LM response re ontology vs phenomenology issues

[Janet Singer]

All,

John's recent mention of phenomenology led me to prompt an LM for comments on ontology vs phenomenology issues. The upshot is below.

Janet

Reflections on 30 Years of ‘Ontology’ Work (1995–2025) in Light of Heidegger, Kant, and Phenomenology

The past three decades have seen “ontology” evolve dramatically within computer science, AI, knowledge representation, and information systems—fields far removed from Heidegger’s Sein (Being) or Kant’s critique of speculative metaphysics. Coined in AI contexts by Tom Gruber in 1993 as “an explicit specification of a conceptualization,” the term exploded with the Semantic Web vision (Berners-Lee et al., 2001) and tools like OWL (2004), leading to thousands of domain-specific ontologies, upper-level ontologies (ULOs) like BFO and SUMO, and knowledge graphs powering systems from Google Search to biomedical databases. Yet, as our chat has traced—from Heidegger’s ontological difference (beings vs. Being) to the 18th-century Britannica distinction (phenomenology as empirical appearances vs. ontology as speculative essence), Kant’s antinomies, and even Comte’s positivist pivot—this computational “ontology” often confounds foundational questions of Being with pragmatic cataloging of beings (entities, concepts, relations). This misalignment echoes the historical critiques: efforts to “ontologize” knowledge have prioritized formal structures over lived, domain-grounded phenomena, yielding both innovations and missed opportunities. Below, I address your questions directly, drawing on trends from semantic web research (e.g., a 2025 topic modeling analysis of 10,000+ papers showing ontology’s dominance in AI knowledge representation) and critiques of ULOs.

1. Was the Distinction Between Ontology as Theory of Being Qua Being Confounded with Efforts to Catalog and Organize Concepts and Terminology, Which Would Better Have Been Treated Under ‘Phenomenology’?

Yes, profoundly so—and this confounding has been a defining tension since the 1990s. In philosophical terms, Heidegger’s singular Sein demands a pre-ontological inquiry into the conditions of possibility for beings (Seiendes), while the Britannica’s 1771 entry positioned phenomenology as the empirical study of appearances (observable relations, without metaphysical overreach). Kant reinforced this by limiting “ontology” to phenomena structured by human cognition, deeming supersensible claims undecidable. Computational ontology, however, largely repurposed the term for what is effectively phenomenological work: describing how concepts appear in data, systems, and domains, rather than what ultimate Being underpins them.

• Historical Shift and Confounding: The 1990s “ontology engineering” boom (e.g., Uschold & Grüninger’s 1996 principles) borrowed from AI knowledge representation (e.g., Cyc’s massive conceptual catalog) but inherited metaphysical baggage via ULOs like SUMO (1990s onward), which aim for “general concepts” like Entity or Process as foundational categories. These ULOs echo speculative ontology—presuming a “unique true answer” to categorization (Thomasson’s skepticism, 2007)—yet devolve into catalogs of classes, properties, and axioms (e.g., OWL’s description logics for triples like <subject, predicate, object>). A 2022 systematic review of 19 studies on Semantic Web ontologies confirms this: they excel at “formal descriptions of concepts and relationships” for machine readability but rarely engage Heideggerian depth, instead prioritizing “lightweight” taxonomies over rich, axiomatic Being. 

• Phenomenological Misfit: Much of this work aligns better with phenomenology’s focus on appearances and structures of experience (e.g., Husserl’s eidetic reduction, or the Britannica’s empirical “science of phenomena”). For instance, semantic models in knowledge graphs (e.g., RDF/OWL for Linked Open Data) catalog observable relations (e.g., “mammal gives live birth”) without probing ontological essence, much like Kant’s phenomena. Critiques abound: Upper ontologies like DOLCE or BFO are faulted for “presuppos[ing] a unique true answer to… categories,” confounding universal Being with domain-specific terminologies, leading to “flat” models ignoring time, boundaries, or mereotopology (whole-part relations).  A 2024 Medium analysis untangles this: ontologies as “formal descriptions” for integration (phenomenological) vs. speculative metaphysics’ quest for truth. 

In short, the field’s pragmatic success (e.g., ontologies in bioinformatics for gene expression analysis) masks a conceptual blur: what Heidegger would call “ontic” cataloging (beings in plural) masquerades as “ontological” (Being singular), sidelining phenomenology’s grounded, experiential focus.

