Re: [ontac-forum] Semantics and Ontology and Semiotics

On Sat, May 27, 2006 at 05:07:18PM +0300, Azamat wrote:
> Chris's sophistries are more shocking. There is a rather clear
> conditional statement, with the plain vanilla meaning:
>
> <AA> the whole enterprise of semantic technology is an otiose
> undertaking and expansive academic mystification without understanding
> of the nature of meaning, its critical dimensions, mechanisms and
> algorithms of representation in computable forms.

The word "conditional" appears to have a meaning for you that it doesn't
have for anyone else; this makes for difficult communication.
Typically, the conditionality of a claim is made explict by the word
"if" or the like, or by the use of the subjunctive mood.  By contrast,
you use the straightforwardly declarative "is" to characterize "the
whole enterprise of semantic technology" as "an otiose undertaking <blah
blah blah>".  That you find floating somewhere in this baldly
declarative statement a "clear conditional" is as mystifying and bizarre
as those who claim to see images of the Virgin burnt into their morning
tortillas.

> In fact, the statement neither derogated nor humiliated anybody, 

Of course it did.  It says, as clearly as can be, that the enterprise of
semantic technology is otiose, steeped in "academic mystification"
(whatever that is supposed to mean exactly), and that it -- i.e., (who
else?) anyone contributing to that technology -- lacks understanding of
the nature of meaning and representation.  I have no objection to your
thinking that, but now to retreat from the plain implications of your
remarks is disingenuous in the extreme.

> It just states [no comprehension of meaning no semantic machines],

But it doesn't "just state" that.  It insults everyone involved in the
enterprise.  Do you really not see that?

> like no meaning no truth. It just states that the meaning (and
> significance) must be viewed with every its elements, denotation cum
> connotation, sense cum reference and representation; with every its
> aspects: syntactic, semantic, and pragmatic. 

These are banal claims that applied most notably to the early days of AI
and knowledge representation when the confusion of sign with thing
signified, denotation vs connotation etc was widespread.  However, these
confusions were also identified early on in the history of knowledge
representation (e.g., Woods's "What's in a Link", Hayes's "In Defense of
Logic, both from the 1970s) and, while they admittedly continue to crop
up as more and more (enthusiastic and well-meaning) people with very
little knowledge of the history jump on the ontology bandwagon, the
AI/KR/ontology community has been quite well aware of them ever since.

> If you dont know this and cant apprehend the distinction between
> meaning and significance, between the formal logical semantics and the
> real ontological semiotics, then refrain from publishing as the W3C
> recommendation such a raw stuff as the RDF Semantics. At least, try to
> avoid the screaming misstatement: This is a specification of a precise
> semantics, and corresponding complete systems of inference rules, for
> the Resource Description Framework (RDF) and RDF Schema (RDFS).
> [1]http://www.w3.org/TR/2004/REC-rdf-mt-20040210/

And thus your central confusion.  You think that a formal semantics for
a formal language is somehow a complete characterization of the
semantical purposes to which the language is put.  In fact, such a
semantics is only a rigorous formal model of the general semantical
properties of the given language.  It provides in particular a precise
characterization of the manner in which the meanings of complex
expressions are determined by the meanings of their syntactically
simpler parts.  But such a formal semantics is in general almost
completely silent on the natures of the entities which the language can
be used to characterize.  And, in application, these will be real world
entities and their properties and relations.  Thus, using your example,
the semantics of RDF only describes how it can be semantically
meaningful that one and the same URI can appear in all three roles of
subject, predicate, and object in different RDF triples.  This syntactic
flexibility is in fact critical in the anarchic representational world
that is evolving on the Semantics Web, where one cannot ensure
conceptual and representational uniformity or enforce a single
ontological perspective.  Thus, in particular, one can find classes
playing not only predicative roles but the roles of individuals that are
themselves subjects of predication (e.g., the class HUMAN is an instance
of CLASS); similarly, relations can also appear as relata of other
relations, or even of themselves (e.g., the MEMBER_OF relation stands in
the MEMBER_OF relation to the class RELATION).  RDF's semantics --
ironically, penned by the very Pat Hayes noted above, who has been
acutely aware of the "confusions" of which you speak as long as anyone
in the AI/KR community -- reveals precisly how these semantical
phenomena can be represented in RDF.

> <CM>... But just what is it you are castigating here?  *No one* thinks
> that a merely formal semantic system with only mathematical or formal
> logical axioms is of any use.
>
> It is not quite honest answer, all formal logicians think just like
> this, read the RDF (formal) Semantics. 

Exhibit A of confusion noted above.

> Or read the promulgation of the formal (dolce sweet) ontologists from
> Italy. Or, here in EU, see the formal semantic web projects
> delivirables, if any,  eating most FP 6 funds allocated for
> Information Society Technologies. 

Exhibits B and C.

> And read this promulgation:
>
> The Laboratory for Applied Ontology (LOA) performs basic and applied
> research on the ontological foundations of conceptual modeling,
> exploring the role of ontologies in different fields, such as:
> knowledge representation, knowledge engineering, database design,
> information retrieval, natural language processing, and the semantic
> web. The group is characterized by a strong interdisciplinary approach
> that combines Computer Science, Philosophy and Linguistics, and relies
> on logic as a unifying paradigm. [2]http://www.loa-cnr.it/index.html

And D.

> How many times it should be said that the formal logic has nothing to
> do with the real world, 

It shouldn't be said at all, as it is an abject confusion.

> but only with special logical forms and constructs: propositions,
> individuals, predicates, and formal inferential rules. The logic is
> nothing but a formal tool (a servant of his masters): of ontology, of
> science, of mathematics, and of semantics. 

That you think anyone disagrees with you is further evidence of the
confusion in question.  Virtually no on in the ontology community
disagrees with this.

> Only ontology can handle the whole world as the largest hierarchy of
> distinct kinds of things organized by distinct types of real
> relationships. For the goal of ontology is to formulate the overall
> patterns and fundamental laws of the universe, with its practical role
> to set the world models, rules, and reasoning algorithms for advanced
> information technology.

Again, no one disagrees.

> Ontology is surely not: 'an explicit conceptual model with formal
> logic-based semantics', the false assumption, a major obstruction to
> the success of the promising ontological projects as the semantic web
> and SUO and ONTAC.

Well now, that depends on what one means by a "conceptual model".
Granted, if it implies that an ontology is just whatever anyone thinks,
and that everything is up for grabs, then indeed that is a misimpression
to be corrected.  Most people don't mean that.  All they mean by it is
something long these lines: a collection of objects, concepts and
relations whose logical properties and connections are expressed through
a set a axioms in a formal language with a clearly-defined formal
semantics.  No one I know thinks that such axioms are not meant to
describe the real world -- though there is interesting and reasonable
debate about what exactly it *means* to describe the real world, most of
which, however interesting, is philosophical sound and fury.

Please note that I am not debating here, and I have no interest
whatsoever in continuing this discussion.  I have simply felt compelled
to respond to your frequent lengthy posts, which I have often found
misinformed and misleading in content and inappropriate in tone.

Sincerely yours,

Chris Menzel

Received on Sunday, 28 May 2006 18:59:04 UTC