn't the same thing.

Actually, it is the opposite: FOL allows you to do basically everything including the things you shouldn't do.

So, the problem is that it does not help you in restricting the ontologies you shouldn't describe.

Sorry for the hurry...I want to add something to my previous email

-G

 

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAK%2BN%2B9inGVRK4XYnzKNy_gW1BHDRfYHu%2Br1YUuM%2Bdc-n213V%3DA%40mail.gmail.com.

[Giancarlo Guizzardi]

John,

This time maybe I pressed the send too fast.

Here is what I propose. Considering the following propositions:

a) An ontology needs to precisely characterize a domain including the state of affairs deemed acceptable according to the laws of that domain

b) Expressive logical languages are extremely useful for writing the constraints that approximate the logical models of our ontology to those state of affairs 

c) logical languages are not that helpful for telling us what those state of affairs are; ontological Analysis is needed for that

d) Ontological theories are fundamental for supporting ontological analysis

e) Every language presupposes some Ontology

f) a language which underlying Ontology matches the Ontological theories we want to use (as per c and d) will make our lives much more productive promoting constraint reuse (from those theories), uniformity and agility in engineering our ontologies based on those theories 

Let's start with which one you would reject.

Since we have had these discussions before (including a long, extremely interesting and pleasant one in Brazil 20 years ago),

I am going one step further:

1) I would be very surprised if you rejected (a). If we go that way, ontology becomes just any logical theory and validation becomes useless 

(only verification counts). Fine. Then we should just choose whatever system of logic is useful for the problem at hand but I think calling

these theories "ontologies" would be a misappropriation of the name. Again, I don't see us going there

2)  I would be surprised if you rejected (b) as well

3)  same for (c) and (d). We might disagree on which ontological theories would be useful or are the most appropriate ones

but that is a different (completely genuine) discussion

4)  Regarding (e), I don't want to enter a discussion now about the ontological neutrality (or lack thereof) of FOL. 

If we consider it to be completely neutral then it has a maximal distance from the typical ontological theories proposed  

by formal and applied ontology (again, theories of parthood, dependence, causality) and my argument is strengthened.

5) Regarding (f), this is an empirical question and, hence, not one we can answer from the armchair. We should look at the data.

I hope structuring the discussion this way helps us to isolate the core of our disagreement (if any) in the present discussion

and have one discussion at a time

best regards,

Giancarlo

PS: again, on the move, so it might take a while for me to reply

[John F Sowa]

Giancarlo,

I agree with every one of your propositions:

a) An ontology needs to precisely characterize a domain including the state of affairs deemed acceptable according to the laws of that domain

b) Expressive logical languages are extremely useful for writing the constraints that approximate the logical models of our ontology to those state of affairs 

c) logical languages are not that helpful for telling us what those state of affairs are; ontological Analysis is needed for that

d) Ontological theories are fundamental for supporting ontological analysis

e) Every language presupposes some Ontology

f) a language which underlying Ontology matches the Ontological theories we want to use (as per c and d) will make our lives much more productive promoting constraint reuse (from those theories), uniformity and agility in engineering our ontologies based on those theories 

My points are very simple and very general: 

1.  An ontological language is just a logical language plus some additional propositions that are specialized for some version of ontology.

2. Readability is a purely syntactic feature.   The same logic can be expressed by an infinite variety of different notations with different levels of readability, writability, analyzability, understandability, etc.

3. Common Logic (CL) is defined by an abstract syntax that makes no assumptions about the infinitely many options for expressing exactly equivalent propositions or collections of propositions.

4.. It's possible to express CL in very readable versions of any or every natural language.

5. It's possible to implement a very user friendly front-end to a CL based ontology that enables people to use their own natural language for input and output.

6. It's possible to make that front end so friendly that it is far superior to any and every syntax that anybody has specified for their ontology.

I'll say more about these issues in a future note.

John

___________________________________________

[Alex Shkotin]

GG,

I think I've grasped the idea: if there's a formal knowledge-representation language, then by examining its keywords and syntactic constraints, we can assess how ontological it is and what ontology (theory of being) should be used when recording certain knowledge in it.

To have a choice of ontology, we can take two extremes: a generally accepted ontology used in everyday life, or a particular philosophical doctrine or mereology.

What constitutes a generally accepted ontology used in everyday life is a separate and interesting question.

Then, a little about OWL2.

It includes a family of three languages (Manchester, Functional style, Turtle) and two ontology representations in XML (RDF/XML, OWL/XML).

Moreover, "Ontologies are formalized vocabularies of terms, often covering a specific domain and shared by a community of users. They specify the definitions of terms by describing their relationships with other terms in the ontology." [1]

And here we at least understand a fairly modest task: a certain subject area uses certain terminology, and an ontology stores it formally, even as a structure.

This structure is described as follows: "The ellipse in the center represents the abstract notion of an ontology, which can be thought of either as an abstract structure or as an RDF graph (see 2.1 Ontologies)."

