ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] Compound nouns

To: <ontolog-forum@xxxxxxxxxxxxxxxx>
From: jmcclure@xxxxxxxxxxxxxx
Date: Sat, 06 Apr 2013 16:13:53 -0700
Message-id: <948b6ac61a300d199057e475bf9cb2fd@xxxxxxxxxxxxxx>

Hello all, first post....

I heartily support an explicit grammatical orientation to an ontology whenever implemented using semantic mediawiki. I claim most singular nouns are best implemented as real or aliased wiki namespaces, e.g., [[Project:ABCD]] where "Project" is the wiki namespace in which pagename "ABCD" is located, and "Project" is the name of a type of page which itself is defined in a "Type" namespace (e.g., on page [[Type:Project]]).

Similarly, I think plural nouns are best represented within SMW's Concept namespace, explicitly identifying criteria for inclusion in this "class" of things.

"Compound nouns" such as "Roadway Project" can be wiki namespaces or can be pages in a "Project" namespace that are of a certain category (say, "Category:Technical") or those having a certain property (say, "Property:Road") or those having a [[Property:Subject]] which are in turn of type [[Type:Roadway]]. Whichever is then named as a criteria in the {{#concept:}} parser function on a [[Concept:Roadway Projects]] page. Whichever applies (or whether some other kind of "implementation" of instances of a type is to apply, e.g., subobjects) is defined by [[Type:Roadway Project]]. Please note that "concepts" may be used like a category within Semantic Forms.

I suggest no nouns are in the Category namespace; rather, adjectives (and past participles and adverbs) are found there. A rigorous facet programming style emerges from this approach: any given facet implies zero or more properties that should be attached to, or can be impled to exist for, a page. This perfectly fits with "semantic gardening bots" whose princiapl function is to imply facts for a page, whether they be categories or whether they be a particular combination of properties, that are to be attached to the page.

I realize a "Type" namespace may be a major departure for those accustomed to the Category namespace being used for all class definitions, but really, singular nouns simply don't belong in the Category namespace which has been meant to be used, since the earliest days of Ward Cunningham, as a LIST of things. This "collection" idea fits well with RDF as it is based on set-theory, aka "extents", but ya know, RDF itself has some tension there, as it calls the global class-defining property "rdf:type", not "rdf:class", when it points to an object of type rdfs:Class or owl:Class.

Here is another way to look at it. Is [[User:John Sowa]] a Person? or a Persons? Clearly John is a Person, John is not a ListOfPersons.

Consequently, my suggestion is to make maximal use of SMW's Concept namespace for plural simple or compound nouns, and a (newish) Type namespace for singular simple or compound nouns. Then, use the Category namespace only for descriptors (adjectives, past participles and adverbs). Then, use the Property namespace only for prepositions, verbs, and a small set of provenance-related properties, such as Dublin Core. In other words, there should be no property named "property:Road" because, by Occam's Razor, such duplication with type-names (i.e., pages in a Type namespace), should be avoided.

Bottom-line, I believe basic grammar can play a major role in de-mystifying ontologies. And I hope my comments are not totally off-topic here! thanks --john

 

On 06.04.2013 14:41, John F Sowa wrote:

On 4/4/2013 12:40 PM, doug foxvog wrote:
What i have been pushing for is that we distinguish when we are referring to ontological concepts and when we are referring to NL concepts. A term in an NL is expected to be fuzzy. A term in a formal ontology should be expected to be far less fuzzy.
Some distinction along those lines is necessary, but you can't draw
a sharp line between NL and formal.  A research report written in
English about chemistry will use many chemical terms in a formal
ontology, but it will also use many other English words for which
much less formal resources like WordNet and Roget's Thesaurus are
useful.

The same is true for professionals in any field.  An article about
chess or bridge will use formally defined terms about those games
with as much precision as any formal ontology.  But it may also
include background info about the players and informal comments
about how they play the game.

In medicine, a physician may use formal terms in talking with
a colleague, and very informal terms in talking with a patient.
Even when speaking with another physician, the level of formality
can vary widely.  When they're trying to diagnose a disease, they
may use very informal terms to describe the symptoms.

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    (01)

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