John Sowa wrote:
1. I don't believe that the definitions in philosophy and
computer science differ in any significant way.
3. If possible, we should adopt a common definition that
is acceptable to both fields
The draft document is written as a logical walk down a set of distinctions,
so that we could discuss the source of disagreements and clearly identify
the point of departure. John's objections to the first and most fundamental
distinction (philosophy vs. computer science word senses) makes evident the
reason why certain topics are never "put to rest" by philosophers and other
dialectic sportsmen. To say there is no difference between what a professor
of Aristotelian ontology means by ontology and what a bioinformatics
computer scientist managing a gene database means is absurd.
There is a new word sense for ontology, just as there are new word senses
for other technical terms in computer science: process, client, server, etc.
While my training in philosophy is surely inferior, I would dare say (with
no loss of irony) that John's argument makes an ontological category error.
The Ontologies of philosophy are theories, ideas, ways of thinking about the
world, and arguments about the nature of Reality. The ontologies that are
the subject of W3C standards, manipulated by software, and used to represent
huge stores of data in databases are material, concrete, objective documents
in the same category as programs, database schemata, and other digitally
stored representations.
Another irony is the rhetoric that we should put the "15 year old"
definition in our literature "to rest" and replace examples of computer
based ontologies that we have been collecting for our summit with examples
from Aristotle and Kant.
There is a reason why a lot of people have stopped reading this list. It is
because of this style of long-winded argument by attrition.
Next week we are going to meet to do something constructive: get a clear
understanding of how the family of computer-based representations that we
work with every day -- from formal ontologies to concept hierarchies to
topic maps to taxonomies to folksonomies and a lot of interesting other
cases -- are alike and differ. If we are successful, we will have come up
with a framework and way of talking about ontologies that will allow normal
people to stop glazing over when they hear the term. We have no need or
ambition to come up with a definition of ontology that is "acceptable to
both fields" (philosophy and computer science). It would be a great
achievement to get an understanding, if not a consensus, of our own.
tom gruber
-----Original Message-----
Sent: Thursday, April 19, 2007 10:35 PM
To: Ontology Summit 2007 Forum
Subject: Re: [ontology-summit] Ontology Framework Draft Statement for the
Ontology Summit
Leo,
I agree with Chris W:
Surely after 15 years we can do better than "specification of
a conceptualization"? Isn't it time we put that one to rest?
A lot of hard work has gone into that draft, but I have some
concerns about the definitions at the beginning:
1. I don't believe that the definitions in philosophy and
computer science differ in any significant way.
2. Where there are differences, they are differences in
emphasis or goals.
3. If possible, we should adopt a common definition that
is acceptable to both fields, and include a few comments
about the way that differences in goals and emphasis may
cause differences in usage.
I'll start with the first point:
There are at least two important word senses for 'ontology':
ontology as a field of study "ontology (philosophy)" and
ontology as a technology for computer and information
scientists. We are talking about the second sense of the
word, "ontology (computer science)".
Suggestion: I would delete the two qualifiers "(philosophy)"
and "(computer science)". Then replace that statement with
the following:
There are two important senses of the word 'ontology':
ontology as a general field that studies what exists,
and a particular ontology that is the result or product
of such a study.
Then follow that with examples of such products, such as
Aristotle's ontology of 10 top-level categories, Kant's 12
top-level categories, and various computer versions, such
Cyc, SUMO, etc.
I agree with Chris that the following definition has some
serious problems:
An ontology, for computer and information sciences, is
a specification of a conceptualization...
A definition is supposed to define a poorly understood word
in terms of other words that are simpler, more common, or
easier to understand. But the word 'conceptualization' is
much harder to define than 'ontology'. It is also a less
common term. (Google has 14.5 million hits for 'ontology',
but only 4.3 million for 'conceptualization' -- or 6 million
if you include the spelling 'conceptualisation'.)
If we define "ontology" as "study of existence" and define
"an ontology" as the result of that study, those definitions
depend only on the three words "study", "existence", and
"result", which have, respectively, 492, 179, and 762
million hits on Google. That meets one criterion for a
good definition: define uncommon words in terms of more
common ones.
I have some quibbles about the remainder of the report,
but my primary recommendation is to make a drastic cut in
the opening section: replace everything up to the heading
"kinds of ontologies" with those simple definitions above.
John
_________________________________________________________________
Subscribe/Config:
_________________________________________________________________