Re: use/mention and reification

Patrick Stickler wrote:

> On 2002-01-23 17:42, "ext Frank Manola" <fmanola@mitre.org> wrote:
> 
>>...  
>>
>>2.  If two people use different non-URIs, like "Superman" and "Clark
>>Kent", to refer to (apparently) different things, the same thing holds.
>>We know they've used different names, but we may still infer they are
>>talking about the same thing if we get enough additional information.
>>
> 
> The same "thing" in the universe, yes, but not the same RDF resource.
> We wouldn't want all nodes with Superman and ClarkKent URIs to
> be merged, even if we determine that they denote the same "thing".


If we really determined that they truly denote the same "thing", we 
certainly would like to try! There are many highly practical examples 
where you'd want to get rid of one of the identifiers and agree to use 
the other.  The situation here, though, is Superman and ClarkKent URIs 
don't, strictly speaking, denote the same thing (they denote different 
personas that the same physical being assumes at different points, but 
we obviously need to keep them apart).  Also, there are practical 
difficulties with "legacy data" using both URIs that is hard to revise. 
  Finally, all this is sort of "theoretical":  we have no way to merge 
the  URIs (or even, in RDF, assert they are equivalent;  you can in DAML).

> 
> 
>>3.  If two people use the same URI, like ex:Superman, to refer to
>>something, we know that they *think* they are talking about the same
>>thing, 
>>
> 
> This equates to my present view, that if two or more folks
> use the same URI, they are agreeing that it denotes the same
> "thing", whatever that "thing" may be.
> 
> That doesn't mean that some "thing" in some universe may not
> be denoted by multiple URIs, or by "constellations" of properties,
> including literal encoded names, etc. Only that, insofar as that
> single URI is concerned, its use represents agreement according
> to what the minting authority says it denotes.
>


We need to be careful here.  I agree that a given "thing" in the 
universe may be denoted by multiple URIs and so on.  However, in my view 
of the way this will probably work, you may create a URI and describe 
what it means (so you're the minting authority).  I may starting using 
it with the idea that I agree with your meaning, but you may never know 
about it (so you don't necessarily agree).  In that case, strictly 
speaking the agreement is only one-way (it would be inaccurate to say 
"we are agreeing").

 
> 
>>but they may not be.  E.g.,
>>
>>a.  One person may have introduced the URI ex:Superman, and someone else
>>starts using it, but the second person actually has something else in mind.
>>
> 
> It is IMO the right of the minting authority to specify what
> a URI means. And if someone else uses it incorrectly, even
> if unintentionally, they shouldn't usurp ownership of the meaning
> of  that URI but should either conform to the specified meaning
> or use another URI. Obviously, that's a social issue (or legal
> one) not a technical issue.


I agree that it's the minting authority's right to specify what a URI 
means, but I don't see how we can be very strict in interpreting this. 
For example, the only way I have of understanding what you mean a URI to 
denote is how you describe it using other properties.  You may have left 
some details out of your description, so I may be misled as to what your 
denotation is, and start using it incorrectly (as far as your intended 
meaning goes).  I wouldn't say I was usurping ownership of the meaning 
in that case;  I may have been *trying* to conform to the specified 
meaning and just not gotten it right.  That's going to happen.  I agree 
it's not a technical issue, but we also need to be careful how 
well-behaved we expect the world to be in designing our technology.


> 
> 
>>b.  The two people involved have the same entity in mind, but have
>>different understandings about that entity's attributes.  Consider
>>ex:Clark Kent.  One person thinks Clark is identical to Superman, and
>>the other (Lois) doesn't.  One person thinks Clark has Xray vision (and
>>conceals it), the other thinks Clark has ordinary vision.
>>
> 
> But both Clark Kent and Superman would presumably be denoted by
> different URIs, and belief could be expressed in terms of either.
> 
> If Superman has Xray vision, and it is later determined that Superman
> is the same thing as Clark Kent, then one many infer that Clark
> Kent also has Xray vision. But even if the are the same thing in
> the universe, they are different resources in RDF space (if
> denoted by different URIs).


