Pat --
Nice emphasis on "getting the job done"
You wrote...
John, or anyone else, please provide what you would consider an
adequate definition of some technical term, in English but structured
so that it also captures what you would want as a logical definition in
an ontology.
Just to throw a different approach into the mix, here's an executable English definition of a technical term "is a part of the Continuant class" from [1].
for all c, t, if some-C c t then there is some c1 such that some-C1 c1 t and c part_of c1 at t
----------------------------------------------------------------------------------------------
that-C is a part of the Continuant class that-C1
(A c,t) [ some-C c t => (E c1) [ some-C1 c1 t and c part_of c1 at t ] ]
----------------------------------------------------------------------------------------------
for all c, t, if that-C c t then there is some c1 such that that-C1 c1 t and c part_of c1 at t
some-C and some-C1 are two different Non-process classes with instances
not : (E c,t) [ that-C c t and not (E c1) [ that-C1 c1 t and c part_of c1 at t ] ]
-----------------------------------------------------------------------------------
(A c,t) [ that-C c t => (E c1) [ that-C1 c1 t and c part_of c1 at t ] ]
some-C and some-C1 are two different Non-process classes with instances
some-c is an instance_of that-C at some-t
not : (E c1) [ that-C1 c1 that-t and that-c part_of c1 at that-t ]
----------------------------------------------------------------------------
(E c,t) [ that-C c t and not (E c1) [ that-C1 c1 t and c part_of c1 at t ] ]
some-c1 is an instance_of some-C1 at some-t
some-c is a part_of that-c1 at that-t
------------------------------------------------------------
(E c1) [ that-C1 c1 that-t and that-c part_of c1 at that-t ]
this-item is an instance_of this-Class at this-t
========================================================
c1 C1 1
c C 1
c1 C1 2
c C 2
this-item1 is a part_of this-item2 at this-t
=============================================
c c1 1
c c1 2
the class this-class is of type this-type
==========================================
C Non-process
C1 Non-process
the class some-C is of type Non-process and has at least one instance
the class some-C1 is of type Non-process and has at least one instance
that-C and that-C1 are different
-----------------------------------------------------------------------
that-C and that-C1 are two different Non-process classes with instances
the class some-C is of type Non-process
some-c is an instance_of that-C at some-t
--------------------------------------------------------------------
the class that-C is of type Non-process and has at least one instance
some-t1 is less than some-t2
-------------------------------
that-t1 is earlier than that-t2
some-C1 is not equal some-C2
-------------------------------
that-C1 and that-C2 are different
This is example is online [2], and you can view, run and change it using the underlying system [3].
Some things that you may consider noteworthy are:
* the definition comes with a test example, in the same spirit as examples in dictionaries
* the definition is built on two kinds of primitives:
(1) internal primitives defined by example, as in the sentences above lines like =====
(2) external primitives such as "equal" and "less than"
* one does not have to rely on a NL-to-logic compiler capturing
the correct intention. It's required that the person writing the
definition should also write the mapping to logic.
A different approach, for sure. But maybe there's a thesis-antithesis-synthesis opportunity lurking here?
Cheers, -- Adrian
[1] ''Relations in biomedical ontologies'' by Barry Smith et al,
Genome Biology 2005. http://genomebiology.com/2005/6/5/R46
[2] www.reengineeringllc.com/demo_agents/RelBioOntDefn3.agent
[3] Internet Business Logic (R)
A Wiki for Executable Open Vocabulary English
Online at www.reengineeringllc.com Shared use is free
Adrian Walker
Reengineering
On 5/4/07, Cassidy, Patrick J. <pcassidy@xxxxxxxxx> wrote:
I completely agree with John Sowa that the existing definitions in the
Longman's, using the current defining vocabulary, are in many case not
adequate to capture all of what we would think if as the "necessary
and/or sufficient conditions" for membership in the categories they are
defining. That's because they were designed to be understandable (at
some level) for people whose native language isn't English. They
nevertheless demonstrate to my satisfaction that a relatively small set
of concepts can be used to define a much larger set of concepts, within
some criterion of satisfaction and purpose.
