W3C home > Mailing lists > Public > www-rdf-comments@w3.org > January to March 2003

Re: summary of reification semantics issues (material for discussion).

From: Seth Russell <seth@robustai.net>
Date: Fri, 14 Mar 2003 11:20:33 -0800
Message-ID: <3E722B81.20209@robustai.net>
To: pat hayes <phayes@ai.uwf.edu>
CC: www-rdf-comments <www-rdf-comments@w3.org>

pat hayes wrote:

>> Pat,
>> You say:
>> [[
>> One objection to the de re interpretation is that it does not allow
>> for the adequate representation of propositional attitudes such as
>> belief. This is controversial (see the discussion of the Russellian
>> theory in http://plato.stanford.edu/entries/prop-attitude-reports/) ,
>> but in any case there is ample experience which suggests that the de
>> dicto interpretation would produce other problems with the
>> representation of such ideas, and that an fully adequate
>> representation of propositional attitudes is unobtainable using
>> reification alone.
>> ]]
>> The only problem I know of for the de re interpertation of the 
>> reification syntax is that you can't substitute equivalents.
> With de re semantics, you CAN substitute equivalents. The problem 
> which arises with propositional attitudes is that such substitution 
> produces unintuitive results (arguably) when it is done using an 
> equivalence outside the propositional attitude report but applies it 
> inside (as in the Superman example). The only adequate way to handle 
> this is to have a mechanism for keeping track of the scope 
> distinctions between propositions asserted inside and outside such 
> attitude reports, eg distinguishing
> s=k  and  P(s)  --->   P(k)
> from
> s=k and LL believes[....P(s)...]  --/--> LL believes[....P(k)...]
> from
> LL believes[.... s=k and P(s)...]  ---> LL believes[....P(k)...] 

Yes I was already assuming that, but thanks for expressing it again here 
so clearly.  

>> It seems to me that the solution of that would be to come up with a 
>> property for equating URIRefs that would exclude substitutions into 
>> reification triples.  In other words, why throw out propositional 
>> attitudes, why not just fix the semantics of equivalence ?
> Because the semantics of equivalence is very basic, and doesnt need 
> fixing. Equivalence just means two names denote the same thing. If 
> they are the same thing, then obviously they have the same properties. 
> The problem arises because when reading prop-attitude claims we often 
> read them as saying that *the person with the attitude* would agree 
> with this if it were presented to them *in this form*, and the basic 
> problem is that this conflates issues of what the facts ARE with 
> issues of HOW the facts are SAID. That is why the middle entailment is 
> considered to be incorrect, not because LL doesn't believe that this 
> actual guy, Superman=Kent, is strong - she does, as a matter of fact - 
> but because if you were to ask her, is Kent strong? she would say 
> 'no'. Thats because she doesnt know who Kent actually is, of course; 
> but the last thing we want is for our basic logic to get as confused 
> as Lois is about the way things really are.
>> If it is too late to fix dmal:equivalentTo, then invent a OWL 
>> property that would exclude sustititions in reifications.
> That amounts to adopting the de dicto interpretation. 

But only for the purpose of substititions in opaque contexts.  The 
object of rdf:Subject is still the girl, not "ex:Mary".  We are just not 
allowed to substitute "ex:MaryOfLamb" for "ex:Mary" in the opaque 
context of reification triples even though they may refer to the same 
girl.  My question is whether disallowing those ~erronious~ 
substitutions would permit us to express some measure of propositional 
attitudes even in the scopeless world of RDF ?  

> What is the justification for not allowing substitutions if those 
> substitutions are in fact valid? 

Because you  *cannot*  substitute two expressions with the same 
reference salva veritate (i.e. keeping the truth value constant) in an 
opaque context.  
And, I am assuming that reification *does* specify a opaque context.  

>> What (if any)  are the other problems with propositional attitudes 
>> exclusive of this one?
> See above. 

May I take it, then,  that variations on this substitution theme are the 
*only* problem with propositional attitudes and RDF reification ?

Seth Russell
Received on Friday, 14 March 2003 14:20:50 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:15:20 UTC