[Top] [All Lists]

Re: [ontolog-forum] Watchout Watson: Here comes Amazon Machine Learning

To: ontolog-forum@xxxxxxxxxxxxxxxx
From: John F Sowa <sowa@xxxxxxxxxxx>
Date: Fri, 01 May 2015 13:32:20 -0400
Message-id: <5543B8A4.8050303@xxxxxxxxxxx>
Ed, Adrian, Rich, and Leo,    (01)

My note was too brief.  I certainly did not want to confuse
temporal modality with alethic modality.  When I said that
'possible' and 'necessary' "can be treated in the same way"
as 'sometimes' and 'always', I was talking about the formalism.    (02)

> The handling of future things is very much about what the ontological
> commitments are.  One "can" treat future things as 'possibilities',
> by making that ontological commitment, but one can also treat them
> as 'facts' by making a different commitment...    (03)

I strongly agree with that point and the "..."    (04)

> In short, this is all about your ontology, and only some ontological
> choices affect the choice of logic.    (05)

I not only agree with that point, I thank you for a very nice
intro to my paper on "Worlds, models, and descriptions":    (06)

  1. Saul Kripke developed the term 'possible world' (from Leibniz)
     into a widely used version of semantics for all the modalities.
     Those worlds may be possible alternatives (alethic modality),
     times (temporal), permissible (deontic), beliefs (epistemic).    (07)

  2. Michael Dunn developed an equivalent semantics based on laws
     and facts:  Each world is characterized by all the facts that
     are true in that world.  Depending on which modality you're
     working with, a law is a fact that is considered necessary
     (alethic), always true (temporal), obligatory (deontic), or
     known (epistemic).  (In deontic logic, an obligatory statement
     might not be true -- because people are often sinners.)    (08)

  3. For knowledge bases and databases, Dunn's version is convenient.
     An ontology may be considered the laws for a family of worlds,
     and a database may be considered the contingent facts that
     describe a particular world (or aspect or time slice).    (09)

For details, see http://www.jfsowa.com/pubs/worlds.pdf    (010)

> My engineering (and semiotics) background objects to your example.
> The design for an aircraft is a design, not an aircraft.  The
> design exists independently of its realization...    (011)

Again, I strongly agree with that point, with your "...", and
with Adrian's example and citation:    (012)

> "Ceci n'est pas une pipe."
> http://en.wikipedia.org/wiki/The_Treachery_of_Images    (013)

> The problem in many engineering disciplines is that the design
> engineers *only* work with designs and prototypes, and they use
> the *terms* for the actual things in describing their design objects.    (014)

I certainly agree with the importance of distinguishing the many
levels of models, metalanguage, and metonymy.  Even logicians are
often guilty of confusing a Tarski-style model of the world (or
some aspect of it) with the actual world we're living in.    (015)

And, by the way, I'm glad that you mentioned semiotics -- because
too many ontologists ignore the fundamental distinctions.  The
first of Peirce's semiotic triads is Mark/Token/Type:    (016)

  1. A mark is anything observable *before* any classification.    (017)

  2. A token is a mark that has been classified as an instance of
     some type.    (018)

  3. A type is specified by a predicate or description that is used
     to classify marks as tokens.    (019)

Yet many linguists and ontologists who adopt Peirce's Type/Token
distinction ignore the marks.  But the same mark may be interpreted
as a token of an open-ended variety of types:  smudge, scratch,
carving, image, glyph, or symbol of a word, syllable, phoneme...    (020)

> the words actually used by people in each sub organization differ;
> they have different perspectives.    (021)

Yes.  At a low level, they may classify the same marks as tokens
of different types that are represented by different words.  When
a patient complains about a "pain" or a "rash", a physician may
have a very different classification.  But even when they use the
same words, a patient, nurse, physician, surgeon, pharmacist...
have different kinds of knowledge and ways of using it.    (022)

> You might look at this recent FOIS 2014 paper,
> And the emerging Information Artifact Ontology by the U. Buffalo et al folks:
> https://code.google.com/p/information-artifact-ontology/.    (023)

That's an important application, but 'information' is an extremely
vague term.  Claude Shannon admitted that the word 'information'
as the name of his most famous theory was confusing.  Adding the
word 'artifact' introduces even more confusion.    (024)

For over 20 years, I've had an ongoing argument with Barry Smith
about the role of semiotics in ontology.  See, for example,
http://www.jfsowa.com/pubs/signproc.htm#static .    (025)

Excerpt from that article:
> [In] Der Logische Aufbau der Welt by Rudolf Carnap... psychological
> “objects” are physical objects that have a spatiotemporal location
> within an individual human object. Carnap “clarified” the notion of
> intention relation by saying it is “nothing but” a relation between
> a psychological object and some other object. Carnap recognized the
> importance of the sign relation and admitted “The construction of
> this relation is more difficult than any of the other relations
> which we have hitherto undertaken.”    (026)

I would go one step further than Carnap:  Attempting to define
"the sign relation" in terms of "elementary experiences" is not
only difficult, but *impossible*.  The reason is simple:  every
perception (elementary experience) is itself based on signs.    (027)

> My other pet peeve is the folk who insist that the design is a
> classifier (predicate, class) of the realizations (in the face of
> the known as-designed vs. as-built distinction in industry practice).    (028)

Again, I very strongly agree.  That's one more example where the
vague term 'information artifact' obscures the semiotic distinctions.    (029)

In terms of the worlds.pdf article, each of the possible worlds can
represent a different stage in the design, redesign, implementation,
maintenance, and use of the system (which may be 40 years or more).
At each stage, the laws (current specifications) evolve -- sometimes
by a drastic overhaul.    (030)

John    (031)

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    (032)

<Prev in Thread] Current Thread [Next in Thread>
  • Re: [ontolog-forum] Watchout Watson: Here comes Amazon Machine Learning - ZDNet- 2015.04.10, John F Sowa <=