What yet needs to be proven, as I have mentioned before, by experiment
(i.e. just doing it) is that the same principle can be applied to the
more demanding criteria of logical specification of meanings sufficient
for automated inferencing to achieve practical goals. That is the
purpose of formulating the COSMO as a conceptual defining vocabulary.
That will take some time - I estimate a minimum of several
person-years, and if anyone knows of a funding agency willing to
support such work, perhaps it can be done by several people more
quickly.
On the linguistic side, it is not idle to wonder whether we can also
create definitions with recursive reference to the base defining
vocabulary, that we can agree *do* satisfy our intuitions about the
necessary conditions, as we might express them logically in an
ontology. This is susceptible to experiment, which I now suggest.
John, or anyone else, please provide what you would consider an
adequate definition of some technical term, in English but structured
so that it also captures what you would want as a logical definition in
an ontology. We can then work through that to see what is needed to
ground that definition in the Longman's base defining vocabulary, and
see if it will be necessary to define additional primitives. Feel free
to distinguish word senses for ambiguous words.
Shall we?
Pat
Patrick Cassidy
CNTR-MITRE
260 Industrial Way West
Eatontown NJ 07724
Eatontown: 732-578-6340
Cell: 908-565-4053
pcassidy@xxxxxxxxx
> -----Original Message-----
> From: John F. Sowa [mailto:
sowa@xxxxxxxxxxx]
> Sent: Friday, May 04, 2007 8:17 AM
> To: [ontolog-forum]
> Cc: Cassidy, Patrick J.
> Subject: Re: [ontolog-forum] A "common basis"
>
> Pat and Patrick,
>
> I agree that the hypothesis is either trivial or false.
>
> PC> What specifically isn't feasible? Specifically, what
> > cannot be done?
>
> PH> Locate a set of basic concepts sufficient that all
> > others can be derived from the base set by writing
> > definitions.
>
> To test the hypothesis, I checked the _Longman Dictionary of
> Contemporary English_, which uses a small defining vocabulary.
> And I looked for the term 'quantum mechanics'.
>
> That term was not in the dictionary, but it did have two
> related terms:
>
> quantum. _tech_ (esp. in PHYSICS) a fixed amount.
>
> quantum theory. the idea that ENERGY (3) travels in
> fixed amounts (QUANTA).
>
> Some points to note:
>
> 1. The word 'tech' in italics is a metalevel term that
> indicates the word 'quantum' is technical.
>
> 2. The capitalized words PHYSICS, ENERGY, AND QUANTA
> are not in the defining list.
>
> 3. The parenthesized 3 indicates that the third
> definition of 'energy' is being used.
>
> Following is definition 3 of 'energy':
>
> the power which does work and drives machines:
> _atomic / electrical energy / the energy of the sun_
>
> These definitions may be useful for human beings who
> have no knowledge of the subject (say, quantum mechanics)
> and want to get a vague notion of what the term means.
>
> They might also be useful for an analogy engine that
> searches for patterns of associations among a family
> of terms. See, for example,
>
> http://www.jfsowa.com/pubs/analog.htm
> Analogical Reasoning
>
> But if you (or a computer program) need to do any kind
> of serious reasoning about these terms, you have to study
> physics. Any associations that a human or a computer
> picks up from such a dictionary are going to be *wrong*.
> They will have to be corrected by a method that is far
> more complex than writing one-line definitions in terms
> of a small vocabulary.
>
> In physics, for example, the words 'power' and 'energy'
> have clearly distinct definitions, and it is false to
> say that energy is power.
>
> I consider these examples to be sufficient *proof* that
> the idea of using a small defining vocabulary is useless
> for defining anything beyond vague associations. Anybody
> who needs further convincing should check other technical
> terms in Longman's.
>
> John
>
>
_________________________________________________________________
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
_________________________________________________________________
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 (01)
|