[Top] [All Lists]

Re: [ontolog-forum] language vs logic - ambiguity and startingwithdefini

To: "ontolog-forum" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "doug foxvog" <doug@xxxxxxxxxx>
Date: Mon, 20 Sep 2010 03:04:39 -0400 (EDT)
Message-id: <51007.>
On Sun, September 19, 2010 16:55, Rich Cooper said:
> Hi Doug, Ferenc,    (01)

> This is an interesting thread.  My comments are below:
> -Rich    (02)

> Sincerely,
> Rich Cooper
> EnglishLogicKernel.com
> Rich AT EnglishLogicKernel DOT com
> 9 4 9 \ 5 2 5 - 5 7 1 2
> -----Original Message-----
> From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
> [mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of doug foxvog
> Sent: Saturday, September 18, 2010 10:26 PM
> To: [ontolog-forum] ct
> Subject: Re: [ontolog-forum] language vs logic - ambiguity and
> startingwithdefinitions
> On Sat, September 18, 2010 6:16, FERENC KOVACS said:    (03)

> > > I believe that core ontology concepts are objects, properties and
> > > relations.    (04)

> > I suppose you mean classes of objects.  Is the distinction between
> > properties and relations that properties relate objects to datatypes
> > while relations relate multiple objects or have more than two arguments?    (05)

I'm not sure what the response below has to do with these questions.    (06)

> There are two ways to go.  You can believe you ALREADY know truth and
> start
> with classes to build instantiated objects, or you can start with objects,
> and group them into classes so that the objects and classes can be
> validated.  IMHO, the second way is far superior, and certainly more
> realistic in most practical applications.    (07)

If one is exploring a totally new topic and not coding from specifications,
using reference works, or working with subject matter experts, and one had
a large data base of instantiated objects, such a bottom-up approach
would have value.  But determining all the possible relations for the
generated classes would be more difficult than inheriting many from
more general classes.    (08)

> Assuming you know everything about the classes to begin with    (09)

This seems like a straw man.  It is certainly not the way an
ontologist would work.    (010)

> is very misleading and causes you to find
> exactly what you imagined instead of exactly what is there.
> That is the way human infants seem to learn.    (011)

Sure.    (012)

> Why should a set of axioms, though usually
> small enough in number to conveniently support a theory, be expected to
> correctly enumerate the details unless the topic is some abstruse, but
> utterly simle mathematical theory.  I prefer a realistic appraisal of
> reality that ends up in a math model, as compared to starting with a math
> model that ends up missing reality and being unable to prove or disprove
> the objects' memberships.    (013)

No such expectation.    (014)

> > This sounds fine to me, so long as the concept of object needs not be
> > physical and can include events and situations.    (015)

> > Personally, i don't see the need for distinguishing properties and
> > relations in an ontology, although many ontology languages do make
> > that distinction.  This seems a language-dependent distinction to me.    (016)

> Agreed, but that is the way convention has proceeded in math logic.  Is
> there a compelling reason to not distinguish between properties and
> relations?    (017)

There can be efficiency improvements if a class of binary relations has
its second argument restricted to data types such as integers, floats,
characters, and strings.  This is language dependent, and i don't find
it compelling to make a distinction in an ontology.    (018)

> The arity is, as you suggest, the only difference between
> properties and relations by established convention.    (019)

This is a second distinction.  Again, binary relations can be more
efficient.  For the Semantic Web, RDF is based on triples, making
higher-arity relations difficult to represent.    (020)

> > > In fact, the intitial state    (021)

> > Are you referring to the initial state of an ontology in the process
> > of creation?  Or are you referring to the state of a reference ontology
> > whose terms are used to make statements in a knowledge base with no
> > additional definitional statements?    (022)

> Consider the initial state of the self generating explosion Ferenc is
> referring to.  The initial state is whatever is needed to hold the axioms,
> and the added state is the automatically generated (unrealistic in many
> cases) set of objects generated from the theory.    (023)

Fine.    (024)

> Again, predefining terms seems like putting the cart before the horse.
> Why
> assume a Poisson distribution (or pick another one) instead of using the
> actual empirical data that we get so much of with current technology?    (025)

Somehow the data does need to be initially classified.  Databases have
columns defined.  System specifications define the types of things they
refer to.    (026)

> > > is an object which is a unity of them and it is
> > > exploded through a number of mental operations. (Examles wanted?)    (027)

> I'm not sure what you mean by an ontology exploding.  Do you mean
> deriving all statements which are derivable from the initial statements
> in the ontology?    (028)

> > If so, i would suggest that this "explosion" could be carried out by
> > formal logical operations -- they wouldn't need to be mental operations.    (029)

