ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] Grover Models

To: "[ontolog-forum] " <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "Barkmeyer, Edward J" <edward.barkmeyer@xxxxxxxx>
Date: Wed, 22 May 2013 11:28:18 -0400
Message-id: <63955B982BF1854C96302E6A5908234417DF2D4E53@xxxxxxxxxxxxxxxxxxxxxxxxxx>
John is spot on.  This illustrates a popular misconception that arises from 
thinking that models follow naturally from careful natural language.
Comments below.    (01)

> -----Original Message-----
> From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx [mailto:ontolog-forum-
> bounces@xxxxxxxxxxxxxxxx] On Behalf Of John F Sowa
> Sent: Wednesday, May 22, 2013 10:32 AM
> To: ontolog-forum@xxxxxxxxxxxxxxxx
> Subject: Re: [ontolog-forum] Grover Models
> 
> On 5/21/2013 11:52 PM, John McClure wrote:
> > Is it actually wrong to model with prepositions, or is it just not a
> > recommended practice?
> 
> There is nothing syntactically "wrong" in using any alphanumeric string you
> prefer.  But English and other NLs resolve ambiguities by looking at a context
> of more than one word.
> 
> In the usual notations for logic, including RDF, the parser or theorem prover
> does not use context.  Any two strings that are spelled the same are treated
> as identical.
> 
> JMc
> > Below I'm keen... to resolve both 'has' and "is-a" Sister kind of
> > queries.
> >
> > [[Person:X]] has [[Sister:Person:Y]]
> > [[Person:X]] of [[Type:Sister:XY]]
> > [[Sister:Person:X]] for [[Person:X]]
> > [[Type:Sister:XY]] has [[Instance:Sister:XY]] [[Instance:Sister:XY]]
> > of [[Type:Sister:XY]] [[Instance:Sister:XY]] for [[Person:X]]
> 
> In these examples, it seems that you are assuming that context (i.e., the
> types of the preceding and following nodes) will somehow distinguish the
> different uses of 'has', 'of', and 'for'.  That is true in English.  It is 
>also true of
> some programming languages that support "polymorphism".
> 
> But RDF and most versions of logic are not polymorphic.
> Therefore, you need to distinguish the different relation types with longer
> names, such as 'has-sister'.
> 
> John    (02)

[EJB]  SBVR makes this mistake:  treating 'has' as a 'verb concept'.  As John 
says, the verb *concept* is 'has sister' or 'has instance'.
And this mistake leads to the second mistake above:  confusing the Role 
'sister' with a Type.    (03)

In dealing with the SBVR lot, I found two excellent bad examples that 
illustrate the problem:    (04)

Example 1:  set has cardinality
'Cardinality' as a role is defined as:  number that is the count of elements of 
a given set.
And the proper verb concept form is then:  (set) has cardinality (natural 
number).  That is:  'has cardinality' is the 'verb concept'.
'Cardinality' as a type is defined as:  number that is the count of elements of 
at least one set.
That 'cardinality' is a type, a subtype of 'number'.  Is 5 a cardinality?  If 
we have a set {a, b, c, d, e} in the world of interest, then 5 is a 
cardinality, but if we don't have a set with 5 members, then 5 is not a 
cardinality in that 'world'.   Now, if 5 is a cardinality in the world of 
interest, can the set {apple, orange} "have" it?    (05)

Example 2:  person has English teacher
'English teacher' as a type is defined as:  person who teaches at least one 
English class.
'English teacher ' as a role is defined as:  person who teaches English to a 
given class or a given person.
Mr. Brown teaches English and History at Hangum High School.  Mr. Brown is an 
(instance of type) 'English teacher'.
Ms. White teaches English at Hangum High School.  Ms. White is an (instance of 
type) 'English teacher'.
Sally is a student at Hangum High School.  Sally has Mr. Brown for History.  
Sally has Ms. White as for English.
If Sally simply 'has' an instance of type English teacher, then Sally 'has' two 
English teachers:  Ms. White and Mr. Brown.
But "Sally 'has English teacher' (person)" is only satisfied by Ms. White.    (06)

Example 3:  I have a wife, and my wife is an instance of the type 'sister' (she 
is sister to someone), but I have no sister.    (07)

The critical distinction here is between 'at least one' and 'a given' in the 
definition.  If you have to 'give' an argument (the set, the person) to 
determine the thing(s) being referenced, then you have a 'role', and the formal 
term for the verb concept (e.g., the RDF verb) must somehow include or imply 
that role.  If you don't care which thing(s) produce the classification, then 
you have a 'type', and the things being referenced are all things that satisfy 
the type.    (08)

-Ed    (09)

> __________________________________________________________
> _______
> 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
>     (010)

_________________________________________________________________
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    (011)

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