[Top] [All Lists]

Re: [ontolog-forum] Foundation ontology, CYC, and Mapping

To: edbark@xxxxxxxx, "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: Mike Bennett <mbennett@xxxxxxxxxxxxxxx>
Date: Fri, 05 Feb 2010 14:53:48 +0000
Message-id: <4B6C30FC.3020404@xxxxxxxxxxxxxxx>
Your example suggests a somewhat trivial illustration of the approach I 
have been trying to take. This approach is to try to use any existing 
ontology or taxonomy that has already been defined by competent 
authorities - I prefer the idea of a competent authority as a setter of 
terms rather than an isolated exercise among ontologists. One competent 
authority relevant to your example is Linnaeus and the taxonomy is the 
Linnaean Taxonomy of Species. I gather this is maintained somewhere. It 
gives us all the axioms we would need (and any changes to axioms) in 
distinguishing one species from another. For example, when dogs were 
reclassified as not being a distinct species but as members of the wolf 
species, it didn't take an ontologist to change the axioms. The 
competent authority updated the taxonomy.    (01)

One thing I have not yet found a satisfyingly authortitative ontology 
for is legal terms, though there is some interesting work in Holland 
which I intend to follow up. In the example you give here, members of 
one species, homo sapiens, have additional legal facts about them. These 
are nothing to do with Linnaeus but are additional axioms which would 
ideally come from another ontology. I think that regardless of local 
jurisdictional differences, it is pretty well established that certain 
members of the species homo sapiens have an additional, legal fact about 
them, that they are able to be a citizen of a country, with certain 
legal rights (the precise age of majority may vary from one jurisdiciton 
to the next). This was obviously only universally applicable since the 
abolition of slavery (finally outlawed in Saudi Arabia in 1954). The 
same set of legal foundational terms would include additional 
definitions of "Legal persons" as defined in different jurisdictions, of 
which there is a great deal of commonality e.g. a company incorporated 
by the issue of shares is a pretty universal kind of legal person. So if 
and when there is a single, semantic standard for basic legal terms, 
these concepts would be present. It would make sense to use these and 
not to either replicate locally (as we have to at present) or to look to 
some overarching FO project.    (02)

So in your example there are simple axioms that can be defined which 
would address the question about whether an elephant can be a citizen of 
India (cows may pose a more interesting question...). These facts would 
be found in a legal ontology and not (as your example seems to suggest) 
through some observation of the differences between a human and an 
elephant or an ape - those are not legal differences so would not be in 
the legal ontology.    (03)

The point being that it should be possible, with a little effort and 
research, to source the relevant axioms all and only from established 
and authoritative sources of facts.    (04)

I've been meaning to chip in to this conversation for a while, but I 
would like to suggest that it would be more productive to identify 
formal sources of semantics that are widely recognised and use these as 
the basis for foundational ontology material. In most cases these will 
be simpler than dictionary terms as Pat C is suggesting, and more useful 
for computer interoperation since many are developed for applications in 
which we use computers (for example XBRL for financial reporting). No 
doubt not all of those standards will have mutually comprehensible 
terms, and many will contain terms which are reducable to something more 
primitive but are not (check out the UN FAO ontology for examples), but 
what they would have is some provenance (and maintenance) of meaning. 
Many industry standards still live in the era of data model or XML or 
other messaging but would benefit from something reverse engineered into 
a formal logical notation and perhaps we can help them with that. Many 
are used in sharing of information among computers about things that 
people care about and have common legal grounds (e.g. commerce, 
financial, insurance) , so we should be unsurprised if the existing 
interoperability of many data standards would be reflected by a useful 
commonality in semantics, including identifable and useful semantic 
"primitives" or simple, extendable terms such as legal person, human, 
goods, services and so on.    (05)

Mike    (06)

Ed Barkmeyer wrote:
> Pavithra wrote:
>> Dr. Sowa,
>> -   An Elephant is an animal
>> -  Clyde is an elephant
>> -  Therefore Clyde is an animal
> Fine.  Now let us use our very limited vocabulary in the following way:
> A citizen of a country is a person born in that country.
> A person is an animal.
> An elephant is an animal.
> Clyde is an elephant.
> Clyde was born in India.
> Is Clyde a citizen of India?
> Maybe.  We can't deny the proposition.
> The problem is that we also need a vocabulary that provides the terms to 
> distinguish "person" from "elephant", and the definition of "person" has 
> to include those "distinguishing characteristics".  A person is an 
> animal with some specific properties that distinguish "person" from 
> "elephant" and, more problematically, from "ape" (or not).  Experience 
> teaches that it takes an enormous vocabulary to explicitly make all the 
> distinctions people's brains have learned to make.  It is in making all 
> the necessary distinctions that the 2000-word vocabulary breaks down.
> The alternative of course is that you only need an axiom:  No person is 
> an elephant.  But then you need a lot of axioms just to sort out 
> persons, elephants, tigers and mongoose.  And the volume doubles when 
> you move to Australia.
> -Ed
>       (07)

Mike Bennett
Hypercube Ltd. 
89 Worship Street
London EC2A 2BF
Tel: +44 (0) 20 7917 9522
Mob: +44 (0) 7721 420 730
Registered in England and Wales No. 2461068    (08)

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
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (09)

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