> So long as the formal logical operations are not confused with the
> reality of objects actually experienced.  Yet again, the theory is
> only approximate,
> and was chosen by someone with limited personal experience in most cases.    (030)

After working with actual data, additional axioms can and should be
added.    (031)

> With the learning curve (see the book "Bionomics"), every doubling of
> experience leads to about a twenty percent improvement in efficacy.    (032)

> > > With the help
> > > of these categories I can semantically analyze natural langauges and
> > > create an ontology that integrates the currently different domains.    (033)

> > > In this approach
> > > axioms and the concept of events are not of primary interest, because
> > > verbs are seen as the representations of relations    (034)

> > This is not necessarily the best way to model verbs.  In languages which
> > use
> > verbs like English does, a verb indicates the occurrence of either a
> > situation or an event, with subject, direct object, indirect object, and
> > prepositional phrases indicating relations between the event and event
> > participants.    (035)

> > Features of the events/situations can certainly be modeled by relations
> > that
> > ignore the events themselves, but this would require multiple relations
> > to be defined for each verb depending upon what other phrases happen to
> > be in the sentence.  Using this technique makes it difficult to add more
> > information about the same event and requires multiple rules to
> > inter-relate the multiple relations that represent the same verb.    (036)

> >   Jill threw the chair.
> >   Jill threw me the chair.
> >   Jill threw the chair through the window.
> >   Jill threw the chair yesterday.
> >   Jill threw the chair to kill the toad.
> >   Jill probably threw the chair at the toad.    (037)

> Jill threw a fit    (038)

This is a different denotation of "threw".    (039)

> and jumped up on the lawn chair when she saw the toad cross
> the road toward her.  Examples abound in common corpora, so why axiomatize
> everything at the beginning when you can take into account the actual
> usage
> of the language, which is dependent on so many factors - the speaker, her
> cultural background, her experiences, goals, values, perceptions, and all
> those things that are NOT axiomatizable in realistic practice.    (040)

I agree.  There is no need to axiomatize everything in the beginning.
I was arguing against the mapping of verbs into relations.  It appears
you would support this.    (041)

> > IMHO, the verb "threw" represents a different relation in each of these,
> > but
> > each use can unambiguously represent a throwing event, with multiple
> > relations to generate depending upon the other parts of the sentence.    (042)

Thus don't create/axoiomatize complex predicates for various uses of
a verb.    (043)

> > > (hence not limited to Boolean operators).    (044)

> > Why wouldn't the relations have Boolean truth values?  Is the point to
> > allow for probability descriptors?    (045)

> Probabilities, fuzziness, certainty, value, cost, goals, suppositions,
> there
> are a never ending set of value types associated with common statements.
> Why limit them to Booleans?  Math logic has always done so, but even the
> Boolean case results in true theorems that can't be proved and false
> theorems that can't be disproved (Godel again).    (046)

No need to limit them to Booleans, but no need to ban Booleans when they
may be appropriate.    (047)

> > > Therefore the issuse of disambiguation as for dictionaries is a
> > > futile exercise, as the defintions used are sometimes incomplete and    (048)

> > Many dictionary definitions are certainly incomplete, but that does not
> > mean
> > that they do not make true statements constraining the meaning of
> > the thing they define.    (049)

> But are the constraints VALID?    (050)

For "accurate" dictionary definitions, they generally would be.    (051)

> > > irrelevant in semantic terms,    (052)

> > I have not seen this.
> > > this is why you cannot "merge" them (should try to integrate
> > > them instead) as they are not in compatible forms (content)    (053)

> If you are referring to multiple definitions from the same source,
> they shouldn't be merged because they are describing different
> denotations of the word.  Each definition should denote a different
> concept.    (054)

> Its a rare occurrence when a concept has only one definition or a
> definition designates only one concept. Consider the many-to-many
> mapping of words to synsets in WN+VN.    (055)

Here you are referring to words and synsets, not concepts and definitions.    (056)

> > If you are referring to multiple definitions from different sources,
> > integrating multiple definitions of the same meaning of a word is
> > certainly appropriate.  However, sometimes such integration can be
> > handled by merging.    (057)

> > > and they are not modular either.
> > > You must accept that such a new ontology should be dynamic as
> > > many of you already suspect.    (058)

> > If the ontology is used to interpret NL text in an open area, the
> > ontology would be incomplete and should dynamically be expanded.
> > If the ontology is to be used to express the information in a data
> > base that has been in constant use for years, dynamaticity is not
> > so crucial.    (059)

