Re: [ontolog-forum] syntax & semantics

["Richard H. McCullough" <rhm@xxxxxxxxxxxxx>]

Hi Jawit

I tried to answer all your questions below.

If I missed something, let me know.

 

Dick McCullough
http://mkrmke.org

----- Original Message -----

From: "Jawit Kien" <jawit.kien@xxxxxxxxx>

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>

Sent: Tuesday, April 28, 2009 10:42 AM

Subject: Re: [ontolog-forum] syntax & semantics

> On 4/28/09, Richard H. McCullough <rhm@xxxxxxxxxxxxx> wrote:
>> ----- Original Message -----
>> Here is a very simple example to illustrate what I mean.
>> My reference for "formal semantics" is the IKL specification document.
>>
>> Consider the English sentence, "Fido is a dog."
>> [The mKR translation is "Fido isu dog;" ]
>> The "real symantics" of English [and mKR] tell us that
>>     a "dog" is a domesticated carnivious mammal, Canis familiaris
>>     "Fido" is the dog identified by Napa, CA animal control license # 1234.
>
> are you coining a term "symantics" to clarify that what you mean by semantics
> may not match other people's expectations?

#### That's a typo.  I meant to say "semantics".
>
>>
>> The IKL translation of this sentence is
>> (and   (isu_rel Fido dog)   (dog Fido)   (individual Fido)  (property dog))
>> The "possible semantics" of IKL tell us that
>>     "dog" is a member of the set of all possible property names.
>>     "Fido" is a member of the set of all possible individual names.
>>     "Fido" is an individual which has the property "dog".
>
> Dick,
> I don't want to be argumentative, but I do want to understand.
> I may not always agree with you, but I value your writing enough to read it,
> and think about what you are saying.
>
> I searched the IKL Guid to understand your IKL translation, at:
> http://www.ihmc.us/users/phayes/IKL/GUIDE/GUIDE.html
>
> but you use terms that aren't in the guide, so the definition
> of those forms, obviously must be elsewhere.
>
> Could you provide some help in understanding your IKL translation?
>
> (and   (isu_rel Fido dog)   (dog Fido)   (individual Fido)  (property dog))
>
> I expect that your statement is a logical formula which is true.
>
> I assume that "and" is the standard logical connective yielding true when
> all the forms within it are true and false otherwise.

#### Yes, see page 4 of IKL Guide.
>
> I will assume that all the forms inside the "and" are also formulas.
>
> The first formula is "(isu_rel Fido dog)"
> I think you said earlier that "isu_rel" is a way of linking an
> instance to a class.

#### isu_rel is the name of the binary predicate which corresponds to

#### the mKR verb isu.  mKR has three verbs related to a genus concept.

#### isu means "is an individual unit of" -- instance

#### iss means "is a species of" -- species

#### isa means "is an individual unit or a species of"  -- instance or species
> I assume that "isa" or "instance" has the same meaning. The link that you
> are stating exists I assume will mean that if there are statements that are
> true for every instance of the class, then they will be known to be true of
> the instance marked in this explicit linking formula.
> So specifically, anything known to be true of all instances of the
> class "dog" is also true about "Fido" as an instance of dog.
> the class "dog" has at least one extension, and Fido is a way of referring
> to that extension uniquely. (is uniqueness guaranteed?)
> you have given us two intensional definitions for "Fido" and for "dog".
>
> looking forward, the next formula is "(dog Fido)"
> I'm not sure what the point of this formula is, since "(isu_rel Fido dog)"
> already says that Fido is an extension of dog. I would have guessed that
> it meant that there is a group/set/class/type named "dog" and "Fido" meets
> the criterion to be considered to be an element/member/instance/_expression_
> for "dog".  But this is exactly what I mean by saying "Fido" is in the extension
> of "dog".   So you will have to clarify, as I am confused. Looking ahead, you
> say that "Fido" is an individual which has the property "dog", so you have some
> idea of a an individual having a property that I don't know yet.
> Do only individuals have a property ?  How is a property different
> than a monadic
> predicate? and of course, how is that different from the extension idea?

#### I didn't use the right words for IKL.

#### Instead of "property", I should have said "monadic predicate".
>
> The third formula is "(individual Fido)", you have said that this means
> "Fido" is a member of the set of all possible individual names.  I'm
> guessing the
> relevant part of that statement must be "individual" rather than "name" since
> I would have said each word in the entire formula was a name,  but you didn't
> so either you are assuming that each word is a name, and not mentioning it,
> or you mean something special by saying "individual".

#### Yes, each word is a name.

#### Looking back at IKL specs again, predicate names and individual names

#### are in the same pool of names.
>
> The fourth and last formula is "(property dog)" which you said meant that
> "dog" is a member of the set of all possible property names. Again, why you

#### One more place where I thought I would make it easier to understand

#### by using "property", and I failed.  My actual translation is

#### "(concept dog)".  I don't want to say "(isu_rel dog concept)" and

#### "(isu_rel Fido individual)" because I consider those statements to belong

#### to a different context.  If I want to include that info. in the same context

#### I use sets:  "(ismem_rel dog concept_set)" and

#### "(ismem_rel Fido individual_set)".
> said this instead of saying (isu_rel dog property), I don't know. are properties
> and predicates the same thing in your mind? I don't see them as the same.

#### As noted above, I should have said "predicate" instead of "property".
>
> By my thought, a predicate is part of a formula, expressing a linking
> relationship.
> John Sowa talks about relationships and concepts, and I think of predicates
> as being relationships used in a logical formula.
>
> Does my intution about all of this make sense to some of the grey beards here?
>
> Anyway, I note that your original statement is what the Cyc folks call a GAF,
> i.e. a Ground Atomic Formula.  There are no variables in it, it is not
> an existential
> formula, ("ThereExists") nor a universal formula ("ForAll"), so it
> just states one
> fact about the universe, presumably for all time, and always true.

#### Yes.
>
> Do you state somewhere else in your computer system that the unique names
> assumption must be known to be true? I can imagine a cat named Fido, and

#### I am assuming unique names.
> If you had your statement in a knowledge base/logical theory then I couldn't
> talk about my cat named Fido in your system, because I assume in your
> system that you have dog and cat as disjoint collections.  I know the
> Cyc folks use
> microtheories to allow both of Fido-the-cat and Fido-the-dog to exist at the
> same time.  Maybe that is what you were trying to do with your context ideas.

#### Yes.
>
> Also how do you handle the fact that Bob's-dog-named-Fido and
> Bill's-dog-named-Fido
> may or may not be the same dog? I can imagine that if Bob and Bill
> were brothers,
> they would both consider the dog-named-Fido as their own, hence having two names
> for the same dog.  But it is just as possible that there are two dogs.
> How do you handle
> this?

#### two names for the same dog:  "name1 is name2;"

#### in IKL, it's  "(= name1 name2)"
>
> JK
> 
> _________________________________________________________________
> 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
> 
>
>

_________________________________________________________________
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)