To answer the comments of John S and Matthew W: (01)
First, I need to mention (again) that the goal of the work on a
primitives-based ontology is twofold:
To provide a basis for accurate semantic interoperability among
databases and applications
To provide a knowledge representation that can at least in part support
computerized human-level language understanding (better than alternatives) (02)
These are very practical *engineering* goals, which are unlikely to
benefit much from broad general theorization. More specifically: (03)
The objections of John and Mathew seem to be based on the misunderstanding
that I am claiming the existence of some finite set of primitives that will
precisely define all words that anyone would ever want to use or invent. Of
course that is improbable in the extreme. This misunderstanding perfectly
illustrates one of the reasons that the NL goal appears to me not to suffer
fatally from the problems described in Kilgariff's paper (which I read more
than once) - that is, if one wants to reproduce human-level language
understanding one has to remember that people don't understand each other
perfectly - even those who are well acquainted with their native language,
and particularly when emotions are engaged and debating points are being
made. The computer doesn't have to be perfect, just as good as people. (04)
To illustrate, consider one point from John's reference slide-set:
[JS] (goal3.pdf):
>> No finite set of words can have a fixed, precise set of mappings
>> to a dynamically changing world (05)
Duh. Of course not, but a finite set of well-specified ontology elements
**can** have a "fixed, precise set of mappings" to a finite set of
databases, the meanings of whose elements are determined by the operations
and goals of their applications. For natural language, we expect
inaccuracies, even among people using language. Any idiot can say things
that no genius could understand. (06)
Of course, people stretch meanings of words in new situations, but unless
all communicating parties are aware of the circumstances, that is what will
often lead to misunderstanding. We have to explain new uses to other
people, as well as to our computers, by providing clarifying definitions. (07)
[MW]
>2. A concept is primitive unless it is the intersection of 2 or more other
>concepts.
>
This is formal-logic definition of 'primitive'. For the COSMO, most
concepts are specified by necessary conditions (not necessary and
sufficient), and I am not overly concerned to identify those that are truly
"primitive" from those that might actually be constructed from others in the
ontology. I am concerned that, at the first iteration, I will have a basic
set of concepts sufficient to specify the meanings of domain concepts used
in multiple different applications, to a level sufficient to support those
applications. (08)
This is a very pragmatic engineering goal. In an engineering task the aim
is to devise an artifact that will accomplish some function. But such
artifacts can rarely if ever be proven by theoretical arguments to have the
proper attributes. Reality is complex, and when artifacts are actually
used, unanticipated problems reveal inadequacies (remember the Obamacare
website? Hardly a rare case of bugs in an artifact). Proof of concept
for an engineering task can only be accomplished by building the artifact
and testing it. That is what the COSMO project is intended to do - test the
notion of an primitives-based ontology to support (initially) semantic
interoperability among databases - and eventually language understanding. (09)
If anyone has a (hopefully simple) test case for database interoperability,
that would be a useful contribution to the discussion. Theorizing about the
flexibility of human use of words does not help - the issue is well known
and does not answer the practical questions involved in this project. The
initial base of 'defining' concepts in the ontology will be supplemented as
required for new applications. The issues involved in supplementation have
also been thoroughly considered. How much supplementation will be necessary
as time goes on can only be determined in practice. I will be very
disappointed if, after a hundred applications have been mapped with the
ontology, *every* new application still requires some new primitives to be
added. At that point, John may be entitled to gloat. (010)
Pat (011)
Patrick Cassidy
MICRA Inc.
cassidy@xxxxxxxxx
1-908-561-3416 (012)
>-----Original Message-----
>From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx [mailto:ontolog-forum-
>bounces@xxxxxxxxxxxxxxxx] On Behalf Of Matthew West
>Sent: Saturday, May 03, 2014 4:23 PM
>To: '[ontolog-forum] '
>Subject: Re: [ontolog-forum] English number of words/concepts that cannot
>be composed of others
>
>Dear All,
>Whilst agreeing with John, I'd like to approach this from the other end of
the
>telescope. First some basics:
>1. For each concept there is a set of objects that it describes.
>2. A concept is primitive unless it is the intersection of 2 or more other
>concepts.
>
>The proposition is that there is some moderate set of primitive concepts
such
>that there is no useful concept that cannot be defined as the intersection
of
>that set of primitive concepts.
>
>Now consider the number of objects there are that we might want to
>describe.
>Consider our universe, the galaxies, planetary systems, stars, planets,
>moons, other bodies, parts of these, life, molecules, atoms, subatomic
>particles. Then there are all the other possible universes, with unicorns
and
>so on.
>
>So you need to prove that there is a number of primitive concepts, n,
where
>n is (a lot) less than infinity, such that it is not possible to come up
with some
>useful set of all these objects that is not the intersection of some of
those n
>concepts.
>
>It does not seem credible to me that there is any such number, but I look
>forward to seeing any proof that such a number exists.
