Semantic-tech through

Descriptions & Situations (my review)

 

I have read and investigated number of research papers by Aldo Gangemi, and I have to say, to my knowledge and limitation of a single person to survey entire field of ontology and semantics, he is the top Ontologist at the present time. By reviewing this particular paper, I was able to aggregate and reinforce some thoughts and ideas from computer science and language theory, Type theory, philosophy of technology, information theory and even Wittgenstein. Let's jump to my commentaries on this outstanding piece of research.

 

Understanding the Semantic Web through

Descriptions and Situations

 

Aldo Gangemi 1 and Peter Mika 2

 

1. Laboratory for Applied Ontology, Institute for Cognitive Sciences and Technology,

National Research Council, Rome, Italy

 

2. Vrije Universiteit Amsterdam, Amsterdam, The Netherlands

 

 

Hossein M. Bojnordi's commentaries and notes on the above paper :

 

(page 2) In connection with the paper "An Information-Theoretic Formalism for Multiscale Structure in Complex Systems" which is one of my favourite papers on "Information Theory" :

 

"We have developed and are exploiting an ontology of contexts, called Descriptions and Situations (D&S), which provides a principled approach to context reification through a clear separation of states-of-affairs and their interpretation based on a non-physical context, called a description."

 

This "non-physical context description" relevant in the D&S plugin to the DOLCE ontology, in my perspective, is an equivalent to the "Augmented reality" object of the paper I mentioned earlier.

 

At the introduction part (pages 2 and 3) and through the following phrase, I can see some patterns (may rather be not terribly obvious) that can connects relatively disparate disciplines and bring some added value.

 

"Integrating theories of communication (linguistic theories) with theories of interpretation (computational semiotics), such an ontology is in fact an attempt to describe ontology-based communication on the Semantic Web. The inclusion of a theory of interpretation in this ontology is crucial in relating the contexts of ontology use to the communication system."

 

This would be my transformation and web of ideas when reading the last paragraph which, in my opinion, connects the ontological works of Aldo Gangemi with a decentralized Internet (enabled with some distributed ledger technology such as the "Web3" and "Internet Computer" initiative), and it also describes the mediation criteria and (possibly) technology (My contribution). Let's do It:

 

An "ontologically" and semantically empowered and enriched "computer Language workbench (a workbench to enable defining and evolution of domain-specific languages), to define and evolve formal language workbenches for those of Ontology definition community (for example, creating a language workbench for D&S plugin of DOLCE ontology), that consequently, the workbench artifacts (Models, Codes and Programs) can be used in a distributed, crowd-sourced Internet (a block-chained Internet for example), to build future-proof applications and services, and an infrastructure that will guarantee our civilization's future.

 

A point to add is, machine learning techniques also can be utilized (they sure will be), but with one condition that they are used with guide and complementation of reasoning and inference engines that are themselves enabled by logic, semantics and ontological methods. In this way we can assure transparency (avoiding black boxes), aligning with evolving laws, regulations and principles underlying human (or extra-terrestrials') ethics. Related to this topic, one reasonable assumption, I think, is that structured, formal human-logic (mathematical logic) will be understood, and more importantly "maintained to be understood" (including tracing backward) for potentially infinite time in the future. The point is that we shall never come to a case as Nick Bostrom explains , a dominantly super-intelligence that by means of some unknown or partly known chains of reasoning plus probably by some amounts of unavoidably random events and selection effects, the dominant agent (super-intelligence) decides to make paper-clips, and ONLY paper-clips as much as physically available material in the universe allows. That would be very unfortunate for us humans, even our tasty flesh will be transformed into paper-clips! How sad and meaningless ending.

 

THEN, let's automate the good stuff! I will elaborate later.

 

 

(page 3) At the "motivation" part of the paper and on its very first statement: "Foundational ontologies in WonderWeb are ontologies that contain a specification of domain-independent concepts and relations based on formal principles derived from linguistics, philosophy, and mathematics."

 

So, I would say, even more important than "formal description (or formal specification)" of an instance, is a "formal description of principles" of at-least "relatively" large domain, if not--a formal description of principles of a whole upper (foundational) ontology.

