- From: Brian McBride <bwm@hplb.hpl.hp.com>
- Date: Tue, 18 Jun 2002 13:37:16 +0100
- To: rdf core <w3c-rdfcore-wg@w3.org>
+ a social dimension to RDF + model theory talks about "possible worlds" + but RDF statements have an "intended interpretation" about the human world and web + also, a social context of publication; e.g. personal home page vs software test cases + social context of use: a program invoked to process some RDF will operate as if the raw statements are true. + truth of RDF statements + some logically truth + some true because people agree they are true + semantic web works with both kinds of truth + logical reasoning applying to assumed truths + "expert" reasoning applied to + semantic web deployment + initially + many assumed assertions + new assertions generated by specialized application software + few formalized deductions + evolving toward + more deduced assertions + more formalized deductions by generic reasoners + trust applies not only to individual assertions, but also to different forms of deduction + What does it mean to assert a graph? + Note: may have some bearing on rdfs:isDefinedBy? .......................... Assertion, truth and the social web ----------------------------------- The web is a shared information space that is navigated and interpreted by people, in support of human affairs. A goal of RDF is to lay a path for offloading some elements of this navigation and interpretation to automated processes, by providing a structure of relationships between the objects of discourse that can support formally defined relationships and logical entailments. To support logical entailments, formal RDF meaning is based on a model theory: a set of formal rules that constrain the possible worlds that can correspond to the truth of some given RDF. The notion of truth here is crucial: a possible world may correspond to some RDF if and only if the RDF statement is true in that world. This leads to consideration of what makes a statement be true. The position of formal model theory is that truth is determined a relationship between a statement and a possible world mediated by some formal rules of interpretation. Not all truths about the world and human affairs can be expressed thus in purely logical terms. Some broader notions of truth can be identified: - logical truths: statements which are true (in any MT interpretation) completely by view of their grammatical form and formal interpretation rules of the language. (Quine: all statements with the same gramatical form are true.) These statements actually tell you nothing about the world. - assumed truths: Quine: "statements that a native speaker of the language would regard as obvious". These might be regarded as axioms of the native speaker's world (?). In general, these will be truths that cannot be described by purely logical means, but rather are assessed by social means and institutions. - assumed deductions: determined by some informal process (assumed to be valid) from some other truths that have already been accepted. - mixed truths: truths that can be determined by some logical entailment from some other truths that have already been accepted. It is presumed here that any interesting statement about the world or human afairs must ultimately depend on assumed truths. Having accepted such an assumed truth into one's worldview, other interesting truths may be deduced by logical means. So we need to have a coherent framework for combining assumed truths with sound logical meaning. Semantic web vocabulary gains currency through use, so also do semantic web deductions ultimately have force through acceptance by people. There is a strong social (non-logical) dimension in which semantic web logic must operate. The semantic web aims to grow the use of automated deduction in the web, but it necessarily starts from a low base. The nature of internet-scale deployments is that there must be a migration strategy -- new capabilities must coexist with older capabilities. Thus, the semantic web deals with mixed truths; in early deployments, most will be assumed and a few deduced; as machinery develops, more truths may be deduced. Currently, RDF provides a way to make simple formal assertions, with very little (no?) machinery for formalizing allowable inferences. Inferences are performed by processes, embedded in software implentations, whose validity is not formally demonstrable, and must be assumed or trusted to be valid (in relation to the world and/or human affairs). In time, it is expected that more of these can be expressed formally, and provable deductions by generic software modules can replace the individual implemenations. What does it mean to assert a graph? ------------------------------------ When an RDF graph is asserted in the web, its publisher is saying something about their view of the world. (The mechanism for deciding whether or not a graph is asserted is not discussed here, but it is presumed that the publisher's intent will be clear in some way -- social convention or logical deduction.) There are a couple of socially grounded ideas here: the publisher of a graph, and the social content in which it is published. When a user invokes an application, there is also a social and technical context of invocation that determines some set of RDF assertions that will be assumed to be true: the application itself, and any RDF files that are passed to it. Garbage-in, garbage-out applies: if the initial assumed facts are wrong or meaningless, the results will have little value. No specfic mechanisms for deciding or evaluating the validity of any such assertions are discussed here. An assertion tells us something about "the world" and human affairs, through the normal model theoretic possible-world constraint mechanisms. Some of the truths that are asserted may be logical truths that can be evaluated using logical machinery. Others may be assumed truths that cannot be evaluated logically, but whose truths can be determined by human (non-logical) interpretation. So when we assert an RDF graph, one is stating a constraint on the world of human affairs, saying that both the logically testable and humanly interpretable truths in the graph are indeed true in that world. Noting that there is no single human opinion about the truth of some statements, the graph may further contain commentary for human interpreters to indicate the realm of human interpretation that should be applied. This mans a graph may contain "defining information" that is opaque to logical reasoners. This information may be used by human interpreters of RDF informaton, or programmers writing software to perform specialized forms of deduction in the semantic web. (cue Brian's test case)
Received on Tuesday, 18 June 2002 08:38:16 UTC