Re: [ontolog-forum] Ontology vs KR

[Ali H <asaegyn@xxxxxxxxx>]

Dear Ed,

Wow, seven days and I have a lot to catch up. Firstly, thank you for taking the time to respond thoughtfully.

I agree that titles can be informative, and one can glean something from Brian Smith's intent given the context in which he wrote those words. 

That said, my interpretation of Smith's use of "process" in his quote is that it applies to the entity that is processing the knowledge, not necessarily a specific process. In the context of computational ontology as it's evolved in the past 30 years, I would posit that the system in which the ontology is situated in, is an adequate analogue to Smith's "embodied intelligent process", which is the "process" he refers to.

I suppose my initial question suffered from an ambiguity as to whether it referred to an ontology as an artefact or ontology as a field. And on reflection this difference was not appropriately acknowledged in the question.

As I'm sure you're well aware, there is copious literature in AI and ontology on differentiating the various levels of representation (see [1], [2], [3] among others) - and I appreciate and see the value in these distinctions. 

I agree that ontological analysis should not muddy the waters by conflating various issues. That said, as an ontological engineer or a knowledge engineer or someone who is creating an ontology supposedly for some purpose other than an end in itself, this is clearly inadequate. That is to say, considering ontological analysis on its own is useful, but inadequate in building an ontology-based system.

Any particular ontology artefact will reside in some system. It will be reasoned upon by some process, which will be at some level of abstraction, procedural. 

If I were to rephrase the original question, I'd wonder whether / where those who practice ontology (in the computational sense) draw the boundaries alluded to in Smith's excerpt. Secondly, I'd wonder whether it makes sense to consider an ontology artefact modulo epistemic and/or procedural constraints. 

Certainly, it would appear that one of the dominant forces in ontology engineering (sem web) has been to internalize and prioritize procedural concerns into the core of ontological analysis and consequently representation.

I'm unsure if I'm articulating my intention clearly, but greatly appreciate your response. 

Basically, I would agree that ontological analysis should focus defining what is (though it seems that process is a core part of what is), but I wonder if this is then projected on the creation and application of the ontology artefact too stringently. It seems many of the issues and difficulties and quests for "true ontologies" or "ontology reuse" become much clearer when one expands the scope of the discipline to allow broader concerns, coming from epistemology, KR or whatever fields inform your view.

[1] Ronald J. Brachman "What's in a concept: structural foundations for semantic networks" International Journal of Man-Machine Studies, 127-152, 1977, 

[2] Allen Newell "The knowledge level" Artificial Intelligence, 1982

[3] Nicola Guarino "The Ontological Level: Revisiting 30 Years of Knowledge Representation" in Conceptual Modelling: Foundations and Applications 52-76, 2009

Best,

Ali

On Wed, Oct 1, 2014 at 7:22 PM, Barkmeyer, Edward J <edward.barkmeyer@xxxxxxxx> wrote:

Ali,

 

I would say that the title of Smith’s paper suggests a major difference in basis:  “Semantics of a Procedural Language” as opposed to “knowledge about some world”.  Ontologies are indeed representations of “a propositional account of knowledge”, but not necessarily knowledge exhibited by any particular process.  And the only behavioral role of an ontology per se is the _expression_ of the knowledge.  It may well be that the purpose of developing the ontology is related to engendering some behavior, but the purpose and the use are not the ontology.  By comparison, the purpose of a corpus in a procedural language (a program) is to engender behavior, and the meaning of every element of the language is tightly coupled to behaviors. 

 

Conversely, if you look at a formal ontology for procedures, like Z, it becomes apparent that the nature of procedural statements is not axiomatic; it is only “propositional” if one defines predicates that state causative relationships.  That is, “if A then B” does not mean “if A holds then B holds”, but rather “when A holds in some state S1 of the world of interest, then B (must) hold in some subsequent state S2”.  And the “(must)” is an important part of the semantics.  It is not a statement that S2 follows S1 “as the night follows the day”, but rather that a behavior that implements state S2 is required/”engendered”.

 

All of this is to say that it is possible to construct ontologies for the formal semantics of procedure (and such ontologies exist in the BPMN semantics and in the Z language semantics, for example), but the formal semantics of procedures is not fundamental to ontologies.

 

-Ed

 

 

From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx [mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Ali H
Sent: Wednesday, October 01, 2014 6:39 PM
To: [ontolog-forum]
Subject: [ontolog-forum] Ontology vs KR

 

Hello all,

 

I know this is a topic that has been tread over many years and many times, but I recently came across this statement from Brian C. Smith in [1]:

 

Any mechanically embodied intelligent process will be comprised of structural ingredients that a) we as external observers naturally take to represent a propositional account of the knowledge that the overall process exhibits, and b) independent of such external semantic attribution, play a formal but causal and essential role in engendering the behavior that manifests that knowledge.

 

Why is this definition never proffered when discussing "what is an ontology"?

 

It seems to me that those in the field of ontology focus on (a). 

 

Do most (formal) ontologists consider Ontology to be (a), and not (b)? If so, why not?

 

Lastly, I understand that in pantheon of AI sciences, Ontology is often suggested as a sub-discipline of KR - yet why is there such little cross over from KR to Ontology - or am I simply misinformed (c.f. FOIS vs KR or CommonSenseReasoning as part of AAAI etc) ?

 

[1] Smith, Brian C. (1985). "Prologue to Reflections and Semantics in a Procedural Language". In Ronald Brachman and Hector J. Levesque. Readings in Knowledge Representation. Morgan Kaufmann. pp. 31–40. 

 

--

.

(•`'·.¸(`'·.¸(•)¸.·'´)¸.·'´•) .,.,

_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
 

--

.

(•`'·.¸(`'·.¸(•)¸.·'´)¸.·'´•) .,.,

_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J    (01)