And here is the reserved OWL2 vocabulary for ontological analysis (see):

"IRIs with prefixes rdf:, rdfs:, xsd:, and owl: constitute the reserved vocabulary of OWL 2. As described in the following sections, the IRIs from the reserved vocabulary that are listed in Table 3 have special treatment in OWL 2.

owl:backwardCompatibleWith	owl:bottomDataProperty	owl:bottomObjectProperty	owl:deprecated	owl:incompatibleWith
owl:Nothing	owl:priorVersion	owl:rational	owl:real	owl:versionInfo
owl:Thing	owl:topDataProperty	owl:topObjectProperty	rdf:langRange	rdf:PlainLiteral
rdf:XMLLiteral	rdfs:comment	rdfs:isDefinedBy	rdfs:label	rdfs:Literal
rdfs:seeAlso	xsd:anyURI	xsd:base64Binary	xsd:boolean	xsd:byte
xsd:dateTime	xsd:dateTimeStamp	xsd:decimal	xsd:double	xsd:float
xsd:hexBinary	xsd:int	xsd:integer	xsd:language	xsd:length
xsd:long	xsd:maxExclusive	xsd:maxInclusive	xsd:maxLength	xsd:minExclusive
xsd:minInclusive	xsd:minLength	xsd:Name	xsd:NCName	xsd:negativeInteger
xsd:NMTOKEN	xsd:nonNegativeInteger	xsd:nonPositiveInteger	xsd:normalizedString	xsd:pattern
xsd:positiveInteger	xsd:short	xsd:string	xsd:token	xsd:unsignedByte
xsd:unsignedInt	xsd:unsignedLong	xsd:unsignedShort

"

I've bolded the parts that don't relate to data.

Perhaps some ontological terms appear among the names of UML diagrams, which are considered the canonical definition of the OWL2 ontology structure.

I think they adhere to the generally accepted ontology used in everyday life.

It was interesting to delve deeper into this topic.

Now I can ask any creator of formal ontologies: what ontological doctrine do you adhere to?

Maybe we should use terms fontology and phontology?

Alex

[1] OWL 2 Web Ontology Language Document Overview (Second Edition)

пт, 21 нояб. 2025 г. в 16:33, Giancarlo Guizzardi <gguiz...@gmail.com>:

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAGG%2B%3DmVi-8eX4Ki1toJCBoya%2BxeZ-8c3Eg4y%3DcenQuP8gQxK5g%40mail.gmail.com.

[Ravi Sharma]

Distinguished researchers on the topic!

1. Knowledge has been represented (KR) over several millennia through sounds, scripts, Glyphs and Graphs and intermediate representations such as information and data (lowest minimum entity?). Thus use of NL for KR.
2. What is role of ontologies (both O and o) in KR, certainly historically first Ontologies defined were themselves in NL (e.g. Greek, Sanskrit). Hence Languages (generally NL) were predecessors to creating Ontologies.
3. What else is required (assuming all good NLs have logic) perhaps internal math representation of FOL?
4. Role of cognition, has to be brought in for specifying discourse domain etc (for what the Knowledge is about).

Will these aspects, coupled with technical programming language inputs from Alex help us integrate the JS and GG descriptions into one consistent definition and use of "purpose of ontology" these days with "o" as we are at least uncovering some aspects of existing KR through AI?

PS: With high regards to OMG, W3C, other standard bodies' enormous work, I am trying to keep use of these under the internal representations!

Regards.

Thanks.

Ravi

(Dr. Ravi Sharma, Ph.D. USA)

NASA Apollo Achievement Award

Former Scientific Secretary iSRO HQ

Ontolog Board of Trustees

Particle and Space Physics

Senior Enterprise Architect

SAE Fuel Cell Standards Member

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAFxxROTLajAy3LkNq6eX9gZ-Eq7RuP6yi0KpOwKLTMGypGqPAA%40mail.gmail.com.

[Nadin, Mihai]

Dear and respected colleagues,

Dear Ravi Sharma—who inspired my message:

Please read: “Translating into and from Mathematics” in  TRANSLATION BEYOND TRANSLATION STUDIES, Edited by Kobus Marais, Bloomsbury, 2022, pp. 5-22 (https://www.nadin.ws/wp-content/uploads/2024/09/translating-into-and-from-mathematics.pdf)

I am quoting, with the purpose of repeating my understanding of the subject:

 

1.                Ontology engineering provides actionable descriptions of reality. The important qualifier ‘actionable’ means in our time computational. (There were ontologies that were actionable in the industrial age, quite different from those we need today. Leibniz created one for translating from Chinese into Latin or French.) In this respect, ontology engineering is the ontology of the age of the Turing machine. I would even call it algorithmic ontology.

2.                Regardless of how this description is done (using knowledge graphs or semantic networks, or anything else), it is actually a translation of ontology (in its broad philosophic sense) in the language of the machine. If the machine would not have been an exclusively syntax processing engine, nobody would have become an ontology engineer.

Mihai Nadin

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAAN3-5c-y_-eZvwS0mc1f9sRPuQ8aFNX2iHTMQR89h8po8Dkmg%40mail.gmail.com.

[Chuck Woolery]

Words matter when they try to represent reality...even if translations vary.

Consider passing this indigenous thanksgiving prayer around the dinner table...any time.  Not just on Thanksgiving...;-)

 

cw

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/BL3PR01MB68971A344BA794CB80ADE208DAD3A%40BL3PR01MB6897.prod.exchangelabs.com.