Knowledge Commitment

Per several knowledge engineering text books that I have read, there is this notion of "ontological commitment".

Ontological commitment is a concept used in philosophy, artificial intelligence, and information systems. It refers to the agreement to use a shared vocabulary, associations, and rules in a coherent and consistent manner within a specific context.

In simple terms, when you make an ontological commitment, you’re essentially saying “I agree that these things exist in the way we have defined them in our shared understanding and I will use these definitions, associations and rules consistently when we talk about those things.”

The same idea applies, in my view, to the knowledge represented in a knowledge graph.

What makes an ontology worth committing to?  Also, what precisely is your definition of the term "ontology"?  There are many ontology-like things that could be used for knowledge representation. My personal favorite is the theory.  A theory is a set of logical statements.

A machine-readable knowledge representation (such as an ontology or theory) for an area of knowledge specifies a formal, rich description of the:

• Important relevant terminology, concepts, nomenclature, jargon (a.k.a. TERMS) of the area of knowledge being represented such that (a) one term can be effectively differentiated from every other term and (b) terms can be organized into useful categories. This includes simple TERMS and complex STRUCTURES (assemblies of TERMS that are grouped for some meaningful purpose as per the Atomic Design Methodology).
• Important relevant relationships among and between TERMS (a.k.a. ASSOCIATIONS such as "is-a" and "has-a") for the area of knowledge.
• Statements distinguishing TERMS, refining definitions and relationships (important relevant assertions, constraints, and restrictions) including logical operations (NOT, AND, NAND, OR, NOR, XOR, XNOR) and mathematical operators (addition, subtraction, multiplication, division).

The above are the "building blocks" or the "elements" used to create a machine-readable representation.  One might expect that the machine-readable knowledge representation, such as an ontology, will be:

1. Encoded formally in a declarative knowledge representation language. (such as XBRL, RDF+OWL+SHACL, GSQL, SQL, PROLOG, DATALOG)
2. Syntactically well-formed for the chosen language, as verified by an appropriate syntax checker or parser. (such as an XBRL processor)
3. Logically consistent, as verified by a language appropriate reasoner or theorem prover for the language chosen. (such as DATALOG or PROLOG or a semantic reasoner)
4. Will meet business or appropriate application requirements as demonstrated though deliberate, extensive, rigorous testing.

One might further expect that if the logic is formal encoding in one language that, if converted to some other formal encoding in another language; the logic would not be affected (i.e. the logic is the same for each formal encoding; the choice of encoding does not affect the logic of the representation).

Additional Information:

• PROOF
• Book of Proof
• All About Knowledge
• The Rise of the Knowledge Graph