If by "different resources in RDF space" you mean RDF has to treat them 
differently, I agree (and in any case, Superman and Clark Kent are 
slightly different, as I've suggested above)).  But they may still be 
the same resource, right?


> 
> 
>>I don't see how we can reasonably expect URIs to be used only in
>>situations where everyone on the Web has a full understanding and
>>agreement on all the attributes of the thing the URI names before they
>>start using it. 
>>
> 
> There need not be agreement about all attributes of a thing, only
> about what thing a given RDF resource denotes.


The *intent* is certainly that use of the same URI indicates that 
agreement.  But I repeat that I only know what someone else denotes by a 
URI (at least, general URIs denoting non-digital things) by their 
description of it in terms of other attributes.  And I may, on the basis 
of the description you present, think we have an agreement when we 
really don't, because we disagree on other attributes that you didn't 
include in your description.

Also, because people determine what names denote on the basis of other 
information about them, there are frequent problems that occur where one 
person thinks that, on the basis of disagreement about descriptive 
properties, they are talking about different things, and another thinks 
that they are talking about the same thing, and the other person is 
simply wrong about that thing's attributes.


> 
> 
>>What we do get using URIs is the extra bit of
>>information suggested by #3 above:  that when people use the same URI,
>>they at least *think* they are talking about the same thing.
>>
> 
> Of course one cannot prevent misunderstanding, or even disagreement,
> on the SW, but agreeing on the same URI to denote the same thing
> seems a reasonable basis to conduct discussions, and we can hope
> that knowledge expressed in terms of RDF and URIs will be less
> ambiguous than knowledge expressed in natural language ;-)


I agree.  My only point is there may be some additional conversation 
necessary to resolve such misunderstandings, even after we *think* we 
have agreement (and the Web provides a medium for conducting those 
conversions).


> 
> 
>>That's 
>>useful information!  You can merge their different beliefs via the URI
>>in their various RDF graphs, and then possibly detect inconsistencies,
>>which could then lead to better agreement after some dialog.
>>
> 
> And that's the beauty of RDF. Users who use URIs are not agreeing
> about all of the attributes of a thing. Only the denotation of
> that thing by which they describe the attributes of that thing
> that *they* know about, or by which they obtain attributes known
> about that thing by others.
> 
> Then knowledge is pooled, including knowledge about
> the equivalences of RDF resources which denote the same things,
> and viola, we get a synergetic merging of knowledge that may
> be far more than the simple sum of its parts.
> 
> And if that merged body of knowledge contains contraditions
> (which it likely will) then as you suggest, these could serve
> to clarify common understanding (or misunderstanding).
> 
> Of course, there can still be disagreement even if everyone is
> in complete agreement about what thing a given URI denotes.


Yes.


> 
> 
>>Another way of putting this is to ask if not allowing URI labeled
>>resources to have different denotations means that different people
>>can't attribute contradictory things to the same URI labeled resource?
>>
> 
> I think that folks can attribute contradictory things about the
> same URI labeled resource without requiring that it denote more
> than one thing.


Yes.


> 
> I consider the N-to-1 mapping of RDF resource(s) to a thing as
> fundamental and essential as the N-to-1 mapping of lexical form(s)
> to a datatype value. If we allow N-to-N mappings then we lose all
> clarity in our statements. I.e. if a given URI can denote more than
> one thing in a universe, then we'd never know precisely what it is
> anyone is talking about.


Strictly speaking though, the problem isn't that we're allowing a given 
URI to denote more than one thing.  The problem is how well (accurately) 
we can transmit that denotation from the minting authority so that it is 
(again, accurately) shared by others who are going to use it.