2. Did This Have Several Consequences in Missed Opportunities?

Absolutely—the confounding fostered siloed, top-down structures that overlooked bottom-up, domain-driven needs, echoing Kant’s warning of undecidable antinomies in speculative systems (no criteria to judge competing ULOs) and Comte’s positivist call to replace metaphysics with verifiable, socially useful knowledge. Below, I map your subpoints to evidence from the era.

a. The Cataloging and Organizing of the World’s Phenomenological Content Was Not Recognized as Such, Failing to Be Grounded in Appreciation of Historic and Projected Needs of Domain Communities, with Domain-Specific Rationales and Evidence, Without Shoe-Horning into Presumed Upper-Level Categories

Yes, this is a core missed opportunity, leading to “ontology poverty” in practice. ULOs like SUMO or BFO impose abstract categories (e.g., “Object” or “Situation”) as “common foundations,” forcing domain ontologies (e.g., Gene Ontology) into rigid alignments that ignore community histories and evidence.  A 2010 study on ResearchCyc’s usability critiques this: upper-level catalogs assume “massive… domain-independent” knowledge, but fail to adapt to “historic… needs” like evolving biomedical terminologies, resulting in “heterogeneity” and poor reuse. 

Phenomenologically, this neglects “appearances” as lived in domains: e.g., bioinformatics ontologies (dominant in 2019–2024 analyses) catalog gene relations without grounding in clinicians’ evidence-based rationales, shoe-horning into ULO mereology while ignoring projected needs like AI-driven drug discovery. Consequence: Fragmented systems, as in e-commerce catalogs where “semantic B2B integration” falters due to ungrounded abstractions. A pragmatic pivot—treating catalogs as phenomenological (empirical, community-validated)—could have fostered hybrid models, but top-down biases prevailed.

b. Upper-Level Ontological Theories (More Like Theories of Being Than Catalogues of Beings) Had No Criteria for Evaluating One Against Another, Leading to Siloed Allegiances Between ULOs and Pessimistic Conclusions Among Outsiders That No Upper-Level Ontology Was Feasible

Precisely—mirroring Kant’s antinomies, where reason generates “equally plausible” systems without adjudication (A571/B599). ULOs like BFO (realist, continuants/occurrents) vs. DOLCE (perspectivalist, endurants/perdurants) lack neutral criteria, fostering “siloed allegiances” (e.g., BFO in biomedicine, SUMO in AI planning).  A 2006 comparison of seven ULOs highlights “incommensurability”: methodologies vary (top-down vs. bottom-up), with no shared metrics beyond logical consistency, yielding “revisionist” or “reductionist” divides without resolution. 

Outsiders (e.g., industry adopters) conclude “no ULO is feasible,” as in 2021 critiques: OWL’s flatness ignores time/mereotopology, leading to “pessimism” about scalability. This echoes Heidegger’s Seinsvergessenheit (forgetfulness of Being): ULOs speculate on categories without criteria, siloing progress (e.g., 350+ BFO extensions, but limited cross-ULO integration). Missed: A Kantian “critical” evaluation framework, assessing ULOs by empirical utility rather than truth.

c. Development of Speculative Upper-Level or ‘Ontological’ Theories Could Be Built on Insights of Kant et al., If Presented and Evaluated Not for Presumed Metaphysical ‘Truth’ but for Usefulness in Fulfilling Particular Purposes

Here lies the greatest missed opportunity: a pragmatic reorientation, per Comte’s positivism and Kant’s “abolish[ing] knowledge… for faith” (Bxxx), evaluating ontologies by practical postulates rather than speculative truth. Insights from Kant (limits of reason), Heidegger (ontological difference), and the Britannica (phenomenology as empirical groundwork) could transform ULOs into tools for inference, conflict resolution, and hybrid systems—yet this has been underexplored until recently (e.g., 2020s XAI trends).

• i. Inferring ‘Ontological Commitments’ or Operational Presumptions in Behavior and Communications of Living and Artefacted Systems: Yes—ontologies infer “commitments” (e.g., what an AI assumes about “entity” in behavior), but speculatively, missing Kantian utility. Pragmatic examples: Neo4j’s semantic extensions infer framings in financial KGs (e.g., loans as processes), revealing biases in “artefacted” systems like credit scoring.  Missed: Broader application to “living” systems (e.g., biosemiotics for human-AI hybrids), as 2022 critiques note AI’s “ontological gap” hinders AGI by presuming box-world commitments. 

