- From: Brian McBride <bwm@hplb.hpl.hp.com>
- Date: Sun, 02 Mar 2003 10:40:54 +0000
- To: RDF Core <w3c-rdfcore-wg@w3.org>, Tim Berners-Lee <timbl@w3.org>, Bijan Parsia <bparsia@isis.unc.edu>
I believe that the differences of view on what we can say about social meaning derive from different conceptual models of the nature of meaning and that a useful step would be to make those models explicit. I'm probably the last person to attempt such a thing, but by including mine, perhaps I illustrate the sort of thing I have in mind. I strongly urge that we stay away from the term, 'meaning'. It is too philosophical in nature. Folks are concerned about two practical, pragmatic things: o the behaviour of machines o the behaviour of societies We should talk about those. o Triples are syntactic - they are langauge. o Model Theory defines structure of interpretations that relate statements in a language to a world, places some constraints on such interpretations and defines truth preserving transformations on syntax. This is what the RDF model theory does for (sets of) RDF triples. o An interpretation assigns a truth value to (sets of) RDF triples. o A system, such as a machine, is permitted (in some cases expected) to apply truth preserving transformations in carrying out its processing of RDF triples. o A society may, in effect, agree on further semantic restrictions on interpretations, e.g. the world is the 'real world' and on the interpretation of symbols, i.e. on what constitutes a definition of the interpretation of that symbol. An interpretation conforming to such social restrictions might be called a social interpretation (I just know some folks will hate that term, but I can't think of anything better). o Social interpretation is determined by the social institutions of a society - often, ultimately by a system of law. Social interpretation is usually not mathematically precise, evolves over time and cannot be precisely processed by an algorithm (is that shooting a member of the class law:crime?). It is usually described in natural language. o There is some relationship between social interpretations and formal model theoretic interpretations. This relationship is determined by society. A social interpretation which conflicted with the constraints on interpretations defined in the RDF model theory would be a very bad thing for the adoption of RDF. o The directors goal is to cause societies to create a social interpretation of RDF. o The W3C do not have the power to determine any society's social interpretation of RDF, but they can encourage a society to develop one and suggest what it should be. It is up to the society to accept, reject or amend that suggestion. Within this framework I reach the following conclusions: o I disagree with the assertion that social interpretation is a bad thing because it cannot be mathematically/philosophically precise. If that was a requirement, society would have no notion of law and there would be no American constitution. o I disagree with the assertion that any formal language can have no social interpretation. A society is free to define a social interpretation for any language it chooses. o The creation of social interpretation will be done through society's institutions.
Received on Sunday, 2 March 2003 05:40:05 UTC