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

Re: the meaning of RDF tokens

From: pat hayes <phayes@ai.uwf.edu>
Date: Wed, 5 Feb 2003 13:41:52 -0600
Message-Id: <p05111b06ba66f95ef1e2@[]>
To: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
Cc: www-rdf-comments@w3.org

>The RDF primer, in Section 2.2, states
>         Using URIrefs as subjects, predicates, and objects in RDF
>         statements allows us to begin to develop and use a shared
>         vocabulary on the Web, reflecting (and creating) a shared
>         understanding of the concepts we talk about. For example, in
>         the triple
>		ex:index.html  dc:creator  exstaff:85740 .
>         the predicate dc:creator, when fully expanded as a URIref, is an
>         unambiguous reference to the "creator" attribute in the Dublin Core
>         metadata attribute set (discussed further in Section 6.1, a
>         widely-used set of attributes (properties) for describing
>         information of all kinds. The writer of this triple is effectively
>         saying that the relationship between the Web page (identified by
>         http://www.example.org/index.html) and the creator of the page (a
>         distinct person, identified by
>         http://www.example.org/staffid/85740) is exactly the concept
>         identified by http://purl.org/dc/elements/1.1/creator. Moreover,
>         anyone else, or any program, that understands
>         http://purl.org/dc/elements/1.1/creator will know exactly what is
>         meant by this relationship.
>This appears to me to state that the meaning of tokens in RDF *is*
>their commonly agreed on meaning, regardless of how that meaning is

Yes, in a sense, but...

>If so, this means that RDF reasoners are responsible for
>implementing this meaning.

....no, because that conclusion follows form a different sense of 
'meaning'. See below.

>Is this actually the case?  If so, how can RDF reasoners be implemented?
>If not, please explain what the above quote means.

(Speaking for myself here, not with the voice of any official WG, but 
what the hell...)

My own reading of this issue agrees largely with yours, Peter, but I 
do not draw the apocalyptic conclusions that you do from it. I think 
you are reading more into it than is intended.

The point is that there is not a single notion of 'meaning' being 
used in these debates. The above passage refers to a broader sense of 
'meaning' than the one specified by the model theory.

Let me take a very simple example to illustrate the point. Suppose 
some RDF is published by A with the following form:

#antelope rdf:type rdfs:Class .
#antelope rdf:comment "This is intended to be the class of all 
antelopes, considered as a species."

and B uses this uriref in some other place as follows:

#Arthur rdf:type A#antelope .

Now, nothing much follows formally from these two taken together: but 
a person (not some RDF software) reading this, and knowing the formal 
meaning of things like rdf:type, might well conclude that B's 
intention was to assert that Arthur was an antelope. I hope you don't 
find this a ridiculous or philosophically risky observation.

The point of all this 'social meaning' talk is only to say that 
'meaning' in this very loose, broad, warm fuzzy human sense (which we 
do not need to make precise: choose your own sense, in fact) is not 
somehow cancelled or abrogated whenever one performs a formal 
inference using RDF. So for example, if B also asserts the RDF:

#PetAntelope rdfs:subClassOf A#antelope .
Joe rdf:type #PetAntelope .

then the formal RDF semantic rules allow one (human or machine) to 
conclude that

Joe rdf:type A#antelope .

and - and this is the point - the fact that some FORMAL inference 
intervenes, as it were, between the assertion and the conclusion 
should NOT be used, or useable, as an argument that this conclusion 
differs in meaning, in any way or sense of 'meaning' at all, from the 
first B assertion above: so the same informal conclusion about being 
an antelope that one might be justified in inferring from the comment 
in the first case, also applies, and in the same way and for the same 
reasons, in the second case. Even though the second case involves a 
formal inference step and the first one doesn't.

There is NO implication intended here that this 'informal' inference 
is somehow magically made formal, or that software is supposed to be 
able to understand English comment strings. The point is rather to 
avoid an argument in the other direction, an argument that I suspect 
may never have even occurred to you (it hadnt to me until I got 
involved in this debate), along the following lines:

"Formal meaning is just a mathematical curiosity and has nothing to 
do with Real Meanings (the kind that really Matter in Human Discourse 
in Society, or whatever), so whenever any formal inferences are done, 
the formal conclusions lose all their Real Meaning and are just 
mathematical curiosities of no real significance, devoid of any Real 
Meaning content outside some narrow abstract mathematical domain."

The point of fussing over this comes from the expectation that RDF 
formal inference processes will in fact be involved in larger 
processes which are indeed social in nature at both 'ends' - such 
things as sending orders, making bids, deciding to buy something, 
giving permissions to charge to an account, that kind of thing; and 
to pre-empt any attempt by someone (who will be nameless but whose 
name is Legion) to claim that RDF/RDFS/OWL/whatever cannot be used to 
do this because they can only convey 'formal' meanings and this stuff 
is all to do with 'social' meanings, and you can't get a social 
meaning down a narrow formal pipe. The point being that although (of 
course) an RDF engine cannot be expected to understand, or have its 
own behavior influenced by, any meaning of any RDF it is processing 
beyond that specified by the model theory, nevertheless this 'other' 
meaning may be in a sense attached  to the RDF (invisibly to the RDF 
engine, but still attached; and maybe visible to other entities at 
either end of the process) and get conveyed from one place to another 

Here's an analogy: think of RDF as something like the post office. 
All it *understands* is addresses; but there can be all kinds of 
things inside the envelopes and packages. It doesn't need to *read* 
that inside stuff - for all the post office knows, the packages could 
all be empty -  but it does get them moved from place to place. The 
formal meaning is the address; the social meaning is the stuff in the 
package. And the point is just to say that even though the stuff is 
inside and not mentioned explicitly in the address, it nevertheless 
still gets delivered.


IHMC					(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.			(850)202 4416   office
Pensacola              			(850)202 4440   fax
FL 32501           				(850)291 0667    cell
phayes@ai.uwf.edu	          http://www.coginst.uwf.edu/~phayes
s.pam@ai.uwf.edu   for spam
Received on Wednesday, 5 February 2003 14:40:09 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 21 September 2012 14:16:31 GMT