[Top] [All Lists]

Re: [ontolog-forum] Thing and Class

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Thu, 28 Aug 2008 09:34:19 -0400
Message-id: <48B6A95B.2030905@xxxxxxxxxxx>
Azamat and Antoinette,    (01)

Developing a good meta-level ontology for talking about ontology
and its relationship to various applications is important, and
poor choices can lead to endless confusion.  (Even initials can
lead to confusion when they're ambiguous, e.g. AA in this case.)    (02)

Azamat> The interrelations of classes as well as classes and things
 > are actually more subtle and deep, than generally presented in
 > various specifications...    (03)

I certainly agree.    (04)

Antoinette> I consider, that in this common world, people are NOT
 > things. People place or thing...    (05)

There is a time-honored international terminology for logic and
ontology that was derived from Greek and Latin.  Some people have
objected because the terms are often long and unfamiliar.    (06)

One such word is 'entity' from Latin 'entitas', which literally
means anything that exists.  It does not have any associated
baggage of familiar associations, and it can be associated with
cognate terms in many other languages.  If some languages don't
have a native word for 'entity', they can just borrow 'entitas'.    (07)

The word 'thing', however, has too many familiar associations,
and the corresponding familiar terms in other languages have
different associations.  Therefore, it is very hard to translate
the word 'thing' to rough equivalents in multiple languages
without creating different confusions in each language.    (08)

The word 'class' is another term that creates multiple confusions
because it also has multiple and confusingly different meanings
in mathematics, programming languages, and common English usage:    (09)

  1. In mathematics, the word 'class' is sometimes used as a synonym
     for set by some authors, and it is sometimes used as a term
     for a set-like collection that is too big to be a proper set.    (010)

  2. In object-oriented programming languages, the word 'type' was
     commonly used for data types.  But the OO languages introduced
     a kind of entity that was different from a traditional datatype
     because it had associated procedural "methods".    (011)

  3. In common English usage, the word class is used confusingly in
     ways that are synonyms for 'set' and in ways that are synonyms
     for 'type'.  For example, one could talk about the class of
     students in a room or the set of students.  But there are other
     uses that refer to the type rather than any particular set,
     as in 'middle class', 'upper class', or 'first class'.    (012)

Some time ago, we had had a discussion on this list about whether
we should use the term 'type' or 'class' for the categories of an
ontological hierarchy.  Both Barry Smith and I were strongly urging
people to use the word 'type' rather than 'class', but many others
wanted to use 'class' because it was, unfortunately, used in OWL.
There was a vote, and the word 'type' won.  But there was a lot
of grumbling by people who were using OWL.    (013)

I believe that the OWL developers made a very serious mistake in
adopting the term 'class' because of its association with OO languages.
That should have been a strong argument against using the word 'class'
because ontological categories are very different from OO classes.    (014)

But the main reason for not using the word 'class' is its association
with the purely extensional set theory:  a set (or class) is uniquely
defined by its instances.  A type, however, is an intensional term,
and two types may be distinct even when they have exactly the same
instances (or no instances at all).    (015)

For example, the empty set is a subset of every other set.  The set
of unicorns is empty; therefore, it is a subset of the set of cows.
But the type Unicorn is very different from the type Cow.  Other
examples, include the set of all human beings and the set of all
featherless bipeds.  Today, those sets are the same, but the types
have very different definitions.    (016)

Many dinosaurs, such as T. Rex, were bipeds.  Some of them had
feathers, but no one knows whether they all had feathers.  But
if some of them didn't, it would be a mistake to call them
human beings.    (017)

So I recommend that we drop the words 'thing' and 'class' when
talking about ontologies, and use the terms 'entity' and 'type'.    (018)

John Sowa    (019)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Subscribe/Config: 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 Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (020)

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