>
>Regards
>
>Matthew West
>Information Junction
>Mobile: +44 750 3385279
>Skype: dr.matthew.west
>matthew.west@xxxxxxxxxxxxxxxxxxxxxxxxx
>http://www.informationjunction.co.uk/
>https://www.matthew-west.org.uk/
>This email originates from Information Junction Ltd. Registered in England
>and Wales No. 6632177.
>Registered office: 8 Ennismore Close, Letchworth Garden City,
Hertfordshire,
>SG6 2SU.
>
>
>
>-----Original Message-----
>From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
>[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of John F
>Sowa
>Sent: 03 May 2014 19:54
>To: ontolog-forum@xxxxxxxxxxxxxxxx
>Subject: Re: [ontolog-forum] English number of words/concepts that cannot
>be composed of others
>
>Gregg, Tom, Pat C, and John B,
>
>GR
>> Have you looked at Natural Semantic Metalanguage?
>> http://en.wikipedia.org/wiki/Natural_semantic_metalanguage
>
>Yes. I cited Anna Wierzbicka's _Lingua Mentalis_ in my 1984 book, and
I've
>followed her other books over the years. I call her primitives 'accordion
>words' -- because you can stretch them and squish them to fit anything you
>please.
>
>They're useful. To quote my favorite philosopher, C. S. Peirce:
>> It is easy to speak with precision upon a general theme. Only, one
>> must commonly surrender all ambition to be certain. It is equally
>> easy to be certain. One has only to be sufficiently vague. It is not
>> so difficult to be pretty precise and fairly certain at once about a
>> very narrow subject. (CP 4.237)
>
>Again, I recommend that every reader of this list *study* the paper "I
don't
>believe in word senses." Adam K and Sue A are *professionals* in
>lexicography and computational linguistics. They know the difference
>between accordion words and precise definitions. Both can be useful for
>different purposes, but it's important to know the difference.
>
>TK
>> I am looking for (I'm going to call it) 'fundamental concepts' and I
>> am making the assumption that there is some basic agreed level of
>> definition of these concepts so we don't end up in Physics and
Chemistry.
>
>Brief answer:
>
> 1. There is no "basic agreed level" whatsoever -- NONE!
>
> 2. The top level of an ontology *must* be vague and underspecified.
> It can be useful, but the real knowledge is in the lower levels.
>
> 3. Please remember that Cyc started out with the assumption that a
> formal ontology of the knowledge of a high-school graduate could be
> specified in 10 years. After 30 years and over $100 million of
> investment, Doug Lenat has emphasized that all the real knowledge
> is in the detailed low levels. The top level is very vague and
> underspecified. It cannot support any kind of detailed reasoning.
>
>TK
>> My criteria for 'fundamental concept' is that it cannot be replaced by
>> a semantic net-let that crosses the agreed level.
>
>If that's your definition, then you're talking about the empty set.
>There is no concept or thought of any kind that cannot be analyzed at a
>deeper level.
>
>> So John S, to take your examples...
>
>I was just trying to give one-line examples. In any case, the terms in
your
>analyses are accordion words. Please study that paper by Adam K.
>
>JB
>> as Lakoff shows us in "Women, Fire, and Dangerous Things" the
>> universals are different for different linguistic environments...
>> But it still comes down to what type of tasks are facing. The "core"
>> concepts for farming are very different from those needed in the office.
>
>Yes! But I would avoid using the word 'core' because it gives the
mistaken
>impression that some kind of core is possible. But even for farming and
>offices, the basic terms are accordion words. Note how we use the
>abbreviation 'cc' in our emails. In office-speak, it used to mean 'carbon
copy'.
>When was the last time you saw a carbon copy?
>
>PC
>> according to Guo, the number of senses used **in the definitions**
>> average to less than 2.
>
>If so, Guo doesn't know how to define words or to count definitions.
>I suspect he was using those terms as accordion words. If you stretch and
>squeeze them enough, you can adapt them to almost anything.
>
>But with every stretch and squeeze, you blur an immense amount of info.
>Please tell Guo to study Adam K's paper. Also study the publications
about
>*microsenses* by Alan Cruse. A microsense is any intermediate point as
you
>stretch and squeeze your accordion.
>
>PC
>> If anyone knows of such a study, I would very much like to get a
pointer.
>
>I've given you many, many pointers over the years. And I beg you to study
>them until you reach enlightenment. For starters, please reread
>http://www.jfsowa.com/talks/goal3.pdf and *follow* every URL to every
>reference in it.
>
>Those other goalX.pdf files are also surveys. You have to dig into the
>references until you get the point. Anything that looks like or smells
like a
>primitive is probably an accordion word.
>
>John
>
>__________________________________________________________
>_______
>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
>
>
>
>__________________________________________________________
>_______
>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
> (013)
_________________________________________________________________
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 (014)
|