Re: model theory for RDF/S

>Pat Hayes wrote:
>>
>>  >"Peter F. Patel-Schneider" wrote:
>>  >>
>>  >>  I am concerned that this model theory locks RDF into a particular
>>  >>  way of interpreting literals, namely that the interpretation 
>>of a literal
>>  >>  can be completely determined from its label, using a fixed mapping to
>>  >>  literal values.
>>  >
>>  >I believe this is by design. It's an important requirement on
>>  >RDF syntax that it be "context free"... i.e. that this
>>  >level of meaning is syntactically evident.
>>
>>  Context free, right, but that 'syntactically evident' may depend on
>>  more than a simple label, as far as the model theory is concerned.
>
>I don't think so... this is the whole point of
>the global XL mapping, no? i.e. once you've
>got the lexical form of a literal, you can
>immediately conclude what it denotes.
>
>>  The MT is (deliberately) agnostic about the exact nature of lexical
>>  items in any particular lexicalization.
>
>Then what am I to make of this?
>
>   "We assume a global set LV of literal values and a fixed
>   mapping XL from the set of literals to LV."

I'm not going back on that.

The issue as I see it (maybe Im confused here) is what exactly counts 
as the lexical form of a literal, which is a question about 
lexicalization syntax.  Once that is decided, the MT says that any 
such literal thingie has to have the same *value* (above quote); but 
it is careful not to take a position about what literals really 
*are*. So here are some possible answers:

1. a certain class of character strings (maybe defined by some BNF somewhere)
2. a certain class of tagged strings, so that eg any string enclosed 
by <rdf:li parseType="literal".> .....</rdf:li> counts as a literal, 
and they can be distinguished by other xml attributes (eg xml:lang).
3. Like 2, but the literal includes the tagging (a special case of 1, in fact)
4. a certain class of abstract structures that can be used to label 
RDF graph nodes, perhaps definable in xml terms but also in some 
other way.

So on view 2, for example, the simple string
123
might or might not be a literal, depending on whether or not it was 
enclosed by <rdf:li parseType="literal".>...</rdf:li> .
On view 3,
123
is definitely not a literal, but
<rdf:li parseType="literal".>123</rdf:li>
is.  The model theory says that (once identified by some syntactic 
rule), the literals have a fixed value, but it can't possibly decide 
what that syntactic rule is.

Pat


-- 
---------------------------------------------------------------------
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 Friday, 28 September 2001 11:33:09 UTC