Re: Examples of Syntax and semantics

On 5 Apr 2006, at 15:37, adasal wrote:
> Henry,
> It seems rather late in the day to be defining the diference  
> between syntax and semantics, with regard to which imagine the  
> following.
> A painting is fully described in mathematical terms. You can go to  
> any level of complexity, devise any scheme you like to describe the  
> texture of the paint, the juxtaposition of colours, what you will  
> to *completely* describe it, but the language and its syntax is so  
> at odds to that of painting you will not be able to convey the  
> meaning of the painting to another human in this way.
> So it is, less dramatically, between one ontology and another  
> although ostensibly about the same subject. Without an underlying  
> framework what are such ontologies predicated on? Trial and error  
> in the popularity stakes?

Yes. Natural selection is the final arbitrer. Though others like to  
refer to it is as God.  Since you like quoting philosophy books you  
may want to check out Ruth Garrett Millikan's "Language, Thought, and  
Other Biological Categories: New Foundations for Realism" [1]

> You said:-
> < http://xmlns.com/foaf/0.1/Person> (Let's abbreviate that to
> foaf:Person from here on) always refers to the class of persons.
> foaf:knows always to the same relation between persons.
>
> And I understand that the point is that the a foaf person in  
> defined is available to all at the URL.
> But the problem comes when I want to redefine that which would  
> manifest in either extending http://xmlns.com/foaf/0.1/Person or   
> by my own definition http://xmlns.com/asaltieldeffoaf/0.1/Person
> or, worse, where there is an implicit change. In fact this has  
> happened in foaf:knows since the way this is used entails a very  
> attenuated sense of "knows" due to foaf based programs (e.g.  
> LinkedIn). One certainly cannot rely on foaf:knows in the way that  
> one ordinarily uses the term "I know June and I know Justin."

foaf:knows is defined at the url http://xmlns.com/foaf/0.1/knows. It  
is explicity not the general term knows
such as defined by Nozick in Philosophical Investigations

	S knows p iff
		p
		S believes p
		if p were false S would not believe p
		if p were true S would believe p

Since for one the foaf:knows relation relates 2 people.

There are an infinite number of relations between things.  You can  
define the relations you wish. But if they are not useful they will  
just be laying around unused.

> (And notice the shades of meaning that can be attached to "I know  
> June" or, again, "I know June and Justin.")

foaf:knows is a specialization of those usages of 'knows'. It need  
not be equivalent. It need just be useful.

> So I may wish intenstionally to adjust foaf to accommodate some of  
> these shades of meaning.
> I then may mistakenly think that I can normalize http://xmlns.com/ 
> asaltieldeffoaf/0.1/Person to http://xmlns.com/foaf/0.1/Person

Perhaps you would better buy your own domain name adasal.info and  
then you would put your new ontology
at http://adasal.info/foaf/0.1/

The point is that because the URLs used to name the things are also  
closely related to the name of the document where
these can be found, it is both easy to find the meanings of the  
names, and also difficult to override the meaning of the legitimate  
owner.

> by mapping some essential property of my definition of knows to  
> foaf:knows and leaving the other properties as (? I don't know the  
> technical details but I will imagine this) anonymous properties in  
> foaf effectively extending foaf to my own purpose but in an  
> unsatisfactory way.

How you do that will be important. You could do it badly and end up  
with an empty class. Or you could end up with something that worked  
differently than what you think.

> What would have happened at this point is that my original meaning  
> would be lost, this is because foaf:knows has its own meaning, due  
> to usage as much as anything, and my sense of asaltieldeffoaf:knows  
> such as knows as good friend, knows as in the brief acquaintance of  
> a collegue and so on, would not fit into foaf:knows.

ok so you want adasal:goodFriendknows . Just put the following into a  
file http://adasal.info/foaf/ontology.owl

adasal:goodFriendKnows rdfs:subPropertyOf foaf:knows;
                        rdfs:subPropertyOf adasal:friend;
		rdfs:comment "a specialised way of knowing between people that are  
good friends" .

you could also do this as an N3 rule

{ ?p adasal:goodFriendKnows ?f } => { ?p foaf:knows ?f;  
adasal:friend ?f . } .

> You said:-
> The trick is that the names you use (URIs) from one language to the
> other are always the same. The semantics is captured in the
> relationship between the names and the things (which is what
> semantics is about).

Does the above help?

> So how would this work in the example I have given?
> Best,
> Adam


[1] http://www.amazon.com/gp/product/0262631156/qid=1144244641/sr=1-1/ 
ref=sr_1_1/103-6405064-4716631?s=books&v=glance&n=283155

Received on Wednesday, 5 April 2006 14:19:01 UTC