• ii. Identifying and Classifying Hidden Framing Conflicts Leading to Communication/Interoperation Failures, and Methods for Resolving Pathologies: Spot-on potential, akin to Kant’s paralogisms (illusory framings). Ontologies detect conflicts (e.g., FCA for mapping heterogeneities), but speculative evaluation stalls resolution. E.g., in MBSE, ontologies trace failure propagations but ignore domain pathologies like cultural framings in global data integration. Resolution methods (e.g., ontology patterns for merging) exist but undervalue pragmatic trials, leading to “interoperability failures” in Semantic Web apps. 

• iii. Using Inferred Framings to Support Interoperability Among Phenomenological Theories/Reference Models, and Construct Foundational World Theories for AI/Human/Hybrid Training, Robot Development, etc.: This pragmatic turn—echoing Comte’s “positive stage”—is emerging but missed earlier. Ontologies enable “interoperability” (e.g., FAIR principles’ “I” for interoperability via ontological commitments) and foundational theories (e.g., OWL for robot reasoning in Industry 4.0). For hybrids: 2025 FMEA enhancements use ontologies for AI-robot failure inference, supporting training via causal chains. Yet, speculative holdovers (e.g., duomining in OOO) limit scalability; a Kant-inspired evaluation (usefulness for “highest good” like equitable AI) could build robust worlds for AGI/robotics, avoiding ethical pitfalls like biased loan approvals. 

In sum, the era’s ontology work innovated phenomenologically (catalogs for AI) but speculated ontologically (ULOs without criteria), missing Heidegger’s call to let Being emerge from beings and Kant’s pragmatic limits. Future paths: Pragmatist ontologies (e.g., Peircean quietism for models ), evaluated by utility in resolving conflicts and enabling hybrids—transforming “missed” into “manifest.”

[John F Sowa]

Janet,

Thanks for that compendium of answers generated by an AI system.   By the way, which system did you use to get those answers?    And which phrases were your inputs, and which were the computer's responses?

My observations:  All the sentences sound plausible.  They seem to be generated from texts found on the WWW.   The authors and systems that are mentioned are well known to anybody who has been reading the literature on ontology and knowledge representation for the past 30 years.

But after reading that, I can't say that there are any new ideas in it -- nor any old ideas for which it provides any new insights.   I can't point to any paragraph that would tell anybody any significant insight about ontology, phenomenology, or their interrelationships.

If anybody asked me for a summary of that article,  I would answer in one word "Blah."

Can anybody else say anything better?   After reading that, can anybody state anything new that they learned from it?  Can anybody find any paragraph in it that they would recommend to anybody else?

John

 

---

From: "Janet Singer" <jsi...@measures.org>

[Janet Singer]

Below is a pdf of the entire session.

I’ll tell you which system I used only if you promise not to get diverted into politics : )

[Ravi Sharma]

John, Janet

Appears to be a query in Chrome Google?

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

On Sun, Nov 30, 2025 at 7:57 PM Janet Singer <jsi...@measures.org> wrote:

Below is a pdf of the entire session.

I’ll tell you which system I used only if you promise not to get diverted into politics : )

--
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/16B047F8-5523-4E14-B105-ABD87E9D2B93%40measures.org.

> On Nov 30, 2025, at 7:29 PM, John F Sowa <so...@bestweb.net> wrote:
>
> Janet,
>
> Thanks for that compendium of answers generated by an AI system.   By the way, which system did you use to get those answers?    And which phrases were your inputs, and which were the computer's responses?
>
> My observations:  All the sentences sound plausible.  They seem to be generated from texts found on the WWW.   The authors and systems that are mentioned are well known to anybody who has been reading the literature on ontology and knowledge representation for the past 30 years.
>
> But after reading that, I can't say that there are any new ideas in it -- nor any old ideas for which it provides any new insights.   I can't point to any paragraph that would tell anybody any significant insight about ontology, phenomenology, or their interrelationships.
>
> If anybody asked me for a summary of that article,  I would answer in one word "Blah."
>
> Can anybody else say anything better?   After reading that, can anybody state anything new that they learned from it?  Can anybody find any paragraph in it that they would recommend to anybody else?
>
> John
> 

> --
> All contributions to this forum are covered by an open-source license.
> For information about the wiki, the license, and how to subscribe or
> unsubscribe to the 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/936292f807d94366a85be597577b08fa%40df935591fe4c43cf99664a31d85f65ff.

--
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/16B047F8-5523-4E14-B105-ABD87E9D2B93%40measures.org.