 

 

(page 5) It is important to note a clarification of minds (as I've done myself) that the whole effort of DOLCE together with D&S extension to it (or any augmented, meta information/data activity, reasoning, processing, storage or communicating through digital channels) about original objects in the world which are organically at the universe level, IS A META-EFFORT, A META DATA effort to a natural, organic, human-related activity which is a or collection of thoughts. The original activity (thoughts, concepts, intuitions) in the brain of an organic agent, is at a PHYSICAL REALITY LEVEL namely THE UNIVERSE. We can define a convention for ourselves that only second-order activity of an organic agent in the universe producing, communicating, storing, and inferring data/new data, CAN BE CALLED META-[DATA/INFO/KNOWLEDGE/PROCESSING], A META-ACTIVITY.

 

Our added value is, a utilization of materials available in the universe (for example to build computation and storage machines, networking etc.) and also utilizing our other important (the most important) assets namely thoughts and intuitions (initially triggered by sensory data, and independent of the nature and philosophical questions around them), conceptualization and abstractions, processing of input data into our brains and reasoning about them and infer new information and knowledge from them, accumulate and preserve them (storage of data whether digitally or physically, however the new-age advantage is to do it digitally) and communicate knowledge (also advantage is to do it computationally), and finally to create noble things and ideas, to innovate future of the civilization and ultimately, --(possibly?)-- to create meaning and purpose for ourselves.

 

(page 5) Here Aldo Gangemi declares: "A description is an entity that partly represents a (possibly formalized) theory T (or one of its elements) that can be “conceived” by an agent."

 

From an Information Theory perspective: A "description" is "An augmented reality information object", "A meta-object about object", "A meta-data about real object", whatever "real" means.

 

(page 5) Referencing to the philosopher of "information systems" Kristo  Ivanov, that has put forward the question that whether "information" in the society is digital or concrete? (I HAVE TO LOOK UP AGAIN THAT MATTER AND THEN ELABORATE)

 

It is important to note a clarification of minds (as I've done myself) that the whole effort of DOLCE together with D&S extension to it (or any augmented, meta information/data activity, reasoning, processing, storage or communicating through digital channels) about original objects in the world which are organically at the universe level, IS A META-EFFORT, A META DATA effort to a natural, organic, human-related activity which is a or collection of thoughts. The original activity (thoughts, concepts, intuitions) in the brain of an organic agent, is at a PHYSICAL REALITY LEVEL namely THE UNIVERSE.

 

Therefor "situation" is a (partial) or collection of (but still partial) "meta-data(s)" , and "description" is a (partial) or collection of (but still partial) definition(s) and description(s)--We might be able to use "representation" only, I think, with considerable amount of forgiveness, because descriptions are almost always incomplete. Only a formal sign ( A linguistic symbol) can totally--totally here means equal to the agent's intuition of the "object" under consideration--

 

I want to argue here that a symbolic (a sign or concatenations of signs, in a formal language) can (isomorphically) represent a real "object" in full (not partial). Because we take the symbol (a sign or concatenation of signs) as a "black box" entity that anyway will represent the real, concrete, physical "object" (whatever "real" means), however, I will attack my own saying of "whatever", and want to argue that "It" is not actually "whatever"! It is an "equivalence relation" between an "object" under consideration of an agent and the totality of "intuition" (a Block-Intuition) of that agent about that "object". This is all we have, the only "channel" available. One might argue about quality of intuitions. Is quality of intuition of Einstein about symbolic elements of relativity equations, the same as a physics undergraduate?--I MUST ELABORATE ON THIS LATER

 

WHAT ABOUT symbolic representation and intuition of "abstract" objects. Is it partial (incomplete) or implicitly total, or absolutely total? (Later)

 