> Databases in constant use constantly uncover situations which cannot be
> handled within the database without expanding it.    (060)

Agreed.    (061)

>>> In math logic domain there is a kind of definition - an abbreviation
>>> when
>>> they introduce new symbol saying for example:
>>> definition
>>> t≤s denotes t<s or t=s
>> In my "semantic analysis" this is formalization, a mental operation of
>> the
>> relation between two objects as indicated.
> > This formalization/definition is a logical operation between two
> > expressions (one of which is a disjunction,  I suppose the expressions
> > could be called objects.    (062)

> Expressions are not objects    (063)

This depends on one's definition of "object".    (064)

> - the interpreters we build postulate imaginary
> objects to fit the theories, but they often miss the valid realities
> especially in language use.    (065)

Sure.    (066)

> > > The commonsense transcript is that an
> > > object (to be specified, otherwise it does not make sense) is smaller
> > > than
> > > another object after comparison and a few other operations also
> > > required to arrive at that result in formalization.    (067)

> > You are stepping beyond the definition as given, to interpret what the
> > definition means.  This is intentionally moving beyond logic.    (068)

> I think that is Ferenc's point - logic is not enough to explain language
> use.  More is needed.  (Ferenc, correct me if my assumption is not what
> you > meant.)
> >> In doing this I used the mental
> >> operation called interpretation, the reverse of formalization.
> >> For any message (statement) to make sense it is necessary to be
> >> complete,    (069)

> > Why can't a message tell merely a portion of a fact, instead of being
> > complete?    (070)

> I have to agree with Doug on this one.  Many statements only make partial
> sense, and some make no sense at all.  So interpretation is a very
> subjective process that, IMHO, may not even be possible to formulate
> without
> great loss of information compared to the glorious ambiguity of language.
> >> which means that if it has (as it should have) a verb in it,
> >> then it should have person,
> > Here, i assume you mean grammatical person, not requiring the message
> > to relate to a person.  Person, in this sense, is a linguistic feature
> > of sentences of many languages, not necessarily relating to a feature
> > of the meaning being discussed.    (071)

> Disagree - I think he means the interpretER, not the declension and
> conjugation sort of person.  The syntactic parts are for convenience of
> information transfer to a DIFFERENT interpretER who may have a totally
> different view of the same sentence in her own interpretation.    (072)

If this is his meaning, i agree.  In the context of "person, number, and
tense", i interpreted "person" as "grammatical person"    (073)

> >> number and tense specified among others to make sense.
> > Similarly number, tense, gender and other linguistic features are
> > allowed or required by different languages "to make sense".  Such
> > features may or may not relate to a feature of the meaning being
> > discussed and might or might not be expressed in the knowledge base
> > using the ontology.    (074)

> "Do you want fries with that?"
> "No."    (075)

> There are no verbs in the second "sentence", no person, number, tense,
> gender or "other linguistic features", but it makes perfect sense to
> customers who drive through a fast food window lane millions of times per
> day.    (076)

I agree.    (077)

-- doug foxvog    (078)

> >> Or in other words "Media is the message" is interpreted as
> >> The message is instruction - in my translation.
> > I find this interpretation curious and don't understand how it
> > relates to your above statements.    (079)

> I think the difference is in acknowledging the subjectivity and inherent
> limits on formalizing language.  John Sowa has explained a lot about how
> language and logic are not fully integratable in the distant past on this
> list.  Now might be a time to bring this information out again, so we can
> look at the Peircean view of logic and language in more depth.
>> Regards, Ferenc    (080)

> ===========================================================
>doug foxvog  doug@xxxxxxxxxx   http://ProgressiveAustin.org
> "I speak as an American to the leaders of my own nation. The great
> initiative in this war is ours. The initiative to stop it must be ours."
>     - Dr. Martin Luther King Jr.
> ===========================================================
> I was at college in Atlanta when King was killed in Tennessee.  We would
> listen to his sermons from Ebenezer Baptist Church whenever they were
> broadcast locally.  He was a very, very subjective person and the world is
> a
> better place for it. He didn't subscribe the logic of his day, and he made
> some amazing changes in society by clearly convincing people that he was
> correct in his moral interpretation of events.    (081)

Thank you for sharing this.    (082)

-- doug    (083)

> HTH,
> -Rich    (084)

doug foxvog    doug@xxxxxxxxxx   http://ProgressiveAustin.org    (085)

"I speak as an American to the leaders of my own nation. The great
initiative in this war is ours. The initiative to stop it must be ours."
    - Dr. Martin Luther King Jr.
=============================================================    (086)

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
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (087)

<Prev in Thread] Current Thread [Next in Thread>