[Alex Shkotin]

Janet,

I find bringing philosophical doctrines into our field counterproductive, for me. It's a different matter when, like Adam, there's a clear interest in philosophy.

If we ask a practicing ontologist what philosophical doctrine they adhere to, the answer will likely be none.

It follows that there is a system of ontological principles accepted by the majority:

Being is the truth of Being.

Being appears.

This is what immediately comes to my mind from Hegel.

It would be interesting to gather them all together, similar to GAAP. Like GAOP 🏋️

We, as a community of practicing ontologists, could work on this. Well, first, we could ask the AI about this.

And perhaps it would be worthwhile to highlight generally accepted methods for describing phenomena as a separate topic.

We begin with a description of the phenomenon, preferably a reproducible one, identify the being that is appearing, and use a theory of this being, if one exists.

How the matter appears to a human being is a fascinating topic. We are surrounded by geometric surfaces glowing in different colors.

Alex

пн, 1 дек. 2025 г. в 05:24, Janet Singer <jsi...@measures.org>:

--
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/2404743B-B7AA-47E6-906C-529D53E65683%40measures.org.

[Chuck Woolery]

Janet,

If a word does not lead to moral clarity of action.  And action isn’t taken.  WTF are words good for?

Just entertaining minds that prefer being RIGHT-  over doing what is right and now urgently NEEDED. 

 

As a biologist who detests the English language for its ambiguity...I joined this listserv for the purpose of prolonging our species sustainably on this goldilocks planet.   If the consensus on this website is only about protecting a word’s accuracy... Take me off.

 

cw

 

 

WARNINGS: FINDING CASSANDRAS TO STOP CATASTROPHES  By Richard A. Clarke and R.P. Eddy,  2017: https://cco.ndu.edu/PRISM-7-2/Article/1401978/warnings-finding-cassandras-to-stop-catastrophes/   The first 8 chapters detail the millions of lives and hundreds of billions of dollars lost to catastrophes,– natural and human engineered – due to people in power failing to act on the advanced warnings of experts.   The last eight chapters estimates the billions of lives and trillions of dollars that could be saved if humanity collectively works to prevent the other dire warnings now being given regarding other threats (some existential).  Chapter 11 “The Journalist: Pandemic Disease”.   Most instructive is Chapter 9.  It outlines three cognitive reasons why humans ignore such warnings.

Here’s a video of optimism if you dare watch it  https://www.rethinkx.com/videos

 

Chuck Woolery

Former Chair, United Nations Association Council of Organizations

Former Issues Director, Global Health Council.

Former Action Board member, American Public Health Association.

Author of 1996 and 1997 Congressional testimony warnings regarding threats to US and global bio- security.

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAFxxROQNhFCeubz%3D%3DM6VJnrTTZzqsaYQgV5mBvC5pZ2qk6DmbQ%40mail.gmail.com.

[John F Sowa]

Alex,

Ontology is a branch of philosophy.   Any and every decision about the top levels of an ontology is a philosophical assumption.  If you happened to notice, Barry Smith was the chief proponent of the top levels of BFO, and guess what his primary job happens to be.  If you haven't checked, he's a professor of philosophy at the University of Buffalo.

Aristotle was a major contributor to logic and philosophy and their application to any and every branch of science.  Kant was teaching both science and philosophy.  In fact, his primary job was natural philosophy, wich included Newton's Principia.

Kant was the first person to propose that the solar system began as a rotating cloud of dust that coalesced into a rotating disk, which then coalesced into the observed planets, of which the earth is one.  His major philosophical publications addressed the issues of providing a universal ontology for all the branches of science,

Summary:  Every theory of ontology is an application of logic and philosophy to an analysis of all the branches of science, including both the natural sciences and the social sciences.

John

_______________________________________

[Alex Shkotin]

John,

The situation with TLO is quite interesting. For a specific theory, for example, graph theory, to be applicable, I need to be able to map its primary terms to reality, so that the model constructed for my graph theory reflects the properties of reality.

In this nexus: theory, model of the theory, mapping the model to reality, where does TLO fit in?

Have a look at https://www.linkedin.com/pulse/theory-model-reality-alex-shkotin-dpnue

We need somewhere written down Generally Accepted Ontological Principles (GAOP), not philosophical doctrines. Is TLO a form of such text?

Alex

вт, 2 дек. 2025 г. в 03:20, John F Sowa <so...@bestweb.net>:

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

To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/ea930908ba1a4aef8e9f56e2ba58bb1a%40b274d66d3bf14b15b6901b52dd838133.