> 
> This leads us right into the heart of the URI/URL/URN/URP debate
> where some folks think that e.g. 'mailto:john.doe@abc.com'
> can denote both an email URL and a person, or 'http://www.abc.com' can
> be used to denote both a web page URL and the company ABC Inc.
> This introduces an N-to-N mapping from the set of RDF resources
> to the set of things in the (or a) universe such that if
> one says
> 
>    <http://www.abc.com> <dc:creator> "John Doe" .
> 
> one cannot know if John Doe is the creator of the web page or
> of the company.
> 
> The "contemporary view" seems to take the view that "a URI is
> whatever it is used for" similar to Perl scalars, yet this
> introduces this N-to-N mapping from resources to things that
> makes inference about knowledge untenable.
> 
> I keep thinking that surely there must exist some axiom
> or "law" in KR or AI that forbids such N-to-N mappings
> in a KR system (my ignorance shames me).
> 
> Anyone?
> 
> 
>>--Frank
>>
> 
> Cheers,
> 
> Patrick
> 
> 
> 
>>Patrick Stickler wrote:
>>
>>
>>>On 2002-01-23 14:10, "ext Graham Klyne" <Graham.Klyne@MIMEsweeper.com>
>>>wrote:
>>>
>>>
>>>
>>>>At 01:46 PM 1/23/02 +0200, Patrick Stickler wrote:
>>>>
>>>>
>>>>>>OTOH, without allowing
>>>>>>that it seems that URI-refs don't offer anything that label strings like
>>>>>>"Lois" and "Jimmy" don't also offer.  The point of my comment was to
>>>>>>suggest that matters relating to personal belief of identity shouldn't
>>>>>>really be expressed in terms of URIs.
>>>>>>
>>>>>>
>>>>>But your examples are not about belief of identity, they
>>>>>are about belief of properties of entities which are given
>>>>>identity.
>>>>>
>>>>>I.e., you weren't saying that <person:Lois> <ex:accepts> that
>>>>>"Superman" <ex:is> <person:Superman>.
>>>>>
>>>>>
>>>>My take on this scenario was that Lois' non-belief that Clark Kent is
>>>>strong is rather bound up with her belief about the identity of the person
>>>>she knows as "Clark Kent";  i.e. that she does not recognize him as also
>>>>being the person she knows as "Superman".
>>>>
>>>>I think some alternative formulation of this scenario could lead to a
>>>>situation in which Clark Kent and Superman have universally accepted
>>>>denotations, and URIs may then be appropriate.  In this case, I think that
>>>>it's not possible that they denote exactly the same thing;  e.g. Clark Kent
>>>>denotes a person X wearing a suit and glasses;  Superman denotes the same
>>>>person X wearing a natty blue-and-red number.  In this formulation, using
>>>>URIs seems less troublesome.
>>>>
>>>>
>>>I see your point. Though I'm not sure that the Dan's do. Talking about
>>>some person X named "Superman" (who might later be deemed to equate to
>>><person:Superman>) is not the same as using "person:Superman" rather
>>>than <person:Superman> to avoid instantiating a URI labeled resource
>>>node in the RDF graph.
>>>
>>>Eh?
>>>
>>>Patrick
>>>
>>>--
>>>               
>>>Patrick Stickler              Phone: +358 50 483 9453
>>>Senior Research Scientist     Fax:   +358 7180 35409
>>>Nokia Research Center         Email: patrick.stickler@nokia.com
>>>
>>>
>>>
>>>
> 
> --
>                
> Patrick Stickler              Phone: +358 50 483 9453
> Senior Research Scientist     Fax:   +358 7180 35409
> Nokia Research Center         Email: patrick.stickler@nokia.com
> 
> 
> 


-- 
Frank Manola                   The MITRE Corporation
202 Burlington Road, MS A345   Bedford, MA 01730-1420
mailto:fmanola@mitre.org       voice: 781-271-8147   FAX: 781-271-875

Received on Wednesday, 23 January 2002 15:53:41 UTC