W3C home > Mailing lists > Public > www-rdf-logic@w3.org > July 2002

RE: questions on assertion

From: pat hayes <phayes@ai.uwf.edu>
Date: Tue, 9 Jul 2002 12:07:21 -0700
Message-Id: <p05111b2eb950e2a8244b@[192.168.0.146]>
To: "Joshua Allen" <joshuaa@microsoft.com>
Cc: www-rdf-logic@w3.org

>  > In this case, I am trying to figure out in that case how the RDF model
>>  theory would cope with expressing the following.
>>
>>  1. my car is red
>
><rdf:Description rdf:about="urn:autos:my-car" rdf:ID="Statement1">
>   <ex:Color>Red</ex:Color>
></rdf:Description>
>

OK, but....

>  > 3. X is not true.
>
><rdf:Description rdf:about="#Statement1" rdf:ID="Statement3">
>   <ex:Veracity>False</ex:Veracity>
></rdf:Description>

....no, that won't do. Unlike ex:Red, the meaning of 'false' has to 
be connected in a particular way to the semantics of the language 
itself: one wants 'X is false' and 'X' taken together to be a genuine 
contradiction. But there are no genuine contradictions of this kind 
in RDF.

Also, by the way, you will need to use reification to get 
'#Statement1' attached properly to that first statement. RDF requires 
full urirefs, not fragIDs.

>  > 4. my car has four wheels
>
><rdf:Description rdf:about="urn:autos:my-car" rdf:ID="Statement4">
>   <ex:WheelsCount>4</ex:WheelsCount>
></rdf:Description>
>
>>  6. X is an assertion made by P
>
><rdf:Description rdf:about="#Statement1" rdf:ID="Statement6">
>   <dc:Author>P</dc:Author>
></rdf:Description>
>
>>  7. Y is an assertion made by Q
>
><rdf:Description rdf:about="#Statement4" rdf:ID="Statement7">
>   <dc:Author>Q</dc:Author>
></rdf:Description>
>
>>  1. If we interpret an assertion to mean "I believe 'my car is red' is
>>  true."
>
>More like "Someone asserted that ('my car is red' is true)".

I think the idea is that if *you* publish it (in what might be called 
'publish mode', eg on your home page and not quoted or attributed to 
others) then *you* are asserting it. Exactly what counts as 'assert 
mode' is a bit murky, but some such principle seems to be needed to 
ground assertions out into some source which can be held responsible 
for making the assertion.

>
>>  "I believe ["I believe 'my car is red' is true"] is false"
>>  Which is a paradox.
>
>Someone asserted that (asserting ('my car is red' is true) is false))
>
>>  So the problem I am getting at, is how can say, without creating a
>logical
>>  inconsistency, that one believes a statement in rdf data is false?
>
>Just because you have two conflicting assertions does not mean that you
>have chosen to believe either one of them.
>
>>  This is in my view a real problem for applications involved in
>reputation
>>  and trust.
>
>Actually, I think that trust in metadata depends on people being able to
>make statements like number 3.  This is exactly what is needed to allow
>you to choose what assertions to trust.  For example, assume that your
>list has a few more assertions:
>
>8. Statement 3 is an assertion made by R
>9. Statement 6 is true
>10. Statement 7 is false
>11. Statement 8 is true
>12. Statement 9 is made by your tamper-proof digital signature checker
>13. Statement 10 is made by your tamper-proof digital signature checker
>14. Statement 11 is made by your tamper-proof digital signature checker
>
>Now, if you can determine that R is someone you routinely trust, you can
>discard assertion #1, and store some internal information about person P
>so that you know to be suspicious of him in the future.  Else, if *you*
>happen to be person P, you can disregard R and put a ding against him in
>your reputation database.  And if P and R turn out to be the same
>person, you can just discard both assertions, since they are coming from
>a schizophrenic (or you could choose to take the most recent one,
>assuming the guy changed his mind or repainted the car).

Right, and wouldn't it be great if we actually could do this kind of 
stuff in RDF?? :-)

>  > 2. If rdf statements implicitly carry assertion, how can I specify the
>>  author of the assertion? That is - does the assertion implied by 1.
>also
>
>As I showed in 6 and 7, and assuming your subsystem can assert 9-14 for
>you.
>
>Also, note that it is not necessary to decorate every assertion like
>this.  You could wrap assertions in collections -- this is how Klyne
>Contexts work.

Almost every deployed application of RDF uses some such mechanism, 
but unfortunately RDF itself does not (yet) include it as a syntactic 
device.

Pat Hayes



-- 
---------------------------------------------------------------------
IHMC					(850)434 8903   home
40 South Alcaniz St.			(850)202 4416   office
Pensacola,  FL 32501			(850)202 4440   fax
phayes@ai.uwf.edu 
http://www.coginst.uwf.edu/~phayes
Received on Tuesday, 9 July 2002 15:07:09 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:52:42 GMT