I think, the Wittgensteinian concept of "common life form" will aid on this matter but be careful about that! Whether abstract or concrete, anyway we eventually construct a common (differs in quality) abstract (intuition) of--NOT--knowledge (According to Bertrand Russel we only have a theory of descriptions not "causes". The only knowledge we have is the knowledge of our "own existence" or existence of our "agentive" thought), but a common abstract "description(s)" again about abstract or concrete entities (no difference--according to my description of "intuition" of "common life form" about "external" (relative to an agent) entities or objects and "Internal" (relative to the agent--common life form internal logic, identity and basically (almost!) "common form inner-life" of species or sub-species or even for example "community of professional physicists". there are levels of abstractions that can be built on top of each other to be able to encapsulate concepts and chain them at higher level abstractions,-- to more efficiently reason about the world). A common, according to the "common life form". Care must come on the "quality" of this "intuition". we might be forced to modify our taxonomy of the bulk "common life form" into more specialized ones, like "community of professional physicists" to account for having the "quality" I mentioned in the Einstein example. But, still, they may not be of (mathematically) equal "quality". We might be able to list a long criteria, pre and post conditions, a comprehensive list of properties, attributes and predicates, to measure the "quality" of intuition (Or understanding) of members of "community of physicists" about a particular element of an equation (A linguistic sign, or/and concatenations of signs). However it will be increasingly difficult to capture ALL necessary of these attributes. It might be practically untraceable, and also impractically contain uncountable (Or even infinite) intricacies that actually could related to the "human life form", structure and necessities of physical reality (including the life form).

 

However, we still might be able to define criteria of a or collection of consistent and coherent frameworks (Or workbenches), for the sole purpose of bringing the "precision" measure to a "enough" level. Common life forms do not need an infinite precision about perception, description and intuition of "objects", but they need an "enough" level (To navigate, reason and live around each other).

 

Here, Model M, itself is a "meta-" object, a (partially) "defining" "augmented" "meta-object" for a concrete "object" in the universe (whatever concrete means).

 

(page 6) Afterwards Aldo Gangemi says: "Intuitively, when a description is applied to a state of affairs, some structure (a “situation”) emerges."

 

I believe my recent comments, is easier and better explain the relationship between "state of affairs" and "description". However, they might slightly be different at some aspects.

 

(page 6) SD or "Situation Description" is formalised in the paper as follows:

 

 

SD ( x ) = D ( x )

 

Satisfies ( x , y ) = S ( x ) ^ SD ( y )

 

satisfiedBy ( y , x ) == satisfies ( x , y )

 

Any x . S ( x ) = thereExist y . SD ( y ) ^ satisfies ( x , y )

 

 

**situationDescription (x)  --->  is itself a description D(x) (actually subset of D)

 

 

(page 7) also,

 

                      PiD ( y ) = D ( y )

 

 

**Allocation (An Intuitive Agent selects a situation by describing it, "description") of (or selection of) subset of D, SD (situation description) predicate(i)D.

 

And,

 

        Negation ( PiD ( y ) ^ SD ( y )  )

 

 

        selects ( x , y ) = PiD ( x ) ^ Pi ( y )

 

 

**This (the "selection") is the intentionality that attributed to an agent. (see previous paragraphs)

 

 

 

(page 7) May It can be called "intentional variables" or "intuitionistic variables", pointer to the intentionality of an "agent", and process of "intuition" or more generally of "understanding".

 

They are temporary because analogues to a computer programming language, they should be put into bin after use (like "Garbage-Collection" concept in computer languages).

 

 

And finally, the last paragraph of page 7 : "... finding a component of SD for each of most general concepts in DOLCE."

 

**Extending DOLCE with D&S: It will enriches the DOLCE theory and concepts with the "new powers" of D&S extension (A Plug-in to DOLCE).

 

 

(page 9) Just before chapter 3, we come across this remark : "Such referencing requires a communication setting including information objects. A communication setting can be understood only within a semiotic framework, and the circle is complete."

 

**That would include : theory of signs, body language and gestures (including absence of action, convey of a meaning by doing nothing), vocal and sound (including a spoken dialectic, noise, distraction, imitation of people, animals or nature, singing), written language (standard in a community, recorded by any means, on trees, rocks, any sign meaningful to a group), utilizing material and things (like sticks, fire and smoke, using (or abusing) plants and animals, language and linguistic techniques in general.

 

 

(page 12) What about theories of communication and interpretation? In this part of the paper, the objection is about absence of a theory of interpretation from Jakobson's model, and the suggestion is to turn to the models of semiotics.

 

**A model of semiotics here, can be a :  Firstly, Model-Transformations or second, Model-Interpretations , in a computer language theory that formalizes the creation and conservation of semantics (meaning).

 

End of the commentary.

Hossein M. Bojnordi