Re: [ontolog-forum] Using controlled natural languages for ontology

[Ali Hashemi <ali@xxxxxxxxx>]

John,

Your slides are useful and do touch on many of the same points, I value them as a resource. And yes, of course a KE ought have working knowledge of the field with a strong dose of humility and awareness of their limitations in a complete/expert understanding of the domain - I believe Ed has also stressed this point.

> There is a large untapped, but potentially more useful (and

> at worst complementary) avenue of capturing SME knowledge via
> examples (i.e. Tarski models rendered in SME readable form).

The most critical part of the ontology is deciding on an
appropriate choice of objects, functions, and relations.

The language required to describe a Tarski-style model would
use exactly the same vocabulary for those items as the language
for describing the logical combinations.

The only difference is in the use of Boolean operators and
quantifiers.  The issue we're debating is whether we should
replace 'and', 'or', 'not', 'if', 'every' and 'some' with
special symbols.  To answer that, please look at Naproche.

While I appreciate what you are saying here, I believe it is missing the point.

What I was trying to bring attention to is that a lot of work is being focused on getting the SME to express their intuitions directly in the logical language (whatever its syntax).

Sometimes (nay, often?), the intuitions that might be necessary for the axiomatization of a particular domain are not easily accessible - whatever the syntax - to the SME.

Take for instance the notion of orthocomplementation or pseudocomplementation, an integral part of many a mereotopology (i.e. http://stl.mie.utoronto.ca/publications/RT-journal.pdf ). Yet expressing either concept, is cumbersome to a lay person, or even someone who has "reading capability" of the logic. Moreover, figuring out that these two notions are important for mereotopology might not even be obvious without a deep understanding and access to a large background of mathematical knowledge!

However, one can, by analyzing the set of models in the extension of the theory (which is after all, is what the SME is working with on a daily basis - the models that satisfy the intensional axioms), we can determine that these models satisfy orthocomplementation or pseudocomplementation.

What I am getting at is that there are many logical constructs or axioms which are integral in the axiomatization / formalization process, which just aren't necessarily intuitive to the SME, no matter how we render the syntax. In contrast, these SME's are often quite intimately familiar with the models that satisfy their desired axioms. Developing a translation from these models to a complete diagram for various fields would go a long way in capturing that elusive knowledge in the SME's head in a machine readable way.

I absolutely agree that the vocabulary of the functions, relations and objects must be aligned; however their application in any representation of the model need not obviously reflect the underlying axioms. CNL's and other approaches to making the language of logic more accessible to SME's is indeed an important and integral part of gaining wider acceptance and usage of ontologies. Unfortunately, it is only part of the story, and misses a whole slew of cases where, no matter the syntax, some concepts just aren't readily accessible to an SME with only a modicum or intermediate level of training in knowledge representation.

Best,

Ali

_________________________________________________________________
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    (01)