[Top] [All Lists]

[ontolog-forum] The KR ontology -- response to some criticisms

To: "cg@xxxxxxxxxxxxx" <cg@xxxxxxxxxxxxx>, [ontolog-forum] <ontolog-forum@xxxxxxxxxxxxxxxx>
From: John F Sowa <sowa@xxxxxxxxxxx>
Date: Tue, 8 Sep 2015 11:32:38 -0400
Message-id: <55EEFF96.60101@xxxxxxxxxxx>
I received an offline note that asked about some criticisms
of the ontology in my book _Knowledge Representation_.  See
http://mba.eci.ufmg.br/downloads/pos/TopLevelOnto-Jansen.pdf    (01)

I'd like to emphasize that the diamond, which I presented in the
KR book (copyright 2000, finished in 1999) was an *example* of how
the product of multiple distinctions can generate a lattice of
categories.  I won't claim that it's the final word, but I'd say
that it's better than most.    (02)

For an analysis of some of the issues, see the the article on Signs,
Processes, and Language Games, which I wrote a few years after the
KR book:  http://www.jfsowa.com/pubs/signproc.pdf    (03)

Among other things, Jansen criticized the way I mixed the distinction
between continuant vs. occurrent with the other two.  In Figure 3,
p. 21 of signproc.pdf, I did not include that distinction.  I still
believe that the distinction can be useful, but it's not fundamental.    (04)

Every continuant is a slowly changing process.  In one context, a
hurricane may be called a process, but it may last for a week or two
and be given a name.  As an extreme example, the Red Spot in Jupiter's
atmosphere is a storm that has been observed for over 300 years.
Whether you call storms (or anything else) a continuant (object) or
an occurrent (process) depends on purpose, context, time scale...    (05)

On p. 22 of Jansen's article, he said
> there are no abstract occurrents... what occurs is never abstract.    (06)

In the KR book, I said that the abstractions describe physical
phenomena.  The categories Script, History, and Purpose describe
physical occurrents.  In my updated ontology, I replace 'abstract'
with 'sign'.  That makes it clear that the signs that describe
an occurrent need not change over time.  Note the most basic of
Peirce's semiotic triads:  mark, token, type:    (07)

  1. A _mark_ is anything perceptible by any sensory organ, possibly
     with the aid of instruments such as microscopes, telescopes, etc.    (08)

  2. A _token_ is an interpretation of a mark as an instance of a type.    (09)

  3. A _type_ is a monadic predicate (AKA universal) that is used to
     classify marks as tokens.    (010)

For occurrents, the sign types Script, History, and Purpose do not
themselves change.  But they describe changes in marks, which they
characterize as tokens of some type.    (011)

Jensen criticizes the KR ontology because it does not mention abstract
particulars.  There are two answers to that claim:    (012)

  1. If you use Common Logic or any other logic that supports quantifiers
     over functions and relations, the distinction between universals
     (relations) and particulars (instances) vanishes.  That distinction
     may be traditional, but it creates more confusion than it clarifies.    (013)

  2. Signs may describe other signs.  Abstract particulars are tokens
     of sign types that describe other kinds of signs.  Note that tokens
     are always physical marks.  Every commentary about abstractions
     refers to some marks (sounds, notations, or diagrams) that
     represent the abstractions.  (If you want a formalism that avoids
     any talk about marks, use Common Logic.)    (014)

Jansen also complains that the KR ontology does not include time and
space as categories.  But note the excerpts from the KR book in the
following web page:  http://www.jfsowa.com/ontology/causal.htm    (015)

In that analysis, *processes* are fundamental.  Time and space are
based on relationships among processes.  For example, see Figure 7,
which shows a Petri net (a kind of script) that describes a process
called a *clock*.  That process may be adopted as a standard for
defining a metric for timing other processes in terms of *ticks*.    (016)

This analysis is consistent with Einstein's Gedanken experiments
about time and space.  If we want to specify a metric over space,
we use a ruler (or other slowly changing continuant) as a standard.
But as Einstein noted, even if our ruler is very rigid, it may still
give different readings in different contexts.    (017)

Jansen repeats Kant's criticisms of Aristotle, but Franz Brentano
replied with a strong defense of Aristotle's categories.  For
Brentano's tree of A's categories, see slide 18 of
http://www.jfsowa.com/talks/patolog1.pdf    (018)

I don't believe that Brentano's analysis can be dismissed by just
repeating Kant's brief comment.  There's solid analysis behind it.
In German, https://archive.org/details/vondermannigfac00brengoog    (019)

And by the way, patolog1.pdf is the first lecture of a 5-day short
course on "Patterns of Logic and Ontology", which I taught in 2013.
Slide 2 of patolog1.pdf points to the other four lectures.    (020)

In slide 26 of patolog4.pdf, I show the top level of the ontology that
John Wilkins specified in 1668.  Jonathan Swift and Jorge Luis Borges
parodied that ontology (and related work by such notables as Descartes,
Leibniz, etc.).  But it is more defensible.  (The references on slide
25 point to the bibliography, http://www.jfsowa.com/bib.htm )    (021)

On slide 30 of patolog4.pdf, I show an updated version of the top level
of Wilkins' ontology.  The most important change is to rename his
category Transcendentals as Signs.  I renamed the category Creator
as Laws, and added "Theists can think of the laws as the Logos, which
John the Evangelist said is God."    (022)

In the physical branch, the category Laws has the counterpart Types
in the sign branch.  Observables correspond to Tokens.  sign types
are specified by humanly-defined laws, which approximate the Logos.    (023)

A computer is a semiotic processor, and the category of Signs is
essential for IT.  See http://www.jfsowa.com/pubs/rolelog.pdf    (024)

For related issues about the relationships among language, logic,
images, and mental models, see the tutorial on "Natural Logic":
http://www.jfsowa.com/talks/natlog.pdf    (025)

In any case, I'm writing another book, which will say more about
all these issues.  (Please don't ask me when it will be done.)    (026)

John    (027)

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    (028)

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